@database A4000HardwareGuide @Index IndexNode @author "Warren Block" @master A4000Hardware.guide @$VER: A4000Hardware.guide 4.0 (08/96) @font courier.font 13 @node "Main" Amiga 4000 Hardware Guide 4.0 (08/96) compiled by Warren Block ----------------------------------------------------------------------------- @{" " link "Introduction"} Introduction @{" " link "Drives"} Drives @{" " link "Common Problems"} Common Problems @{" " link "Monitors"} Monitors @{" " link "Common Questions"} Common Questions @{" " link "Sources"} Sources @{" " link "Tips"} Tips @{" " link "Editor"} Editor @{" " link "Internals"} Internals @{" " link "Credits"} Credits @{" " link "Boards"} Boards @{" " link "IndexNode"} Index ----------------------------------------------------------------------------- @{"What's New With This Version" link "What's New With This Version"} @endnode @node "Introduction" Introduction ----------------------------------------------------------------------------- Copyright This document is Copyright © 1996 by Warren Block. Distribution The Amiga 4000 Hardware Guide is freeware. It has been copyrighted to assure its availability to all. Fred Fish and the Aminet are explicitly allowed to include this document in CD-ROM or floppy compilations. Others should contact the @{"Editor" link "Editor"}. Disclaimer Working on computer hardware can be dangerous, both to the computer and to yourself. If you are not a qualified technician, please do not attempt to perform any of these procedures yourself. Neither myself nor any of the people listed in the @{"Credits" link "Credits"} make any claim that any of these tips actually work. In fact, they will probably destroy your computer or your self- confidence. Neither is any claim made that any of the information presented here is error-free, so if you do attempt any of these modifications or fixes and damage yourself or the computer, neither myself nor any of the persons listed in the Credits section will be held responsible. Introduction From The @{"Editor" link "Editor"} The Amiga 4000 Hardware Guide was compiled from online messages posted by many different folks, various hints and tips I've collected elsewhere, and from my own experiences with the 4000, so it is by no means complete. Corrections or additions are welcomed. The purpose of the guide is to help make A4000 troubleshooting easier by gathering all kinds of Amiga 4000 hardware information into a single, easy-to-use guide file. This guide is specifically for the Amiga 4000; however, at the prompting of several other individuals, I've compiled a short A1200 hardware FAQ to address those same old questions I see popping up in c.s.a.hardware all the time. Both are available in the hard/misc directory of Aminet. All of the people who have contributed are listed under @{"Credits" link "Credits"}. I can't thank them enough! At this point in time, information of this type can be very valuable in keeping A4000 systems alive and running, and they have been gracious enough to freely share this information with everyone. Comments on this document should be addressed to the @{"Editor" link "Editor"}. @endnode @node "Common Problems" Common Problems ----------------------------------------------------------------------------- @{" " link "Fan Problems"} Fan Problems @{" " link "Battery Problems"} Battery Problems @{" " link "-5V Power Problems"} -5V Power Problems @{" " link "Zorro-III Problems"} Zorro-III Problems @{" " link "Video Banding Problems"} Video Banding Problems @{" " link "SCSI Reselect Problems"} SCSI Reselect Problems @{" " link "Green Display Problems"} Green Display Problems @{" " link "Dead Machine Problems"} Dead Machine Problems @{" " link "Other Video Problems"} Other Video Problems @{" " link "Slow A2091 Problems"} Slow A2091 Problems @{" " link "Backplane Problems"} Backplane Problems @{" " link "IDE Drive Problems"} IDE Drive Problems @{" " link "Cable Routing Problems"} Cable Routing Problems @{" " link "SCSI Drive Problems"} SCSI Drive Problems @{" " link "Floppy Drive Cable Problems"} Floppy Drive Cable Problems @{" " link "Expansion Cards Not Recognized"} Expansion Cards Not Recognized @{" " link "Memory SIMM Problems"} Memory SIMM Problems @{" " link "MultiFace and FastLane Problems"} MultiFace and FastLane Problems @{" " link "Card Guide Problems"} Card Guide Problems @endnode @node "Fan Problems" Fan Problems ----------------------------------------------------------------------------- A4000 makes rattling noises, fan stalls on powerup, or fan does not turn. Solution: replace fan. Replacement fans: Radio Shack #273-243 or Panasonic FBK-08A12M, available from Digi-Key, Hosfelt, and others (see @{"Sources" link "Sources"}). Please be aware that I've seen two styles of power supplies in the 4000; the fans may differ slightly. Some older 4000s had a power supply with a large hole for the fan, and a bolt-on grill protecting it. This power supply used an actual Panasonic FBM-08A12M. A newer power supply design has a built-in grill formed by holes in the side of the power supply; in this case, the fan is held in place by four odd-looking screws that are tapped directly into the holes in the fan's plastic frame, and the fan itself is a "MAX FLOW" generic. Christopher Laprise has suggested that the A4000 power supply hot air vent is "horribly wrong," and I have to agree. Removing one or more of the horizontal strips of metal can make for a dramatic drop in temperature inside the A4000. I'd suggest using a nibbling tool for this, as it won't leave metal filings inside the power supply. As Chris notes, if you remove more than one section, make them alternate ones to leave something there to keep things from falling into the power supply. @endnode @node "-5V Power Problems" -5V Power Problems ----------------------------------------------------------------------------- Problems with large hard disk transfers, discolored Toaster output (pink or magenta display of white areas), system crashes, or Emplant diagnostic failures. (See @{"Emplant Reference" link "Emplant Reference"}.) Solution: U198 7905 -5V regulator is probably bad. Replace with new 7905 1-amp -5V regulator (see @{"Sources" link "Sources"}). To locate U198, look near the bottom expansion slot at the back of the machine. The easiest way to replace this component may be to clip off the leads close to the body of the defective part, then trim the leads on the replacement and solder it to the old pins. New 7905 regulator Solder (side view) here | Old leads on motherboard \\ | \\ \\ \\ -------- -------| |--- XXXXXXXXX -------- X X X @endnode @node "Video Banding Problems" Video Banding Problems ----------------------------------------------------------------------------- Display on monitor has faint, darker vertical bands or stripes. See the Tips section for the @{"Video Banding Modification" link "Video Banding Modification"}. @endnode @node "Green Display Problems" Green Display Problems ----------------------------------------------------------------------------- Video output from the A4000 has a greenish tint. This may be caused by the Sync On Green jumper (J500) being in the wrong position. Unless the monitor is set up to receive sync signals piggybacked on the Green video input, pins 1-2 of J500 should be jumpered. See Internals/@{"Motherboard Jumpers" link "Motherboard Jumpers"}. @endnode @node "Other Video Problems" Other Video Problems ----------------------------------------------------------------------------- Video problems can be caused by improperly-wired peripheral cables or malfunctioning peripherals, since some signals are shared between the video slot, video connector, and peripheral ports. Disconnect peripherals to isolate video problems to the motherboard. @endnode @node "Backplane Problems" Backplane Problems ----------------------------------------------------------------------------- After adding or removing expansion cards, system no longer boots, displays yellow screen. Solution: too-long resistor and capacitor leads on solder side of backplane daughterboard are bent and shorting together. Remove the backplane and trim leads. @endnode @node "Cable Routing Problems" Cable Routing Problems ----------------------------------------------------------------------------- Make sure that signal and power cables aren't blocking the power supply fan air intake. The floppy ribbon cable can be routed from the motherboard between the power supply connector and the power supply itself. If there is only one drive installed, excess ribbon can be pushed under it. @endnode @node "Floppy Drive Cable Problems" Floppy Drive Cable Problems ----------------------------------------------------------------------------- Many (perhaps most) A4000s were shipped with improperly-wired floppy cables. These cables had wires 3-5 twisted, instead of wires 4-6. These cables will work fine for a single drive, but will not properly connect a second drive. To use two drives that are both jumpered as DS0, the floppy cable should have both wires 4-6 and 10-12 twisted. A cable with only wires 4-6 twisted will require the drives to be jumpered as DS0 and DS1. An improperly-wired or failing floppy cable can cause a high-density drive to only work with double-density disks. Pin-outs for the internal floppy connector are shown in Internals/Connector Pin-Outs/@{"Internal Floppy Connector Pin-Outs" link "Internal Floppy Connector Pin-Outs"}. @endnode @node "Battery Problems" Battery Problems ----------------------------------------------------------------------------- Battery (BT176) is "furry." Batteries can actually leak and destroy part of the motherboard, so replacement of corroded batteries is advised. @{"Dalco" link "Dalco Electronics"}'s 3.6V 3-pin battery, part #46875, is an almost-identical part: the pins and size are identical, but it is rated for 60 mAh rather than the A4000 stock battery's 40 mAh. @endnode @node "Zorro-III Problems" Zorro-III Problems ----------------------------------------------------------------------------- Problems with transfers when using 4091 SCSI-2 controller or other Zorro-III boards. Check for revision of Super Buster; the revision 9 chip had problems with Zorro-III bus arbitration. Revision 11 of the Super Buster fixed this problem. Note that this problem is not due to DMA transfers, but Zorro-III bus arbitration, so it is possible to see it with any Zorro-III board. Problems may be encountered with the @{"A2091" link "A2091 Reference"} or GVP Series II SCSI controller boards. To isolate this problem, check disk transfers to Chip RAM with a program like Diskspeed 4.2. This problem can be fixed by replacing PAL U209 on the @{"A3640" link "A3640 Reference"} daughterboard with a revision 3 version. There can be a software component to these problems, also. Check that libs:68040.library is at least version 37.30. Note that due to the way the library version numbers are handled, version 37.4 is an earlier version. The later 37.30 library correctly handles Zorro-III transfers and mapping of Zorro-III boards (for instance, Commodore's SCSI drivers require that copyback caching be disabled during DMA, and this version of the library does that). @endnode @node "SCSI Reselect Problems" SCSI Reselect Problems ----------------------------------------------------------------------------- System reports "SCSI Bus Phase Error" or copying from one SCSI device to another doesn't seem to work, but copying from the same device to a floppy does work. Check that all SCSI devices support Reselect; if not, disable Reselect mode in the drive's RDB. Some SCSI controllers, notably GVP, let you disable Reselect mode on a per-address basis. @endnode @node "Dead Machine Problems" Dead Machine Problems ----------------------------------------------------------------------------- A4000 does nothing on power-up; keyboard light, power light, and hard drive spin-up show that power supply is working, but screen is gray or black, and disk drive doesn't click. Possible cause: the processor board is not firmly connected to the mother board, or has worked its way loose (the power LED may flash quickly in this situation). Reseat the connector strip that is parallel to the floppy drive. Since this connector is quite long, it is possible for the ends to be firmly attached while the center is not making adequate contact. Be careful not to flex either board too much while reconnecting them; using a large rubber eraser or other flat object will protect your fingers while pressing the boards together gently. In extreme cases, it may be necessary to use a contact cleaner to clean accumulated gunk (like silicon heatsink compound) from the connectors. Possible cause: the motherboard is not supported very well near the internal IDE connector. Solution: gently lift the motherboard near this connector. Installing an insulating support underneath the motherboard near here will be a more permanent solution. Possible cause: badly-trimmed pins of the soldered power supply connector have poked through the insulating plastic under the motherboard and shorted to the case. Solution: trim the pins and install more insulating plastic. Possible cause: power supply failure. A company called Micro R&D repairs and upgrades power supplies. Contact them at: Micro R&D 721 'O' Street Loup City NE 68853 (308) 745-1243 @endnode @node "Slow A2091 Problems" Slow A2091 Problems ----------------------------------------------------------------------------- A2091 SCSI controller in A4000 performs very slowly. Solution: the A2091 controller has problems with the A4000 environment (see @{"A2091 Reference" link "A2091 Reference"} in the Boards section). Several utilities are available in the "hard" directory of Aminet that can help speed up the A2091's performance in an A4000. @endnode @node "IDE Drive Problems" IDE Drive Problems ----------------------------------------------------------------------------- Files larger than about 128K are corrupted when copied to or from an IDE hard drive. Solution: set MaxTransfer for every partition on the hard drive to 0x0001FE00, or 128K. Older IDE drives may even require setting MaxTransfer to 0x0000FFFF, or 64K. It's even possible that some newer drives may need a setting of 0x0000FE00. This problem is due to the way some IDE hard drives respond to requests for large blocks of data. (Note that MaxTransfer is *NOT* "maximum transfer rate," but rather "Maximum Transfer Size.") Some IDE drives may not be consistently ready on powerup; this can cause the A4000 to fail to boot. The machine may just sit there with the drive light on or off, or it may come up with the purple "Insert Floppy" screen. The @{"Seagate ST3144A" link "Seagate ST3144A Reference"} drives do this once in a while; you may hear a clicking as the drive does its self-test then decides it's not up to booting today. It may work after a reset, or may require a power-down to revive it. This problem may occur more often with a master-slave drive setup. @endnode @node "SCSI Drive Problems" SCSI Drive Problems ----------------------------------------------------------------------------- Errors occur on transfers to or from SCSI hard drives, CD-ROM drives, or other SCSI devices: Possible causes: improper termination or bad cables. See Drives/@{"SCSI Examples" link "SCSI Examples"} for example SCSI configurations. @endnode @node "Expansion Cards Not Recognized" Expansion Cards Not Recognized by Rhett Rodewald ----------------------------------------------------------------------------- Problems with cards not being recognized? Your problem may be the auto- config chain. This works similarly on most Amigas, but this description is very precisely intended for the Amiga 4000. (Not the Tower, but I'm sure it's similar.) First some background. I originally intended only to add SCSI to my Amiga 4000. My friend Bill lent me an old GVP SCSI +8 card from his A2000. (He now uses an A3000, and so had little need for it.) I checked the manual carefully, set the jumpers, and plugged it into my 4000. It just hung up on boot. Remove the SCSI card, and all is well, but it will not boot with the card in. Looking at the manual, I noticed that one jumper on the card, next to the described jumpers for setting the amount of RAM was undocumented. After trying many things, I tried shorting that jumper. I plugged in the card, and booted. This time it booted, great! (or so I thought) But the card was unseen by the Amiga. Oh, well, I removed the card, and put my old cards back in. Boot up and ***AAAAAAARRRRRRRRRGGGGGGGGGHHHHHHHH***... NO CARDS! My A4000 functioned flawlessly, but it would not recognize a single card. Eventually I traced it to the autoconfig and realized that I had actually *burned a trace* on the daughtercard. Not one that can be seen, mind you, but one on an inside layer. After determining how auto-config *should* work, (and replacing the 7407 chip, which likely didn't need to be replaced) I soldered a jumper across where I inferred the blown trace should be and, lo and behold, it worked! While I hope that no-one else has this problem, there are a few things that may be learned from my experience. 1) ***DO NOT, UNDER ANY CIRCUMSTANCE, PLAY WITH UNDOCUMENTED JUMPERS!*** (Unless you know exactly what you're doing... or if you are trying to configure an unmarked PC card... or if you *really* want to... but don't blame me (especially don't sue me.) -- YOU HAVE BEEN WARNED!) Dave Haynie has stated (I'm paraphrasing) Don't play with jumpers unless you know exactly what they do, Even if they are marked as "Free sex and beer" 2) An understanding of how Autoconfig works, which is the main point of this article. Auto-config is a wonderful thing, it keeps us from pulling our hair out trying to get a new card to work without hardware conflicts. (We still have to deal with installing support software if the card needs it, but it is still a million times easier than configuring a PC.) If you have access to the Amiga Hardware reference manual, you will find an excellent explanation in the back in the appendicies. There are many issues involved when designing a card, and they are detailed here. However, since we aren't designing a card, only two paragraphs apply. Amiga autoconfiguration is surprisingly simple. When an Amiga powers up or resets, every card in the system goes to its unconfigured state. At this point, the most important signals in the system are /CFGIN and /CFGOUT. As long as a card's /CFGIN line is negated, that card sits quietly and does nothing on the bus. ... As part of the unconfigured state, /CFGOUT is negated by the PIC immediately on reset. The configuration process begins when a card's /CFGIN line is asserted, either by the backplane, if it's the first slot, or via the configuration chain, if it's a later card. The configuration chain simply ensures that only one unconfigured card will see an asserted /CFGIN at one time. ... (ed. note: after the system has read and configured this card) ... will cause the PIC to assert its /CFGOUT, enabling the next board in the configuration chain. (Quoted from the Amiga Hardware Reference Manual, 3rd edition, page 431) The magic is in how the configuration chain works, or if you have a situation like mine, in how it does not work. Here's what to check. NOTE: The following applies to my A4000. Errors may have occured in the entry, transmission, or otherwise, and I will not be held responsible. In addition, do not attempt to actually repair or even trouble-shoot your Amiga unless you are confident of your electronics skills and willing to take full responsibility for your actions. Your Amiga may differ from mine, and so the following may apply only in concept, not particulars. Make sure that you understand not only what is connected to what, but why it works as it does. This is only an overview of *one* system, and should not be construed as being a common problem or necessasarily related to your machine/problem. Note that even trouble-shooting without getting out your soldering iron can still cause damage-- Especially when the machine is on! The Autoconfig chain is on the daughterboard (riser) card in the 4000. The daughter card plugs into the motherboard using two seperate connectors. The connector towards the rear of the 4000 carries the signals for the video slot, among others. The connector towards the front of the A4000 is basically a single Zorro III slot. (Note: Do NOT unplug the riser card, and plug a Zorro card directly into this slot. Although it *may* work, the proper bus termination is not present, and I have not personally verified that the signal connections are 100% compatible.) When you reset (or start) your Amiga, it starts by asserting the /CFGIN line on this connector. (Note: Signal preceded by an "*" or "/" are active-low. This means that 0V means active, and 5V means inactive. Therefore, on power up, all /CFGIN lines should be 5V except for the first slot, (the active one) that will be 0V.) This line should be connected to pin 12 on the first slot. Note: Pins are numbered starting at pin 1, towards the back of the Amiga. If you look at the silk-screen on the Motherboard, you should see a large "1" near the corner of the slot. Pin 2 is directly across from pin 1. Therefore, all the odd pins (1, 3, 5, ...) are on one side, and all the even pins are on the other side. Pin 1 usually has a square solder pad (if you look at the back of the board) instead of a round one. This is useful on the daughter card. Disassemble your 4000, and pull out the daugher card. (If you have trouble doing this, you probabally should *NOT* be doing any of this.) Check continuity between pin 12 of the plug in edge and pin 12 of the first slot connector. Pin 12 is the sixth pin on the back-side of the riser, counting from the middle gap towards the front of the Amiga. (The even pins are on the back side, so count 2, 4, 6, 8, 10, 12 to find pin 12.) Note: The first slot is the bottom one, and the last slot is the top one. This is the trace that I burned when playing with the SCSI card. If you've managed to do the same thing, just solder a jumper between these two pins. (Don't get the wrong pins!!!) When the first card gets the config-in signal, it goes through the configuration process with the Amiga, and then, when done, it asserts /CFGOUT on pin 11. Pin 12 (enable) and Pin 11 (done) are fed to an "or" gate. Note that since these are active-low signals, both must be low before the output becomes low. (The "or" effectively works like an "and" gate because of the inverted signals.) This output is fed to the next slot's /CFGIN and the next or gate. This process is repeated for each card until the last one, and then the output from the final /CFGOUT is fed back to the motherboard's /CFGOUT, indicating that there are no more cards to be configured. Take note that this doesn't have any particular limitations on how many card slots there are. This is why you can buy super-risers that have more slots available. It's a chain that ends when it gets back to the motherboard. If for some reason this chain is broken, your Amiga should continue to function, but it will not recognize all the cards in the slots. Here's a diagram. (Slot 4, pin 11) (Slot 3, pin 11) \\ ____ (Slot 2, pin 11) \\ ____ \\---(4)\\ \\ \\ ____ \\---(1)\\ \\ )OR >(6)-+ ____ \\---(12)\\ \\ )OR >(3)---+---(5)/___/ | A--(10)\\ \\ )OR >(11)--+---(2)/___/ | | )OR >(8)---+--(13)/___/ | | | B---(9)/___/ | | (Slot 4, pin 12) | | (Slot 3, pin 12) | (Slot 2, pin 12) | | /| | (Pin 11 on the card edge.)---------(7407,pin 12)--< |--(7407,pin 13)------+ \\| (7407, Driver chip) Notes: A = /CFGIN = Pin 12 of the card edge, also connected to Slot 1, pin 12 B = /CFGOUT(1) = Slot 1, pin 11 All numbers in parenthesis indicate pin numbers on the 74HCT32N chip. In English: /CFGIN (pin 12 on the card edge *and* pin 12 on slot 1) connects to the input of the first "or" gate (pin 10 on the 7432 chip). /CFGOUT(1) (Slot 1, pin 11) connects to the other input of the same gate (pin 9 on the 7432 chip). The output from this gate (7432, pin 8) connects to both the /CFGIN(2) (Slot 2, pin 12) and the input to the next "or" gate (7432, pin 13)... and so on. If you suspect that your auto-config chain is screwed up, check all the direct connections in the diagram, and see if any are broken. Finally, if you still haven't found anything, install the riser and check the inputs and outputs at each card slot. Do this *very* carefully! Sticking a probe into a running computer can easily short together two wires that were never meant to touch, and *really* blow something up. While any voltages on the motherboard are at most +/- 12V (mostly +5V) and should not harm you, they will harm your computer if the wrong wires are shorted together. If you don't understand how to trouble-shoot the above circuit, please take your computer to someone who does. Here's how it works. At reset, all /CFGIN lines (pin 12 on each slot) are pulled high. (with a pullup resistor) All unused (no card in that slot) /CFGOUT lines (pin 11 on each slot) are pulled low. If there is no card, in the slot, the low signal coming in on (A in the diagram) /CFGIN is immediately "or"ed with the low /CFGOUT and propagates the signal to the next slot. If a card is present in a slot, it pulls the /CFGOUT line high (inactive) until after it gets a /CFGIN signal, and configures itself with the system. Then it releases /CFGOUT so the card in the next slot may configure itself. Finally, the /CFGOUT signal of the last card is fed back to the motherboad's /CFGOUT, and it knows that configuration is done. This hopefully provides some insight to the autoconfig circuitry, and may be of use when trouble-shooting board-recognition problems. Best of luck. @endnode @node "Memory SIMM Problems" Memory SIMM Problems ----------------------------------------------------------------------------- SIMM sockets may develop solder fractures or internal breaks that will cause problems, either when the machine warms up or when it tries to boot. Resoldering may help, or it may require replacement of the socket. The plastic-tabbed SIMM sockets used in the A4000 are fragile, and if the tab breaks off, you'll need to either replace the socket or come up with some other mechanical method to hold the SIMM firmly into the socket. @endnode @node "MultiFace and FastLane Problems" MultiFace and FastLane Problems ----------------------------------------------------------------------------- When using both the MultiFace III and FastLane boards in the A4000, the FastLane should be in a lower slot, closer to the motherboard. If the MultiFace board is lower than the FastLane, using the MultiFace software's PIT0: device will cause the machine to lock up. See also: @{"MultiFaceCard III Reference" link "MultiFaceCard III Reference"} and @{"FastLane Reference" link "FastLane Reference"}. @endnode @node "Card Guide Problems" Card Guide Problems ----------------------------------------------------------------------------- Some of the newer A4000T models come with oddly shaped card guides that can damage components mounted on expansion boards. Phil Wright provides this diagram: /\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/ | | | | | | | | | | | | | |B| | |A| | |C| | |K| | Zorro Card |P| | Seen From Above |L| | (NOT TO SCALE) |A| | |N| | Pins of IC come _____ |E| | dangerously close _|I|_ |_| | to card guide------> _|C|_ || | _____ _____ _|_|_ || | ______| |_______| |______ || |_| Plastic Card Guide |__|| __|_____________________________|___| @endnode @node "Common Questions" Common Questions ----------------------------------------------------------------------------- CIA Questions Question: Which CIA is associated with the parallel port? Answer: U350 is the parallel port CIA; the other CIA is U300. Question: Can the surface-mount CIAs used in the A4000 (and A1200) be removed and replaced with sockets for easier replacement in the future? Answer: [Courtesy of Dale Currie of AmiTrix] "These sockets are sometimes called Ultra Low Profile SMT sockets for PLCC Packages. The ones we get are AUGAT PCS-SMU Series. I believe ASSMANN also make them now, called SMT-PLCC-sockets. There are probably others. They are only about half the height of a normal PLCC socket, and have flat pins turned underneath with slots in the bottom of the socket all around the inside edge, such that the pins sit on the SMT pads and you can solder them through the slots in the bottom of the socket. Needless to say, it's a very delicate operation, requiring a good iron with a very fine tip, a steady hand, and quite often the assistance of a magnifying glass/lamp. Getting the CIAs off intact in the first place is another story. That requires some rather expensive gear (we use a PACE station, $1500-2000 CAN with all the atachments), but does a nice job. The only other way destroys the existing chips by cutting their pins and removing them one by one after the chip body has been freed. I wouldn't recommend it, unless you're very experienced, it's quite easy to destroy a pad/trace by overheating." Mr. Currie may be contacted at AmiTrix Development 5312 - 47 Street Beaumont, Alberta, T4X 1H9 Canada Phone or Fax: 1+ 403-929-8459 Email: sales@amitrix.com or support@amitrix.com http://www.networkx.com/amitrix/index.html ----------------------------------------------------------------------------- Memory Questions Question: What type of memory does the A4000 use? Answer: The A4000 comes with 2M of Chip RAM, either in a single SIMM or surface-mounted on the Motherboard. There are four SIMM sockets for expansion memory (Fast RAM). These sockets hold 72-pin SIMMs, either 1M or 4M in capacity, 80 ns or faster. It is not possible to mix 1M and 4M sizes, although it has been reported that 8M SIMMs can be used in place of two or four 4M SIMMs. To fit properly, these SIMMs must be single-sided modules. The total motherboard Fast RAM limit is 16M, regardless of SIMM combinations. (These specifications describe the motherboard memory; expansion boards may use other types of memory.) Question: Can the A4000 use 36-bit SIMMs, instead of 32-bit? Answer: Yes. The extra parity bits are ignored. ----------------------------------------------------------------------------- Floppy Drive Questions Question: Why does the floppy light flash every so often, even with no disk in the drive? Answer: This is a result of the system polling the drive to see if a floppy has been loaded. Question: Why doesn't my second floppy drive work in high-density mode? Answer: Check for the proper setting of jumper J351 (see Internals/ @{"Motherboard Jumpers" link "Motherboard Jumpers"}). See also @{"Floppy Drive Cable Problems" link "Floppy Drive Cable Problems"}. Question: Will normal PC-type *double-density* (720K) 3.5-inch floppies work in the A4000? Answer: Yes, but you may encounter a couple of problems. First, many PC-type floppies are not jumpered to support a diskchange signal. Enabling this may be as easy as moving a jumper, or it may require unsoldering, moving, and resoldering a surface-mount part. Secondly, many of the PC-type floppy drives connect the diskchange signal to pin 34 of the connector; however, the Amiga expects this signal on pin 2. A re-routing of the conductors in the cable can solve this, or you can use the DiskChange command to manually notify the system of disk changes. Question: Will normal PC-type *high-density* 3.5-inch floppies work in the A4000? Answer: No. [The following text courtesy of Gene Heskett] A commodity PC drive runs at normal spindle speed for the drive, or 300 rpm. To move data in and out of it in high-density mode requires a 500 kilobaud data pump in the floppy path. The Amiga chips are only able to handle around 400 kilobauds. The Amiga actually runs its floppy data rate at the older double density standard of 250 kilobauds. To do high density on the older drives, special drives were ordered by Commodore that could run a fairly stable spindle speed of 150 rpm. If you watch it, you'd swear the drive was going to stop, it's that slow. These are the high-density drives for an Amiga, and until the chips get a refresh for a higher data rate, are the only type of drives that can be used in high density mode on an Amiga. [Editor's note: some people have managed to modify standard drives; however, all reports indicate that these perform unreliably at best.] I might add that since these custom Amiga drives run at half speed, the read signal from the head is only about one-fourth of what a standard drive has, and they do require an electrically quiet environment for a usable error rate. There are a couple of other problems with using the PC drives, too: lack of a ready signal for the automatic diskchange detection being one of them, and the lack of talkback identify to tell the Amiga what kind of a drive it is is another. Even if you could diddle the spindle speed down to 150 RPM (as its digital, that's doubtfull) to make a high-density drive out of it, you'd still have to have a custom driver that puts twice as many sectors on a track. The Amiga would otherwise only do the normal double density sectors/track, and simply fill the remaining space up with "trailer" bytes till the next index pulse came by, thereby wasting half the track. The Amiga defaults to double density if the drive doesn't talk back. ----------------------------------------------------------------------------- IDE Hard Drive Questions Question: Can the A4000 support two IDE hard drives? Answer: Yes, since IDE supports a master/slave drive setup. Make sure the @{"drive jumpers" link "Seagate ST3144A Reference"} are set properly for two drives. You might have problems with two different brands of drives working together; this depends on the age and type of the drives. Question: Can the A4000 use IDE hard drives larger than 512M? Answer: Yes. The supposed "limit" of 512M is a limitation of the BIOS in MS-DOS machines, and the A4000 is not subject to this limit. The maximum supported partition size is 2G, and the maximum drive size is 4G. If you want to fully use larger drives, you'll have to look into alternative filesystems like AFS. Question: Can "EIDE" or "Fast ATA" hard drives be used in the A4000? Answer: Yes. These are just different names for revised versions of IDE, and should work fine with the A4000's on-board IDE controller. ----------------------------------------------------------------------------- SCSI Hard Drive Questions Question: Does the A4000 come with a SCSI or SCSI-2 hard drive controller? Answer: No. The A4000T (tower) model did (and does) come with a built-in SCSI-2 controller, though. The A4091 and FastLane expansion boards are common Fast SCSI-2 controllers for the A4000, and the @{"A2091" link "A2091 Reference"} is a fairly common SCSI-1 controller. Question: Why doesn't SCSI work on the A4000? Answer: It does. But because of a bug in early versions of the Zorro-III DMA controller (the "Buster" chip), DMA SCSI controllers didn't work properly. This problem can be fixed (by replacing the early revision 9 Buster with a revision 11 version) or avoided altogether (by using a SCSI controller that doesn't use the Zorro-III bus, like the one built into the @{"Warp Engine" link "Warp Engine Reference"} accelerator). Question: What are SCSI terminators? Answer: Terminators are resistor packs attached at both ends of the SCSI chain. The resistance reduces ringing and noise on the bus, and is necessary for reliable operation. See @{"SCSI Examples" link "SCSI Examples"}. Question: What is the proper setup for SCSI termination? Answer: Both ends of the SCSI chain need to have terminators, and there should be none in the middle. See @{"SCSI Examples" link "SCSI Examples"}. Now, the catch: some SCSI devices are not very compliant with the SCSI specification, and require oddball setups. Devices made within the last couple of years are usually pretty compliant. Question: Can a SCSI-1 drive be used with a Fast SCSI-2 controller? Answer: Yes, and the reverse will also work. SCSI-1 and SCSI-2 (and Fast SCSI-2) are compatible. Of course, a SCSI-1 drive won't go any faster when plugged into a Fast SCSI-2 controller; neither will a Fast SCSI-2 drive connected to a SCSI-1 controller go any faster than SCSI-1. See @{"SCSI Examples" link "SCSI Examples"}. Question: Why does my hard drive "stutter" every fifteen minutes or so, without the drive light coming on? Answer: The drive is performing thermal recalibrations ("t-cals") to make certain that the heads remain centered over the tracks as the platters heat up. The so-called "AV" drives spread out t-cals so this momentary interruption doesn't occur during "live" video or sound recording or playback. Question: What is the pin-out for a DB25 SCSI connector? What about the standard 50-pin SCSI header? Answer: See the Drives/@{"SCSI Pin-Outs" link "SCSI Pin-Outs"} section for both. Question: Will SCSI hard drives meant for the Mac work on an A4000? Answer: Yes. The cable included with most external Mac hard drives is a DB25-to-Centronics 50 adapter, and this will work on a SCSI controller with DB25 external SCSI port. Software is available for accessing an existing Mac filesystem, so file interchange with a portable SCSI device is possible. ----------------------------------------------------------------------------- CD-ROM Questions Question: Will an EIDE CD-ROM drive work with the A4000's IDE interface? Answer: Yes. See the disk/misc and disk/cdrom sections of Aminet for drivers to work with these CD-ROM drives. Beware of pseudo-IDE CD-ROM drives, like the older Mitsumi, Panasonic, and Sony drives, which will not work unless you have a special interface board for them. @endnode @node "Tips" Tips ----------------------------------------------------------------------------- @{" " link "Connecting VGA Monitors"} Connecting VGA Monitors @{" " link "Video Banding Modification"} Video Banding Modification @{" " link "Processor Board Mounting"} Processor Board Mounting @{" " link "Processor Cooling"} Processor Cooling @{" " link "Speeding Up IDE Boot-Up"} Speeding Up IDE Boot-Up @endnode @node "Connecting VGA Monitors" Connecting VGA Monitors ----------------------------------------------------------------------------- VGA monitors can be connected to the A4000; however, since the special circuitry in the Amiga video output can mistakenly identify a monitor as a genlock and thus cause problems, a special cable or adapter is the best way to hook them up. All this cable really does is buffer the horizontal and vertical sync signals by double-inverting them through a TTL gate. Commodore's DB23-to-HDD15 adapter (supplied with most A4000s) used a 74HCT08 for this, but you can use a 74LS04 or other low-power TTL-level inverters (or other gates wired as inverters, of course). Remember that most VGA monitors won't sync at 15.75 kHz, so you'll have to set the A4000 to use the Double NTSC (or Double PAL) or Multiscan video modes. Even using these modes, the video output may not work with some VGA monitors, since some of the modes use horizontal sync frequencies as low as 23 kHz, and normal VGA starts at 31.5 kHz. Using the VGAOnly monitor driver will bump these frequencies up a bit, perhaps enough to make them usable with picky monitors. A4000 VGA Video Cable A4000 DB23 VGA HDD15 Red (Pin 3) ----------------------------------- Red (Pin 1) Green (Pin 4) ----------------------------------- Green (Pin 2) Blue (Pin 5) ----------------------------------- Blue (Pin 3) Grounds (Pin 16-20) ----------------------------- Grounds (Pin 5-8,10,11) and ground for inverter. Also use a .01 uF ceramic decoupling cap between +5V and ground at the chip power supply pins. |\\ |\\ Horizontal Sync (Pin 11) ----| >o---| >o--------- Horizontal Sync (Pin 13) |/ |/ |\\ |\\ Vertical Sync (Pin 12) ------| >o---| >o--------- Vertical Sync (Pin 14) |/ |/ +5V (Pin 23) ------------------------------------ Power supply for inverter chip. @endnode @node "Video Banding Modification" Video Banding Modification ----------------------------------------------------------------------------- Video Banding Modification: Many (most?) A4000s show some faint vertical bands on the display. There is a modification to prevent this, but, like anything else, it has advantages and disadvantages. Advantages Removes vertical bands from display. Simple modification. Relatively easy to disable. Disadvantages Interferes with operation of attached genlocks and deinterlacer boards like the Amber A2320 board (see Boards/@{"A2320 'Amber' Reference" link "A2320 Reference"}). The modification involves connecting pin 15 of the video port connector through a 100 ohm, 1/4 watt resistor to ground. This can be done inside the A4000, but it's safer and more versatile if the modification is performed on the DB23-to-HDD15 adapter that was included with the A4000. That way the adapter can be unplugged from the machine if a genlock or Amber board is to be used. There are two grounds on the adapter PC board: shield ground and video ground (pins 16-20). Since all the other components on the board use the video ground, it seems reasonable to use it for this modification, rather than the more-easily-reached shield ground. The resistor fits nicely on the bottom of the adapter PC board, running from pin 15 to pin 18 of the connector's soldered pins. @endnode @node "Processor Board Mounting" Processor Board Mounting ----------------------------------------------------------------------------- The nylon standoffs that hold the A3640 processor daughterboard in place grip the board very, very tightly, making removal difficult and prone to flexing this fragile board. Pushing the two halves of the standoff together with needle-nose pliers makes it somewhat easier to remove the processor board, but it may still be a fight, putting stress on the motherboard, the processor board, and the person removing it. After removing the board, you can clip off one "prong" of the side that plugs into the processor board. If you leave one of these prongs in place, the processor board will still be held firmly in place, but removal will be much easier, next time, anyway. If you are afraid that the processor board will not be held firmly enough, do this modification on only two of the standoffs. -- -- / | | \\ / | | \\ <-- Clip off one "prong" of this \\ | | / side to make processor board \\ | | / removal easier. -------------- | | -------------- | | | | | | | | | | | | | | | | ----------------------- | | <-- Note that the larger disk is the ----------------------- bottom end (clips into motherboard). / | | \\ / | | \\ \\ | | / \\ | | / -- -- @endnode @node "Processor Cooling" Processor Cooling ----------------------------------------------------------------------------- A company called PC Power & Cooling makes a stick-on cooling fan for the 486 and Pentium processors that just happens to fit the 68040 perfectly. The fan uses a stick-on backing to attach to the chip, so you don't have to mess with little clips or spring wires. It spins on ball bearings instead of a bushing, so life expectancy is claimed to be 50,000 hours instead of the ordinary bushing fan's 5000 hours. A disk-drive power connector powers the fan. Finally, it's only .6 inches high, no more than the heat sink that comes with the A3640 processor board. PC Power & Cooling calls it by several names; the best one to use is the "PENTA-COOL 54" for the 90 MHz Pentium or the "CPU-COOL 1.9." Some earlier versions of the fan had "corners" cast into the aluminum heat sink to better fit the 486; these versions don't fit the 68040 without modification, so ask for either of the versions above, or make sure they know that the one you're purchasing must be of the flat variety. In the United States, you can usually save a few dollars by ordering these fans from large mail-order companies like PC Connection, but please make sure of the model and brand first. PC Power & Cooling 5995 Avenida Encinas Carlsbad CA 92008 (800) 722-6555 (619) 931-6988 Fax @endnode @node "Speeding Up IDE Boot-Up" Speeding Up IDE Boot-Up ----------------------------------------------------------------------------- Even when an A4000 is configured without one, the system will still look for an IDE hard drive on power-up or reset. To allow IDE drives time to spin up, there is a long search built into the 3.0 ROMs, and an even longer one in the 3.1 ROMs (about 30 seconds). A BattMem bit to disable the extra delay with 3.1 ROMs has been reported and finally confirmed, but it only enables or disables eight seconds of the wait, and defaults to off. With software, it is possible to remove the delay between resets, but not on power-up (Matthew Frost's "NoIDE" program is available from the disk/misc directory of Aminet). The only currently-known way to completely disable this delay is by building a small piece of hardware (the circuit is courtesy of Sean Riddle; the instructions are mine, so errors are entirely my fault). With this "IDE ignorer," the search for IDE drives is skipped altogether. Parts Required 2 4.7K resistors 1 40-pin crimp-on IDC connector 1 male-to-male 40-pin header (see below) Procedure: The two resistors are connected between pin 39 and pin 3, and pin 39 and pin 5 (the effect is to tie IDE data bits 7 and 6 high). Rather than using a piece of whole 40-conductor ribbon cable, I stripped short pieces of individual conductors from one and used fine needle-nose pliers to press these wires into the contacts for pins 3, 5, and 39. Then the crimp-on cover was clipped on top of these. Since the relevant pins are at either end of the connector, I ran the resistors flat along the side of the connector, using heat-shrink tubing and tape to insulate the leads. The only 40-pin male-to-male header I had didn't have pins long enough to make contact with both connectors, so I just cut three pins off a wire-wrap socket and stuck them in the connectors at pin locations 3, 5, and 39. Physically, this seems to be strong enough to hold the two connectors together. _______________________________ |-----------####----------------| <--two 4.7K resistors on side |---------------####------------| of 40-pin IDC connector \\_______________________________/ ||||||||||||||||||||||||||||||| =============================== <--40-pin male-to-male header ||||||||||||||||||||||||||||||| _______________________________ / \\ | | <--A4000 IDE cable connector |_______________________________| @endnode @node "Internals" A4000 Internals ----------------------------------------------------------------------------- @{" " link "Motherboard Jumpers"} Motherboard Jumpers @{" " link "Connector Pin-Outs"} Connector Pin-Outs @{" " link "Power-Up Self-Test"} Power-Up Self-Test @{" " link "Keyboard Self-Test"} Keyboard Self-Test @{" " link "Definitive Buster"} Definitive Buster @endnode @node "Motherboard Jumpers" Motherboard Jumpers ----------------------------------------------------------------------------- A4000 Motherboard ------------------------------------------------ | ooo || oo | | | Internal || DF1: | | | Audio || Enable | | | Connector || | Power | | || | Supply | | :: Mystery || | | | :: Header || ooo | | | :: (see below) || ooo | Fan | | || ooo -------------| | || Power Supply | | || Connector | | || | | || -------------------| | || | | | || | | | || | | | || | Drive | | || | Bays | | || oo | | | || SIMM | | | || Size | | | || | | ------------------------------------------------ Jumpers J351: DF1 Enable Closed: Enable low-density (880K) floppy as DF1. Open: No DF1 *OR* for high-density (1.76M) DF1. J852: Fast RAM SIMM Size (Chip RAM is always 2M.) 256K: 1M SIMMs. 1M : 4M SIMMs. CN404: Internal Audio Connector Audio signals attached here will be mixed with the A4000 audio output. The A4000 audio is somewhat louder than the normal line level output from most CD-ROM drives, presumably to make sound effects audible over background music. Setting the software- controlled A4000 audio level lower (to 32 instead of 64) will help match the levels. Pin 1: Audio In (Left) Pin 2: Ground. Pin 3: Audio In (Right) "Mystery" Header (courtesy of Dave Haynie) "There is a 12 pin header (DIL, J975) near the mouse ports. This feeds the 'extra' shift register in Lisa. Unlike OCS/ECS systems, which simply multiplex the four quadrature signals from each mouse port with one another, the AA systems serialize all eight bits of mouse data. While they were at it, the AA designers added a second 8-bits to the mouse port registers. These don't hook into mouse logic or anything, but they can be read by the CPU. So we used them for configuration. The first two bits (not on that header) tell the OS about the AA system -- is it 16 or 32 bit, is it single or double access per cycle. These are hard-wired on the motherboard. The remaining six bits read high when unjumpered, low when jumpered. I had recommended using one to say 'VGA only,' but there wasn't enough ROM space." Other Jumpers (Not Shown Above) J100: CLK90 Clock Source 1-2 Closed: Internal (68020/68030) 2-3 Closed: External (68040) J104: CPU Clock Source 1-2 Closed: Internal 2-3 Closed: External J151: ROM Speed: 160 or 200 ns ROMs. 1-2 Closed: 200 ns ROMs (default). 2-3 Closed: 160 ns ROMs. J213: Chip RAM: 2M or 8M 1-2 Closed: 2M Chip RAM (default). 2-3 Closed: 8M Chip RAM. This option was apparently for use with the never-released AAA chip set, and won't work in a normal A4000. J500: Sync On Green 1-2 Closed: Sync on green disabled (default?). 2-3 Closed: Sync on green enabled (see the Common Problems section for the @{"Green Display Problems" link "Green Display Problems"} note on this jumper). J501: Lisa Sync (Wide input on the Lisa chip.) 1-2 Closed: CSync from Agnus Pin 80. 2-3 Closed: +5V (default). J502: Select DAC Sync 1-2 Closed: DAC syncs on green. 2-3 Closed: DAC uses standard signal (default). J850: Enable DSACK (Used with 68020) 1-2 Closed: DSACK Enabled for 68020. U860 and U152 also required. 2-3 Closed: No DSACK. "Haynie Kludge" Jumper (courtesy of Dave Haynie) "That jumper enables the "early sizing" mode on the Zorro III bus. One of the complaints about Zorro III is that the size of a transfer isn't known until the data phase of the cycle. So in the Zorro III addendum I added an optional mode that allows data size information to be latched in the address phase. Just in case any existing boards have a problem with this (if they go by the spec, they don't, but who really knows what they're doing), it's shipped disabled. The idea was to test this out in the A4000T, bless it as standard for future Zorro III controllers as long as it did what designers wanted." @endnode @node "Connector Pin-Outs" Connector Pin-Outs ----------------------------------------------------------------------------- Note: Signals shown with a star (*) in front of them are active-low. Please check the pin-out information shown here with a meter before using it. See also: Drives/@{"SCSI Pin-Outs" link "SCSI Pin-Outs"}. ----------------------------------------------------------------------------- @{" " link "Serial Port Pin-Outs"} Serial Port @{" " link "Parallel Port Pin-Outs"} Parallel Port @{" " link "Video Port Pin-Outs"} Video Port @{" " link "Keyboard Port Pin-Outs"} Keyboard Port @{" " link "Joystick Port Pin-Outs"} Mouse/Joystick Ports @{" " link "External Floppy Port Pin-Outs"} External Floppy Port @{" " link "Internal Floppy Connector Pin-Outs"} Internal Floppy Connector @{" " link "Internal IDE Hard Disk Connector Pin-Outs"} Internal IDE Hard Disk Connector @{" " link "VGA Monitor Pin-Outs"} VGA Monitor Pin-Outs @{" " link "Power Supply Pin-Outs"} Power Supply Pin-Outs @endnode @node "Serial Port Pin-Outs" Serial Port (DB25 Male) ----------------------------------------------------------------------------- Pin 1: Shield Ground Pin 2: Transmit Data Pin 3: Receive Data Pin 4: RTS Pin 5: CTS Pin 6: DSR Pin 7: Data Ground (Do not connect to shield ground.) Pin 8: CD Pin 9: +12V (20 mA maximum.) Pin 10: -12V (20 mA maximum.) Pin 11: Amiga Audio Out (Left) Pin 12: Unused Pin 13: Unused Pin 14: Unused Pin 15: Unused Pin 16: Unused Pin 17: Unused Pin 18: Amiga Audio In (Right) Pin 19: Unused Pin 20: DTR Pin 21: Unused Pin 22: RI Pin 23: Unused Pin 24: Unused Pin 25: Unused @endnode @node "Parallel Port Pin-Outs" Parallel Port (DB25 Female) ----------------------------------------------------------------------------- Pin 1: *Strobe Pin 2: Data 0 Pin 3: Data 1 Pin 4: Data 2 Pin 5: Data 3 Pin 6: Data 4 Pin 7: Data 5 Pin 8: Data 6 Pin 9: Data 7 Pin 10: *Acknowledge Pin 11: Busy Pin 12: Paper Out Pin 13: Select Pin 14: +5V Pull Up (10 mA maximum.) Pin 15: Unused Pin 16: *Reset Pin 17: Ground (Do not connect any of these grounds to a shield.) Pin 18: Ground Pin 19: Ground Pin 20: Ground Pin 21: Ground Pin 22: Ground Pin 23: Ground Pin 24: Ground Pin 25: Ground @endnode @node "Video Port Pin-Outs" Video Port (DB23 Male) ----------------------------------------------------------------------------- Pin 1: *External Clock Pin 2: *External Clock Enable (47 ohm) Pin 3: Red Video (75 ohm) Pin 4: Green Video (75 ohm) Pin 5: Blue Video (75 ohm) Pin 6: Digital Intensity (47 ohm) Pin 7: Digital Blue (47 ohm) Pin 8: Digital Green (47 ohm) Pin 9: Digital Red (47 ohm) Pin 10: *Composite Sync (47 ohm) Pin 11: *Horizontal Sync (47 ohm) Pin 12: *Vertical Sync (47 ohm) Pin 13: Ground Return (Digital ground return for pin 2.) Pin 14: *Pixel Switch (Genlock overlay, 47 ohm) Pin 15: *Clock Out (47 ohm) Pin 16: Video Ground (Do not connect any of these grounds to pin 13.) Pin 17: Video Ground Pin 18: Video Ground Pin 19: Video Ground Pin 20: Video Ground Pin 21: -5V (10 mA maximum.) Pin 22: +12V (100 mA maximum.) Pin 23: +5V (100 mA maximum.) @endnode @node "Keyboard Port Pin-Outs" Keyboard Port (6-Pin Female Mini-DIN, PS/2 Type) ----------------------------------------------------------------------------- Pin 1: Data 6 --- 5 Pin 2: Unused Pin Layout: | | Pin 3: Ground (Index key | | Pin 4: +5V (100 mA maximum.) at top.) 4 --- 3 Pin 5: Clock Pin 6: Unused 2 1 Note: A PS/2 keyboard will not work with the A4000. @endnode @node "Joystick Port Pin-Outs" Mouse/Joystick Ports (DB9 Male) ----------------------------------------------------------------------------- Mouse: Light Pen: Pin 1: Mouse Vertical Pin 1: Unused Pin 2: Mouse Horizontal Pin 2: Unused Pin 3: Mouse Vertical Quadrature Pin 3: Unused Pin 4: Mouse Horizontal Quadrature Pin 4: Unused Pin 5: Mouse Button 3 (Middle) Pin 5: Light Pen Press Pin 6: Mouse Button 1 (Left) Pin 6: *Light Pen (Capture Beam Pos) Pin 7: +5V (50 mA maximum.) Pin 7: +5V (50 mA maximum.) Pin 8: Ground Pin 8: Ground Pin 9: Mouse Button 2 (Right) Pin 9: Unused Digital Joystick: Analog (Proportional) Joystick: Pin 1: *Forward Pin 1: Button 3 Pin 2: *Back Pin 2: Unused Pin 3: *Left Pin 3: Button 1 Pin 4: *Right Pin 4: Button 2 Pin 5: Unused Pin 5: Pot X (Horizontal Control) Pin 6: *Fire Pin 6: Unused Pin 7: +5V (50 mA maximum.) Pin 7: +5V (50 mA maximum.) Pin 8: Ground Pin 8: Ground Pin 9: Fire Button 2 Pin 9: Pot Y (Vertical Control) @endnode @node "External Floppy Port Pin-Outs" External Floppy Port (DB23 Female) ----------------------------------------------------------------------------- Pin 1: *Disk Ready Pin 2: *Disk Read Data Pin 3: Ground Pin 4: Ground Pin 5: Ground Pin 6: Ground Pin 7: Ground Pin 8: *Disk Motor Control Pin 9: *Select Drive 3 Pin 10: *Disk Reset Pin 11: *Disk Change (Latched Low) Pin 12: +5V (250 mA maximum.) Pin 13: *Select Disk Side (0=Upper, 1=Lower) Pin 14: *Write Protect Pin 15: *Track Zero Pin 16: *Disk Write Enable Pin 17: *Disk Write Data Pin 18: *Step (Pulse: Low, then high.) Pin 19: Direction (0=Inner, 1=Outer) Pin 20: Unused Pin 21: *Select Drive 2 Pin 22: *Disk Index Pulse Pin 23: +12V (160 mA maximum, 540 mA surge.) @endnode @node "Internal Floppy Connector Pin-Outs" Internal Floppy Connector (34-Pin Male Header) ----------------------------------------------------------------------------- Pin 1: Ground Pin 18: Direction Pin 2: *Change Pin 19: Ground Pin 3: Unused Pin 20: *Step Pin 4: *In Use 1 Pin 21: Ground Pin 5: Ground Pin 22: *DKWD Pin 6: *In Use 0 Pin 23: Ground Pin 7: Ground Pin 24: DKWE (Write Enable?) Pin 8: *Index Pin 25: Ground Pin 9: Ground Pin 26: *TRKD Pin 10: *Select 0 Pin 27: Ground Pin 11: Ground Pin 28: *Write Protect Pin 12: *Select 1 Pin 29: Ground Pin 13: Ground Pin 30: *DKRD Pin 14: Unused Pin 31: Ground Pin 15: Ground Pin 32: *Side Pin 16: *MTRI Pin 33: Ground Pin 17: Ground Pin 34: *Ready Note: some 3.5-inch drives may use different pins for the *Change (diskchange) signal. @endnode @node "Internal IDE Hard Disk Connector Pin-Outs" Internal IDE Hard Disk Connector (40-Pin Male Header) ----------------------------------------------------------------------------- Pin 1: *Reset Pin 21: Unused Pin 2: Ground Pin 22: Ground Pin 3: Drive Data 7 Pin 23: *I/O Write Pin 4: Drive Data 8 Pin 24: Ground Pin 5: Drive Data 6 Pin 25: *I/O Read Pin 6: Drive Data 9 Pin 26: Ground Pin 7: Drive Data 5 Pin 27: I/O Channel Ready Pin 8: Drive Data 10 Pin 28: Unused Pin 9: Drive Data 4 Pin 29: Unused Pin 10: Drive Data 11 Pin 30: Ground Pin 11: Drive Data 3 Pin 31: Interrupt Request Pin 12: Drive Data 12 Pin 32: Unused Pin 13: Drive Data 2 Pin 33: Disk Address 1 Pin 14: Drive Data 13 Pin 34: Unused Pin 15: Drive Data 1 Pin 35: Disk Address 0 Pin 16: Drive Data 14 Pin 36: Disk Address 2 Pin 17: Drive Data 0 Pin 37: *IDE_CS1 Pin 18: Drive Data 15 Pin 38: *IDE_CS2 Pin 19: Ground Pin 39: *Active (LED driver output.) Pin 20: Unused Pin 40: Ground For completeness, the extra four pins used on the mini 44-pin 2.5-inch IDE connector (as found on the A600 and A1200): Pin 41: +5 volts Pin 42: +5 volts Pin 43: Ground Pin 44: Unused @endnode @node "VGA Monitor Pin-Outs" VGA Monitor Connector (HDD15 Male) ----------------------------------------------------------------------------- Modern VGA or multisync monitors use a high-density 15-pin D-connector. This HDD15 connector is the same overall size as a DB9, and in fact early VGA monitors used DB9 connectors. After a while, the connector standardized on the HDD15, probably because many people were damaging their new VGA monitors by connecting them to MDA (mono) video cards. On some cables, pin 9 is not only not connected, but not even present in the connector. Male HDD15 VGA Connector _____________________________ | | | 1 2 3 4 5 | | | | 6 7 8 9 10 | | | | 11 12 13 14 15 | |_______________________| Female HDD15 VGA Connector _____________________________ | | | 5 4 3 2 1 | | | | 10 9 8 7 6 | | | | 15 14 13 12 11 | |_______________________| Pin 1: Red Video Pin 2: Green Video Pin 3: Blue Video Pin 4: Ground Pin 5: Unused Pin 6: Red Ground Pin 7: Green Ground Pin 8: Blue Ground Pin 9: Unused Pin 10: Ground Pin 11: Ground Pin 12: Unused Pin 13: Vertical Sync Pin 14: Horizontal Sync Pin 15: Unused Older Female DB9 VGA Connector _____________________________ | | | | | 5 4 3 2 1 | | | | 9 8 7 6 | | | |_______________________| Pin 1: Red Video Pin 2: Green Video Pin 3: Blue Video Pin 4: Vertical Sync Pin 5: Horizontal Sync Pin 6: Red Ground Pin 7: Green Ground Pin 8: Blue Ground Pin 9: Unused @endnode @node "Power Supply Pin-Outs" Power Supply Connectors ----------------------------------------------------------------------------- Motherboard Power Connector +------+------+------+ |Orange| Red |Brown | Information still needed: how does the Power Good | +12V | -12V |PwrGd | signal work? |------+------+------| |Yellow| Blue | Blue | | +5V |Ground|Ground| +------+------+------+ Disk Drive Power Connectors Large Connector __________ / \\ | O O O O | Connector viewed from front (open end). +---+--------+ The two center pins are grounds. +12V +5V Small Connector | | +-------+ Connector viewed from front (open end). |o o o o| The two center pins are grounds. -----+- +12V +5V @endnode @node "Power-Up Self-Test" Power-Up Self-Test ----------------------------------------------------------------------------- Test Status Color Shown Description Of Error Passed Light Gray Initial hardware configuration tests passed. Initial software tests passed. Final initialization test passed. Failed Red ROM Error: Make sure ROMs are seated properly. Green Chip RAM Error: Make sure Agnus is seated, and check Chip RAM SIMMs for proper seating. Blue This is not an "official" error color. It may have existed in very old (pre-1.0) versions of the system software, but doesn't any longer. Blue screens may be produced by a system failure, but are not really error codes. Yellow Processor detected error before software trapped it. Purple Not an "official" color, but may be caused by bad ROMs. @endnode @node "Keyboard Self-Test" Keyboard Self-Test ----------------------------------------------------------------------------- Number Of Caps Lock Blinks Description Of Error One Keyboard ROM failed. Two Keyboard RAM failed. Three Watchdog timer failed. Four Short detected in keyboard. @endnode @node "Definitive Buster" Definitive Buster by Dave Haynie ----------------------------------------------------------------------------- Editor's Note The following information is a text on interactions between the various versions of Buster chip, the Zorro III bus, and the A3640 processor board, kindly provided by the author himself. ----------------------------------------------------------------------------- System Buster RAMSEY DMAC CPU A3000/16 -07 -04 -01/02 68030-16/68881-16 A3000/25 -06/07 -04 -01/02 68030-25/68882-25 A3000T/030 -07 -04 -02 68030-25/68882-25 A3000T/040 -07 -04 -02 A3640 3.0/3.1 (some have '030 too) A3000+ -09 -07 -04 68030 (68040 planned) A4000/030 -09/11 -07 N/A 68EC030-25 A4000/040 -09/11 -07 N/A A3640 3.0/3.1 A4000T -11 -07 N/A A3640 3.2 Nyx (AAA proto) -11 -07 N/A Any CPU module The A3640 The A3640 had two problems in its Rev 3.0 form. The first was a marginality -- its sampling of the local bus STERM* signal was marginal. This is fixed on Rev 3.1 with a cut and jumper. But beware, some boards marked 3.1 didn't get this fix, though apparently they're a small number. The second problem on the A3640 Rev 3.0 is a real live bug. This was a flaw in the bus arbiter that could allow the '040 and any local bus master on the local bus at the same time. Rev 3.1 incorporates -02 of the U209 PAL to fix this problem. It's not a perfect solution, though, in that it creates a potential for the local bus master to be locked out of the local bus for 10's of microseconds, even if the '040 isn't using the bus. This was corrected in -03 of the U209 PAL, which makes your Rev 3.1 A3640 into a Rev 3.2 A3640. Clearly, if you're not using cards with a DMA problem, this is not an issue. The technical detail on this is that, originally, the A3640 didn't handle a state of the 68040 bus arbitration scheme called "implied mastership". Most of the time the '040 wants the bus, it will assert either BR* (bus request) and/or BB* (bus busy); the former requests the bus, the latter holds it on the bus until it's ready to get off. However, the '040 can still claim the cycle after it negates both BB* and BR*. This is called implied mastership. The idea is that the '040 arbiter figures the current cycle will probably hit in cache, and decides to let another '040-like device on the bus one clock sooner than it might have. Other '040s understand this, and (when their arbiters are properly designed, at least) they can start taking the bus, but stop if the relinquishing '040 really isn't giving the bus back. The Rev 3.0 A3640 didn't handle this condition at all. So the implied mastership condition, which is fairly rare, would cause the bus arbiter to give the cycle over to a pending local bus request. The Rev 3.1 version of the bus arbiter prevented this, by holding the bus in this case. The problem with that is that when it happened, the '040 would usually hit in cache, but the bus would be locked against any other DMA device until the '040 needed the bus. A big waste, though fortunately rare. This is why the GVP PhonePak fails with Rev 3.1; it requires a fairly fine grained determinism when recording from the phone, and the Rev 3.1 card, when it locked up, could be off long enough to overrun any buffering it had available on-card. I was called in to fix this, and the Rev 3.2 board is the result; it handles implied mastership properly. Buster Next we consider the Buster chip. The Buster in the A3000, Rev 6 or Rev 7, is a well proven design. The difference between the two is only that there was a small bug in Rev 6 that caused it to fail at 16MHz, but it works fine at 25MHz. These are what we called Level I Busters; they don't support Zorro III DMA or Quick Interrupts, and they don't attempt to translate local bus burst cycles into Zorro III burst cycles. Starting with the unreleased Rev 8 Buster, we went to Level II, which is roughly twice the size of the Level I design. Level II Buster supports Zorro III bus arbitration, DMA, Quick Interrupts, and translation of local bus burst cycles into Zorro III "Multiple Transfer" cycles. There are two of these parts released: Rev 9 and Rev 11. The Rev 9 Buster has a few flaws. The primary flaw, and the main reason the part was revised, is that the Zorro III bus arbiter can jam under the right conditions. Some DMA cards, like FastLane Z3, use a workaround for this (they avoid the lockup condition), others don't, and will lock up when used with a Rev 9 Buster. There is also a potential problem with end-of-cycle synchronization in the Rev 9 part. Some Zorro III cards will demonstrate this problem, some won't. This is made worse by the STERM* sampling problem on the Rev 3.0 A3640. A final problem with Rev 9 Buster was introduced by the A4000 architecture. The integrated bus buffer, Bridgette, used in the A4000 can't quite guarantee the propagation times required by the Rev 9 Buster design (done before Bridgette was proposed). In the typical case it works fine, in the worst case some Zorro III cards will have a problem with this condition. The Rev 9 part was the unfortunate victim of the wheels of "progress." The first problem was a changeover at CSG (Commodore Semiconductor Group) from channeled arrays to sea of gates. They had a number of screwups on these first parts (the Rev 5 or 6 RAMSEY was also affected), first some mask problems, then speed problems. About six months after I released the Buster design, I got back parts that ran at about 1/4 normal speed; during the A3000 project, we got back parts in more like one month. These problems were eventually fixed, just in time to suffer the change in engineering administration. I had a test bed project to prove all of the features of the Buster Level II chip, a multiprocessor board called Gemini, with two '030s, 4MB of RAM and an independent Zorro III hookup each. I had a prototype of this around the time of the slow Rev 9 Busters, but when the new administration took over, they wouldn't hear of any "Research Projects." Or projects, for that matter; they also tabled the AA project for 6 months, and killed the A3000+. But that's another story... Anyway, after the Rev 9 problems were discovered, I got to fix them, with the Rev 11 Buster. The Zorro III bus arbiter is fixed. All synchronization problems were fixed, and Rev 11 uses a double-strength driver on its STERM* line. Because of this, Rev 11 sometimes cures non-DMA Zorro III problems seen with the Rev 3.0 A3640 card -- that card's flawed STERM* sampling is right on the hairy edge, and the stronger driver makes Buster's STERM* fast enough, at least potentially, to avoid this problem. I still recommend upgrading to Rev 3.1, though, since it fixes the DMA problem, and the STERM* sampling may still be a problem in worst-case, or when RAMSEY or Gary drive STERM*. The bus buffer controls on Rev 11 Buster have been adjusted to support the A4000's buffering scheme perfectly; no properly designed Zorro III cards will have a data setup problem with Rev 11. This is especially critical for burst write cycles. Since it was the last chip of the A3000 architecture that we could revise, I figured a way to solve another A3000 problem in the Rev 11 Buster. There's a race condition between the end of a Zorro II DMA cycle, Gary, and the Amiga chips. When lost, you have problems with Zorro II devices reading Chip RAM during DMA. This was solved with external logic on the A4000 series, and in Rev 11 I figured a way that Buster could play essentially the same trick on Gary. So Rev 11 Busters are a fix for Zorro II DMA problems on an A3000, but aren't needed for that alone on the A4000. Still Having Problems So maybe you're still having problems with Zorro III cards on the A4000, even with Rev 11 Buster and Rev 3.1 or 3.2 A3640, eh? I can think of a few things, though most would lie with the card design. The Rev 11 part runs a somewhat faster Zorro III cycle than Rev 7 did. This isn't a problem if the card was designed to the Zorro III spec; the A3000 architecture only allowed Zorro III to go at about 1/2 its potential rate. It might be a problem for any card designed more to "observed behavior," as defined by how an A3000 first behaved when released. Some cards may have a problem supporting burst cycles on Zorro III, since they couldn't be tested with the Rev 7 Buster. However, this is rare, since the only stock system from Commodore that could run burst on Zorro III is the A4000/030. That's because the A3000's Buster didn't translate burst cycles from the local bus, and the A3640 card doesn't translate burst cycles to the local bus. Also, most Zorro III cards identify themselves as "essentially I/O." These will get mapped as noncachable by the 68040.library, which means they don't get burst, even if you have an '040 card that bursts. On an '030, data burst is disabled by default (you can set it with a SetCPU-like tool), and no I/O card lives in instruction space, so still, no burst. A final Zorro III problem exists on some cards, including the A4091s from Commodore, though not necessarily DKB (eg, I don't know). Originally, there were a couple of ways for a Zorro III card to terminate a bus cycle. It could give the bus back during its last cycle or after its last cycle. This former mechanism can cause some problems, including bus lockups, when multiple masters are present. So I only recommend the latter mechanism -- the card runs its last cycle, then unregisters the bus. This takes longer, but it's safe. This is only an issue when multiple bus mastering Zorro cards are working together. @endnode @node "Boards" A4000 Boards ----------------------------------------------------------------------------- @{" " link "68030 Processor Board Reference"} 68020/68030 Processor Board Reference @{" " link "A2060 Reference"} A2060 Reference @{" " link "A2065 Reference"} A2065 Reference @{" " link "A2091 Reference"} A2091 Reference @{" " link "A2320 Reference"} A2320 'Amber' Reference @{" " link "A3640 Reference"} A3640 Reference @{" " link "A4091 Reference"} A4091 Reference @{" " link "Ariadne Reference"} Ariadne Reference @{" " link "DKB 3128 Reference"} DKB 3128 Reference @{" " link "Emplant Reference"} Emplant Reference @{" " link "FastLane Reference"} FastLane Reference @{" " link "Hydra Reference"} Hydra Reference @{" " link "MultiFaceCard III Reference"} MultiFaceCard III Reference @{" " link "Oktagon Reference"} Oktagon Reference @{" " link "Retina Z2 Reference"} Retina Z2 Reference @{" " link "Warp Engine Reference"} Warp Engine Reference @endnode @node "68030 Processor Board Reference" 68020/68030 Processor Board Reference ----------------------------------------------------------------------------- This processor board is the board supplied with the A4000/030. It may contain a 68030, 68EC030 (functionally equivalent to the 68030 but without a memory management unit), or even a 68020 processor. The 68020 option was apparently for an extremely low-cost version of the A4000; it is unlikely that any boards using the 68020 were ever sold. A possible cost-reduced variation of this processor board has no jumpers. This type of board has a PLCC socket for the math coprocessor, which runs at the same speed as the processor. (See also Tips/@{"Processor Board Mounting" link "Processor Board Mounting"}.) Jumpers J100: FPU Select 1-2 Closed: Use FPU in the PLCC socket. 2-3 Closed: Use FPU in the PGA socket. J101: FPU Clock 1-2 Closed: Use optional on-board oscillator at U103 for FPU clock. 2-3 Closed: Use CPU clock as FPU clock. J103: MAPROM Enable 1-2 Closed: MAPROM disabled. 2-3 Closed: MAPROM enabled (requires U100). J201: 68020 Select 1-2 Closed: 68020 not selected. 2-3 Closed: 68020 selected. J202: 68030 Select 1-2 Closed: 68030 selected. 2-3 Closed: 68030 not selected. J203: 68020/68030 Select 1-2 Closed: 68030 selected. 2-3 Closed: 68020 selected. @endnode @node "A2060 Reference" A2060 Reference ----------------------------------------------------------------------------- The A2060 is a full-length Zorro-II network card that supports the Arcnet standard. While Ethernet is far more popular, Arcnet has recently become very cheap on the Amiga due to a surplus of these cards. While Arcnet does not transfer information as quickly as Ethernet, tests of actual transfers on the Amiga suggest that it can move information at rates up to 150K bytes per second, which is adequate for many purposes. Arcnet can be configured in a bus arrangement where each machine is linked to the next, or in a star, where all the machines are connected to an active hub. The A2060 will work with both setups. The A2060 has some bugs. First, the "hybrid" chip that forms the electronic interface to the Arcnet network comes in two different versions: HCY 9058 (for bus networks) and HCY 9068 (for star networks). As the A2060 manual describes it, the board is for a bus network, but many A2060s come with the 9068 (star) hybrid installed. A bus network needs 93-ohm terminators at each end, and this works fine with the 9058 (bus) version of the hybrid. With the 9068, however, the hybrid itself performs the termination. To connect two machines with 9068 hybrids, run coax from one machine to the other, without using terminators. Using T-connectors to attach more machines in the middle of the bus may or may not work, due to each one adding its termination to the bus. To connect a 9068-version A2060 to a bus network of 9058-version A2060s, place it at the end of the chain and connect the cable directly, without a terminator (this may limit the network to only being operational when the 9068-equipped machine is on). Both versions of the card should have no problems when attached directly to an active hub. It is also possible to replace the HCY 9068 hybrid with the 9058 version, provided you can locate one. There are also several well-known problems with version 37.2 of the "a2060.device" driver software. Replacements for this driver are available in the comm/net directory of Aminet. Some commercial networking packages like Envoy 2.0 also include much better replacement drivers. Other Notes Arcnet requires RG62 coaxial cable, *not* the RG58 that Ethernet uses, and has a minimum cable length between stations of three feet (0.9 meter). Active hubs used for a star layout are self-terminating, so cables are connected directly between the hub and the Arcnet cards. If the A2060 does not perform reliably even with updated driver software, check the board for cold solder joints on hand-soldered components like the BNC coax connector and DIP switches. Some or all of these components may need to have overly-long leads trimmed to prevent interference with adjacent cards or connectors. Finally, the Arcnet address switches on the back of the board are labelled incorrectly in the manual (or on the board, depending on how you look at it). At least some A2060's have a sticker stuck onto the DIP switch, which may disagree with both other references. Ignore all of these: the correct layout is described in the Switches section below. (Assign Arcnet ID numbers starting with 254 and decreasing from there. This will provide a slight performance increase due to Arcnet's token-passing setup.) Despite all the problems, the A2060 works quite well once the bugs are corrected. Board Layout _________________________________________________________________ | _________ ... |___ | | ROM | LED | | |_________| ______ | | | | | | | | Bit0# | |Hybrid| . # Arcnet | | | . # Node ID | (Hybrid version number is | | . # Switches | labelled on the back side | | Bit7# | of the potted "chip.") |______| |_ | |_| BNC | | |_________________________________________________________________| ||||||||||||||||||||||||||| | Jumpers LED: Access LED. Attach a hard disk access LED here to see activity on the Arcnet bus. The left pin of the connector is positive, and the board provides a current-limiting resistor. Switches Arcnet Node ID: This switch is used to set the Arcnet address of the board (refer to the board diagram above). Bit 0 is the switch farthest from the BNC connector; bit 7 is the closest to the BNC connector. Switch settings: 1: Down (toward the solder side of the board) 0: Up (toward the component side of the board) Note: Zero is reserved, and not a valid Arcnet address. Example Arcnet Node Address Settings ID Binary Bit7 Bit6 Bit5 Bit4 Bit3 Bit2 Bit1 Bit0 ----- -------- ---- ---- ---- ---- ---- ---- ---- ---- 254 11111110 down down down down down down down up 253 11111101 down down down down down down up down 252 11111100 down down down down down down up up 128 10000000 down up up up up up up up 3 00000011 up up up up up up down down 2 00000010 up up up up up up down up 1 00000001 up up up up up up up down @endnode @node "A2065 Reference" A2065 Reference ----------------------------------------------------------------------------- The A2065 is a full-length Zorro-II network card that supports the Ethernet standard. Both BNC (coax) and DB15 (AUI) connectors are provided. While the A2065 hasn't been produced for some time, it's still the most common Ethernet board found in Zorro-bus Amigas. Ethernet Configurations 10Base-2 (coax cable, or "thin net") and 10Base-T (twisted pair) are now the most common variations of Ethernet. Thin Net (10Base-2, or coax) The coaxial cable variant of Ethernet uses a bus arrangement, where a "t-connector" is attached to the BNC connector of the Ethernet board, and RG58 coaxial cable is connected to the two BNC connectors on it. (Note that "stub" extension cables between the t-connector and the card are *not* allowed.) At the ends of the bus, 50-ohm terminating resistors are installed on the unused BNC connectors of the t-connector. Twisted Pair (10Base-T) Twisted-pair Ethernet uses two pairs of Category 3 (or higher) twisted- pair wire, although the wire usually used actually contains four pairs. Eight-conductor RJ45 phone-type connectors are used (the actual pins connected are 1,2,3 and 6). Topologically, 10Base-T uses a star arrangement, where wire is run between each workstation and a "concentrator," or hub. Two cards can be directly connected with a "null- modem" version of the standard cable, in which the pairs are swapped (1-2 connect to 3-6, and 3-6 connect to 1-2). While the A2065 does not directly support 10Base-T, inexpensive transceivers are available to plug into the AUI port. Board Address The firmware Ethernet address of the A2065 can be determined by taking the registered CBM block number (00-80-10) and appending the hex value shown on the sticker on chip U4. Board Layout _________________________________________________________________ | |___ | | | | | |_ | |_| BNC | | | JP7 ABC |_ | : ABC | | | ABC | | AUI | JP1-JP6 | | | :::::: |_| | | |_________________________________________________________________| ||||||||||||||||||||||||||| | Jumpers JP1-JP6: Interrupt (default is JP2). JP7: Ethernet Type Off: Type 1. On: Type 2 (default). ABC Jumpers: Select Thick/Thin Ethernet. AB: Select BNC connector (thin Ethernet). BC: Select AUI connector (thick Ethernet). @endnode @node "A2091 Reference" A2091 Reference ----------------------------------------------------------------------------- The A2091 is a full-length Zorro-II DMA SCSI hard drive controller that was originally introduced with the A2000HD. Because of the A2091's high availability, it is often found in A4000s, even though it performs very slowly in them. The A2091 can't use DMA to transfer data to 32-bit Fast RAM, and if Zorro-II DMA memory is not available, the driver falls back to programmed I/O transfers. In either case, this makes for extremely slow disk transfers (at worst, 50K/second; at best, no more than 1M/second). There are utilities in the "hard" directory of Aminet that may help alleviate this problem. (Side note: with some of these, adding some memory on the A2091 may help by providing Zorro-II DMA-accessible memory for buffering.) ROM revisions are a common problem with the A2091; 6.6 or later ROMs are desirable, with 7.0 being the last version (strongly recommended, and necessary for 68040 machines). Most A2091s have revision 04 of the Western Digital 33C93 SCSI controller chip; replacing this with the 08 version may clear up SCSI bus problems. (Despite common belief, the "PROTO" marking is insignificant on either version of the WD chip; version 04 or version 08 are the only significant values.) Sixteen sockets allow for the addition of up to 2M of 16-bit Fast RAM, using CMOS 256K x 4 DRAMs (44C256) rated at 120 ns or faster. (This is not normally very useful on the A4000, since a SIMM added to the motherboard RAM sockets is simpler to obtain and install, and will operate *much* more quickly.) The hard drive power connector on the board is not a reliable way to power a card-mounted hard drive. Use the connectors attached directly to the power supply instead. Usually looked on as a cheap (sometimes very cheap) way to access SCSI peripherals from the A4000, the A2091 is certainly better than no SCSI controller at all. A little-known and mostly-unused function of the A2091 is a built-in XT-IDE controller. The IDE connector and LED are not installed, but visible on the board to the left and upper-right of the SCSI chip. Unfortunately, this port will only work with 8-bit (XT) IDE hard drives, not the common 16-bit AT-IDE hard drives. This option was apparently mostly used for the 20M drives supplied with the A590, the A500 hard disk option which shared the 2091 design. Board Layout _________________________________________________________________ | DRAM _________ ::Memory ... |___ | |||| | Odd ROM | ::Size LED | | |||| |_________| :: | | _________ :: :: |# | |||| |Even ROM | :: :: 50-Pin SCSI |# External | |||| |_________| Options:: Connector |# SCSI | :: |# Port | |||| :: |# (DB25) | |||| :: |# | :: | | |||| . Power | | |||| : Connector | |_________________________________________________________________| ||||||||||||||||||||||||||| | Jumpers JP1: Memory Size 0K: Set for no memory. 512K: Set for 512K of RAM. 1M: Set for 1M of RAM. 2M: Set for 2M of RAM. JP2: Autoboot Enable AUTO: Set to enable autobooting from the A2091. DIS: Set to disable autobooting from the A2091. JP3: Interrupt Select INT6: Select interrupt 6. INT2: Select interrupt 2. JP5: Options Option 1: LUN Enable. When jumpered, enables scanning for Logical Unit Numbers numbers at each physical SCSI address. Default value: Off. Option 2: Time-Out Length. When jumpered, enables longer time-out for slow-starting drives. Default: Off. (Note: When on, this jumper disables parity during the message in phase.) Option 3: Reserved. Factory default: Off. JP201: Factory use only. For pin-out information on the internal 50-pin header and external DB25 SCSI connectors, see Drives/@{"SCSI Pin-Outs" link "SCSI Pin-Outs"}. @endnode @node "A2320 Reference" A2320 'Amber' Reference ----------------------------------------------------------------------------- The A2320 is a video deinterlacer board originally built for the A2000. It is essentially the motherboard deinterlacer circuitry from the A3000 on a board. Based on the Amber chip used in the A3000, the board is often referred to as the Amber board. Physically, the board is designed to fit into the video slot of an A2000. Electronically, it works fine in an A4000. Why would you need a separate deinterlacer board when the A4000 already has AGA circuitry that can scan-double? If you have a VGA or multisync monitor, there are two main reasons: A. Not all programs can be mode-promoted to "double" screens through software (games, for instance). The Amber board will scan-double all 15.75 kHz screens. B. The AGA "double" modes are not truly double in frequency. A 640x200 "doubled" screen syncs at about 27.5 kHz, not the 31.5 kHz that you'd expect. Some multisync monitors can't sync this low. With an Amber board, the output is 31.5 kHz, the same as "stock" VGA. Physical Mounting A modified "slot cover" can be attached to the back panel of the Amber board to allow it to be attached securely to an A4000 slot. You'll also need to trim a bit off the "top" of the Amber's metal panel to allow clearance for the A4000 case (a nibbling tool is useful here). The board will only fill part of A4000 video slot; it looks funny this way, but it works. Don't remove the enable/disable switch! The Amber gets confused by some of the "doubled" screen modes, and rather than passing them through, tries to double them to 55 kHz or above! On these modes, you'll need the disable switch to force the board to pass the video through. (Productivity mode is passed through correctly, since it was part of the ECS chip set that was around when the Amber board first came out.) Disadvantages The Amber board was designed before AGA came out, and doesn't really under- stand AGA. As noted above, some modes are not passed through properly unless the board is disabled with the switch. According to Scott Hood, the designer of the A2320, it samples 12 bits for each color. On the A4000, this is the upper 12 bits of the 24-bit AGA information. So AGA screens with more than 32 colors or HAM-6 will have the colors quantized to a certain degree. This hasn't been a problem so far, although it can be seen on things like ImageFX preview screens. Games that use the AGA color abilities but don't allow for promoting their screens to doubled modes are the only likely sources for this trouble. @endnode @node "A3640 Reference" A3640 Reference ----------------------------------------------------------------------------- The A3640 is the stock 68040 processor board that comes with most A4000s. It contains a 25 MHz 68040; some boards came with the 68LC040, which is a 68040 with no built-in math coprocessor functions. The A4000 User's Guide has instructions on upgrading from a 68EC040, which has no math coprocessor or memory management unit (if any A4000s were ever shipped with 68EC040 processors, there were very few of them). Revision 3.1 or 3.2 boards with U209 marked as "-02" or "-03" can be used in A3000 or A3000 tower computers. (For more information on A3640 board revisions and bugs, see Internals/@{"Definitive Buster" link "Definitive Buster"}.) The "cut and jumper" patch mentioned by Dave Haynie in the Definitive Buster section is: lift pin 6 of U200, and short pads 6 and 7 on U200. This, in combination with an upgrade to the U209 PAL, converts a 3.0 board to 3.1. The only difference between 3.1 and 3.2 boards is another upgrade to the same PAL. Information on PAL upgrades can be found at: ftp://iaehv.nl/pub/users/paul/amiga/A4KPALS.LHA (See also Tips/@{"Processor Board Mounting" link "Processor Board Mounting"}.) Jumpers J100: Enable *CacheDisable *MMU Disable 1-2 Closed and 3-4 Closed: Enable CDIS* MDIS* (caches and MMU disabled at powerup and reset: default). J400: Enable MAPROM: Enable remapping circuit for loading Kickstart into Fast RAM with a developer utility program. 1-2 Closed: MAPROM enabled (default). 3-4 Closed: MAPROM disabled. @endnode @node "A4091 Reference" A4091 Reference by Bob Emery ----------------------------------------------------------------------------- The A4091 is a full-length Zorro-III DMA Fast SCSI-2 hard drive controller designed for the A4000. It is based on the NCR 53C710 chip. (Editor's note: this is the same Fast SCSI controller chip used in the Warp Engine.) Unlike some similar products, it is only a Fast SCSI-2 controller; it has no sockets for additional RAM. The A4091 package includes an active terminator and a custom ribbon cable, which is nicely folded to accommodate up to five internal drives, one on-board and a pair in each of the front and rear bays. The A4091 requires a revision 11 Buster, which Commodore supplied loose with the card. It works with A3640 3.1 and 3.2 revisions, but not 3.0. The board is theoretically capable of 10 Megabytes per second transfer rates. Board Layout _________________________________________________________________ | ... ... LED ::::::::::::::: |___ | 50-Pin SCSI | | ||| ||| ||| Connector | | ||| ||| ||| ||| |# | ||| |||||| |# External | ||| ||| ||| ||| |||||| |# HiDensity | ||| ||| ||| ||| |||||| |# SCSI-2 | |||||| |# Connector | ||| ||| |# | ||| ||| ||| |=sw 8 | ||| . Power |= : | : Connector |= : |_________________________________________________________________|=sw 1 |||||||||||||||||||||||||||||| | LED Connectors To DSPLY To MB CN351 CN308 ------------------------- | o o o o o o | | N R B B R N | C e l l e C | d a a d | c c | k k | Unplug the Front Panel LED cable from the connector on the motherboard and plug it into the CN351. To make the LED operate for IDE drives as well, use the extra cable supplied with the A4091 to connect CN308 to the motherboard. Rear Connector View --------------------------------------------------------------- | | | DIP Switch External SCSI-2 Connector | | _________________ | | ON | | 1 25 | | | 1 2 3 4 5 6 7 8 | {}:::::::::::::::::::::::::{} | | OFF |_________________| 26 50 | | | --------------------------------------------------------------- DIP Switch Settings, defaults are all OFF SCSI Address of Card 0 1 2 3 4 5 6 7 SW1 ON OFF ON OFF ON OFF ON OFF SW2 ON ON OFF OFF ON ON OFF OFF SW3 ON ON ON ON OFF OFF OFF OFF SCSI Fast Bus SW4 OFF indicates that the SCSI Fast Bus feature is enabled. Set this switch to ON if none of your SCSI devices support SCSI Fast Bus. Short/Long Spinup SW5 OFF indicates that your system uses the standard spinup (booting) time. Set this switch to ON to request a longer booting period. If one of your SCSI devices has a long power-on cycle, the Amiga may not recognise it during the standard booting period. Synchronous Mode SW6 OFF indicates that the synchronous mode feature is enabled. Set this switch to ON to disable synchronous mode. Synchronous mode does not require acknowledgement for every byte transmitted, which can mean improved response time with most SCSI devices. External SCSI Termination SW7 OFF indicates that you do not have any external devices. This activates the terminator on the board since this is one end of the SCSI bus. Set this switch to ON when you install an external device. This disables the termination on the board since it is now in the middle of the SCSI bus (not at the end). Logical Unit (LUN) Enable SW8 OFF indicates that unit 0 is the only unit recognized. Set this switch to on to enable the system to recognize 1-6 as LUNs. @endnode @node "Ariadne Reference" Ariadne Reference ----------------------------------------------------------------------------- The Ariadne is a combination Ethernet and parallel port board. Physically, it is a 3/4 length card, with BNC and RJ45 connectors for 10BASE-2 and 10BASE-T Ethernet connections, and a female DB25 parallel port connector. An internal 26-pin header provides a connection for a second parallel port, and four LEDs display connection information. By default, the Ariadne auto-selects the type of Ethernet media based on what is attached. This can be overridden by setting an environment variable called Sana2/ariadne0.config to the preferred value (10BASET or 10BASE2). For example, to force the Ariadne to use the 10BASE-T connection: setenv Sana2/ariadne0.config 10BASET Remember that environment variables will only be saved if they are copied to the ENVARC: logical device. The Ariadne software claims to support up to ten Ariadne boards in one Amiga (although which Amiga has this many slots remains unclear). Software is also provided to redirect printing to the Ariadne parallel ports. Board Layout _____________________________________________________________ | Parallel Port 2 1234 1 _____ |___ | ############ OOOO ::: | | | | JP1 .. 1 LEDs LED |EPROM| |# | Boot Connector| | |# DB25 | | | |# Parallel | |_____| |# Port 1 | |# | | | #| RJ45 | #| 10BASE-T | |_ | |_| BNC |_____________________________________________________________| 10BASE-2 ||||||||||||||||||||||||||| | Jumpers JP1: Enable Boot ROM? Off: Set to disable autobooting (default). On: Set to enable autobooting from an on-board EPROM. Connectors LED Connector (6-pin header) Pin 1: (LED 1) Twisted pair MAU link status Pin 2: (LED 2) Transmit status Pin 3: +5V Pin 4: (LED 3) Collision Pin 5: +5V Pin 6: (LED 4) Receive status Parallel Port 2 (26-pin header) Note that these connections are set up for use of a standard 26-pin header to a DB25 female mounted on a slot cover. This should be a standard part, but check connections before using it. Pin 1: *Strobe Pin 2: +5V Pull Up (Current limit unknown on Ariadne, normally 10 mA.) Pin 3: Data 0 Pin 4: Unused Pin 5: Data 1 Pin 6: *Reset Pin 7: Data 2 Pin 8: Ground (Do not connect any of these grounds to a shield.) Pin 9: Data 3 Pin 10: Ground Pin 11: Data 4 Pin 12: Ground Pin 13: Data 5 Pin 14: Ground Pin 15: Data 6 Pin 16: Ground Pin 17: Data 7 Pin 18: Ground Pin 19: *Acknowledge Pin 20: Ground Pin 21: Busy Pin 22: Ground Pin 23: Paper Out Pin 24: Ground Pin 25: Select Pin 26: Unused Parallel Port 1 (DB25 Female) Parallel Port 2 (DB25 Female when connected) Both Ariadne parallel ports have the same pin-out as the standard A4000 parallel port. See Connector Pin-Outs/@{"Parallel Port Pin-Outs" link "Parallel Port Pin-Outs"}. @endnode @node "DKB 3128 Reference" DKB 3128 Reference by Phil Wright ----------------------------------------------------------------------------- The DKB 3128 is a full-length Zorro-III memory expansion board for the A3000(T) and A4000(T) Amiga computers. The DKB 3128 accepts up to 4 32-bit SIMMs of 4, 8, 16 or 32 megabytes in size for a total of 128MB of add-on memory. The SIMMs you choose should be 80 ns or faster. The SIMM sockets hold the installed SIMMs parallel with the DKB 3128 circuit board so the DKB 3128 remains low profile even with SIMMs installed. If you install SIMMs larger than 4MB you will need to remove one or more jumpers from the SIMM size jumpers. If you install an 8 or 32 megabyte SIMM module, you must remove the shorting block on J1. If you install a 16 or 32 megabyte SIMM module, you must remove the shorting block on J2. Jumper J3 is not used and should not have a shorting block on it. Note: The DKB 3128 circuit board is silk-screened with this information. You can mix SIMM sizes on the DKB 3128 but you must put the largest SIMM in the second SIMM socket labeled "BANK 1" below. Additionally, if you mix SIMM sizes, you must install the included 3128 program at the beginning of your startup-sequence. If you do not, the system will assume every module is the same size as the largest and add "phantom" memory which isn't actually there. Running your system in this state is not a good idea. The disk that comes with the DKB 3128 includes the 3128 and an installer script to install it. Memory on the DKB 3128 is not as fast as motherboard memory in the A4000, about 80% of normal A4000 memory speed according to AIBB. Board Layout _________________________________________________________________ | 123 ____ ____ ____ ____ |___ | ::: | | | | | | | | | | SIMM | | | | | | | | | | Size B| | B| | B| | B| | | | A| | A| | A| | A| | | | N| | N| | N| | N| | | | K| | K| | K| | K| | | | | | | | | | | | | | 0| | 1| | 2| | 3| | | | | | | | | | | | | | | | | | | | | | | | _|_| _|_| _|_| _|_| | | | |_________________________________________________________________| ||||||||||||||||||||||||||| | SIMM Size Jumper Settings 4MB 8MB 16MB 32MB J1 L R L R J2 L L R R R = Remove L = Leave Installed J3 should never be installed. @endnode @node "Emplant Reference" Emplant Reference ----------------------------------------------------------------------------- The Emplant is a Zorro-II board that, in combination with the appropriate software, makes it possible to emulate other computers on the Amiga system. The Emplant hardware performs several functions. Sockets are provided for both DIP- and SIMM-packaged ROMs (the ROMs of the "target" computer are installed on the board, then copied to image files, then removed). Also provided are two Macintosh-type serial ports and a basic, no-frills, non- autobooting SCSI-1 interface based on the 53C80 SCSI chip. (There are also empty sockets for audio digitizing chips on the Emplant board, but the software has never been implemented.) Dual serial ports and the SCSI controller are optional, and some models of the board come without one or both of these options. A "RsrvMem" command added to the the beginning of the startup-sequence sets up the MMU for use with the Emplant board. If you wish to remove the Emplant from the A4000, it's best to remove this command first, or getting the machine to start may be impossible. Early versions of the board came with "ST" brand serial chips, "LOGIC" brand SCSI chips, or custom GALs with date codes earlier than 4693. All of these chips caused problems; the serial and SCSI chips with their respective ports, the GAL chips with general operation. Utilities Unlimited has offered free replacements for these problem chips in the past. Contact them at: Utilities Unlimited International, Inc. 790 N. Lake Havasu Avenue #16 Lake Havasu City AZ 86403 (602) 680-9004 Sales (602) 453-6407 Fax (602) 680-9234 Tech Support A commonly-defective -5V regulator on the A4000 motherboard can cause the Emplant board to fail diagnostic tests and not perform properly in other ways, including unreliable operation of AppleTalk devices. See the Common Problems section on @{"-5V Power Problems" link "-5V Power Problems"} for more information. The included diagnostic program is often a source of concern; reports indicate that it doesn't work properly on PAL machines, and often it can't locate the Emplant board, even if there are no problems. This may be a problem specific to the A4000. Board Layout _________________________________________________________________ | JP1 __ _ JP6 _ RCA____ |___ | xxxxxxxxxxxxxxxxxxxxxxxxxxx JP2 | | | | | | | | | | ROM SIMM Socket ------------ | | | | | | | | ##| Serial | | Autoboot | | | | | | | | | ##| Port B | | Socket | | | | | | | | | | | ------------ ---------------- | | | | | | | | ##| Serial | | ROM/RAM | | ROM/RAM | JP3 |__| |__| |__| |____| ##| Port A | | Socket-1 | | Socket-3 | JP4 _________________ | | ------------ ------------ | 53C80 SCSI Chip | |# External | | ROM/RAM | | ROM/RAM | JP5 |_________________| |# SCSI | | Socket-2 | | Socket-4 | |# Port | ------------ ------------ Internal SCSI Connector |# |_______________________________________XXXXXXXXXXXXXXXXXXXX_JMP1_| ||||||||||||||||||||||||||| | Jumpers JP1: ROM SIMM Address Line Selection Right: Default. Left: ???. JP2: ROM SIMM Write Enable Right: Default (gated write select). Left: Signal pulled up to +5V. JP3: Auto-Boot ROM/SRAM Socket Power/Address Select Upper: Supply power to 28-pin DIP. Lower: Supply address line for 32-pin DIP. JP4: ROM/RAM Socket 3/4 ROM/SRAM Socket Power/Address Select Upper: Supply power to 28-pin DIP. Lower: Supply address line for 32-pin DIP. JP5: ROM/RAM Socket 1/2 ROM/SRAM Socket Power/Address Select Upper: Supply power to 28-pin DIP. Lower: Supply address line for 32-pin DIP. JP6: Mac Emulation Audio Mode Select Upper: Mono. Lower: Stereo. JMP1: SCSI Terminator Power Enable On: Supply SCSI terminator power. Off: Do not supply SCSI terminator power. RCA: Input Connector For Audio Digitizing Circuitry @endnode @node "FastLane Reference" FastLane Z3 SCSI-II Controller by Ian A. Clark ----------------------------------------------------------------------------- The Z3 FastLane is a Zorro III Fast SCSI-II DMA controller for the A3000 or A4000 with transfer rates of approximately 7MB/sec Asynchronous and 10MB/sec synchronous operation. It will support devices with SCSI, SCSI-II and Fast SCSI-II interfaces. It also has provision for up to 64MB memory (256MB as an option) in the form of standard 30-pin SIMM modules. RAM speeds supported are 60, 80 or 100ns. It is compatible with Buster revisions -09, -10 or -11. For fitting inside the A4000/040 with version 3.0 68040 processor boards, a 74FCT240 clock driver chip is also provided. This replaces the original 74FCT244 chip at position U103 on the motherboard, directly under the processor daughter-board. For fitting inside the A3000, the Buster revision level MUST be -09 or higher. Software is also provided, on a single floppy, in the form of Workbench and CLI utilities for configuration and control of SCSI devices, a CDROM filesystem and a memory cache utility. (Editor's note: a Usenet post by David A. Newman states that board revision 2.4 is the most current, with 2.3 and 2.2 also available. Revision 2.2 boards under serial number 31001650 may require a patch by Phase 5 Germany to work with a Broadcaster Elite or DBC32. Use with a 68060 may require a new ROM from Phase 5.) See also: @{"MultiFace and FastLane Problems" link "MultiFace and FastLane Problems"} Board Layout _______________________________________________________________ | ------------------- ------------------- |___ | ------------------- ------------------- CONFIG | | Memory bank 4 Memory bank 4 ::::: | |CPS ------------------- ------------------- 43210 |::|# |: ------------------- ------------------- |::|# |:: Memory bank 3 Memory bank 3 |::|# External | ------------------- ------------------- |::|# SCSI | ------------------- ------------------- |::|# port | Memory bank 2 Memory bank 2 |::|# (Centronics) |RSIZ ------------------- ------------------- |::|# |:::: ------------------- ------------------- RAMSPD |::|# |:: Memory bank 1 Memory bank 1 :::::: ^^ |# |____________________________________________________________||_| |||||||||||||||||||||||||| || | |Internal SCSI | Terminator SIPs Jumpers CONFIG: SCSI Device Configuration Jumpers 0 - Debug mode open default 1 - Reserved open default 2 - Slow inquiry mode. open default closed Lengthens time to wait for device response. Used with older drives. 3 - Slow cable mode open default closed Only if transmission problems occur with cables that are longer than 5M. 4 - Syncron Auto-enable open Z3 examines RDB information on SCSI disks to closed determine whether to operate in Synchronous mode. This jumper is closed by default. RAMSPD: RAM Speed And Bank Size Jumpers This jumper block allows speed settings to be set (40, 60, 80 & 100ns). The 40ns setting is currently not supported. The two right-hand jumpers are used for memory configurations, which are detailed below. o o o o o o o o o o o o ^ ^ ^ ^ | | | | | | | 100ns | | | | | 80ns (default) | | | 60ns | 40ns It is possible to use 70 ns SIMMS by setting the jumper to 60 ns. It can be dropped back to 80 ns if problems arise due to the tolerance levels of the modules. The 60 ns setting will provide memory speeds which are approx 95% of that on the motherboard. Memory Configurations Standard 30-pin SIMMs are used, either 36-bit (PC) or 32-bit, so SIMMS generally advertised as either 1Mx9 or 1Mx8 can be used. 4MB and 1MB SIMMS can be mixed (see below) but not in the same bank. Memory must be fitted in groups of four SIMMS to completely fill each bank, and must be the same size in each bank, therefore the minimum configuration for 1 bank would be 4MB (4x1MB SIMMS) and the maximum 16MB (4x4MB SIMMS) An upgrade kit can be provided which will allow the use of 16MB SIMM modules. However, once fitted, you can no longer use 1MB SIMMs, but you can mix 4MB & 16MB. Total --------Memory installed on------- Config Memory Bank 1 Bank 2 Bank 3 Bank 4 Setting 4 4MB 1 8 4MB 4MB 1 12 4MB 4MB 4MB 1 16 4MB 4MB 4MB 4MB 1 16 16MB 2 20 16MB 4MB 2 24 16MB 4MB 4MB 2 28 16MB 4MB 4MB 4MB 2 32 16MB 16MB 3 36 16MB 16MB 4MB 3 40 16MB 16MB 4MB 4MB 3 48 16MB 16MB 16MB 4 64 16MB 16MB 16MB 16MB 5 CPS: Config Page Select jumpers RSIZ: RAM size jumpers These two jumper blocks determine the memory configuration. Each configuration setting has been allocated a number on the chart above and are detailed below. Setting 0: Memory off, or no memory fitted. o o o o o o o | | | | | | o o o o o o o o o o o o o o o o o o o o o | | | | o o o o o o o o o o CPS RAM size RAM speed Setting 1: 4MB, 8MB, 12MB, 16MB exclusively with 1MB SIMMs o o o o o o o | | | | | | o o o o o o o o o o o o o o o o o o o o o | | | | o o o o o o o o o o CPS RAM size RAM speed Setting 2: 16MB, 20MB, 24MB, 28MB with 1x4MB & 1MB combinations o o o o o o o | | | | | | o o o o o o o o o o o o o o o o o o o o o | | | o o o o o o o o o o CPS RAM size RAM speed Setting 3: 32MB, 36MB, 40MB with 2x4MB & 1MB combinations o o o o o o o | | | | | | o o o o o o o o o o o o o o o o o o o o o | | | o o o o o o o o o o CPS RAM size RAM speed Setting 4: 48MB with 3x4MB SIMMs o o o o o o o | | | | | | o o o o o o o o o o o o o o o o o o o o o | | o o o o o o o o o o CPS RAM size RAM speed Setting 5: 64MB exclusively with 4MB SIMMs o o o o o o o | | o o o o o o o | | o o o o o o o o o o o o o o | | | | | | o o o o o o o o o o CPS RAM size RAM speed @endnode @node "Hydra Reference" Hydra Systems AmigaNet 1.1 Ethernet Board by Calum Tsang ----------------------------------------------------------------------------- The Hydra board is a Zorro II compatible full length Ethernet network adapter for all Zorro bus compatible machines. It is widely acknowledged as a superior board to the A2065, having faster throughput. (Editor's note: I've heard this, but have yet to see a benchmark.) Physically, it has an AUI port for transceiver and a pair of BNCs for RG58 ThinNet Coax. This strange configuration has a plus: it means you won't need an extra T connector for each board! Just daisychain the ThinNet cables in and out of each Hydra. This means you won't need an extra T connector for each board; just daisy-chain the ThinNet cables in and out of each Hydra. There is no difference between the two BNCs, just make sure that there is either another station attached to the other BNC, or a terminator if you're on the end of the chain. Onboard, there is only one set of jumpers, block F1, which has six pairs of pins. When all are in the down pair jumpered position, it is set to ThinNet 10Base2 (CheaperNet). When all are jumpered on the up pair, the Hydra is set to use the AUI port for 10Base5. (Meaning you can attach a Thick transceiver, a FOIRL optic box, or a 10BaseT twisted pair adapter.) This shouldn't really be needed-it's already printed on the board and VERY clearly marked. It uses a NatSemi 32490CN Ethernet chip. Software-wise, a SANA2 driver, hydra.device, is provided with the board. Board Layout _________________________________________________________________ | |___ | | | :::::: F1 | | |# 10Base5 | |# /AUI | |# Port | |# (DB15) | |# | |_ | |_| BNC | |_ | |_| BNC |_________________________________________________________________| ||||||||||||||||||||||||||| | Jumpers F1: Up: AUI Down: BNC @endnode @node "MultiFaceCard III Reference" Multifacecard III Reference by Thomas Huber ----------------------------------------------------------------------------- The MultiFaceCard III (MFC) is a Zorro-2 card providing 2 extra serial ports and 1 extra parallel port. The buffered serial ports of the MFC are superior to the A4000's stock serial port, as RTS/CTS-handshake is done by the hardware itself. This makes high-speed transfers more reliable and reduces the CPU load, making the Amiga more useable during the transfers. (Editor's note: the MFC card uses the 68681 serial UART, which has a three-byte FIFO buffer.) The parallel port is capable of higher speeds than the stock A4000 parallel port. As this port is bidirectional, it can be used with a special version of ParNet found on the install disk. There's no ROM or EPROM on the card, and thus no drivers that autoconfig. You have to run the MFC program which provides the "duart.device" for the serial ports and the "pit.device" for the parallel port. See also: @{"MultiFace and FastLane Problems" link "MultiFace and FastLane Problems"} Board Layout |___ | ______________________________| | Ser0 :::::::: | | | | | _______________________________| Ser1 :::::::: |# Serial | |# Port 0 | |# (DB9) | | | :JP |# | |# Parallel | |# (DB25) | |# | |# |______________________________________________________________| ||||||||||||||||||||||||||| | Jumpers JP: Board Disable On: Disable MFC Off: Enable MFC Connectors Ser0: Serial Port 0 (External DB9) (Standard PC 9-pin serial port.) Pin 1: DCD Pin 2: RXD Pin 3: TXD Pin 4: DTR Pin 5: Ground Pin 6: DSR Pin 7: RTS Pin 8: CTS Pin 9: RI Ser0, Ser1 (External DB25 connected to internal 26-pin header) Pin 1: Shield Ground Pin 2: TXD Pin 3: RXD Pin 4: RTS Pin 5: CTS Pin 6: DSR Pin 7: Data Ground Pin 8: DCD Pin 9: +12V (Probably current-limited to about 20 mA.) Pin 10: -12V (Probably current-limited to about 20 mA.) Pin 11: NC? (Stock A4000 serial: Amiga Audio Out (Left)) Pin 12: SI (Stock A4000 serial: Unused) Pin 13: NC? Pin 14: NC? Pin 15: NC? Pin 16: NC? Pin 17: NC? Pin 18: NC? (Stock A4000 serial: Amiga Audio In (Right)) Pin 19: NC? Pin 20: DTR Pin 21: NC? Pin 22: RI Pin 23: NC? Pin 24: NC? Pin 25: NC? The internal Ser0 connector can be used instead of the external DB9 if signals not present on the DB9 connector are needed. Parallel (DB25 compatible with standard A4000 parallel port.) @{"Parallel Port Pin-Outs" link "Parallel Port Pin-Outs"} @endnode @node "Oktagon Reference" Oktagon Reference ----------------------------------------------------------------------------- The Oktagon 2008S is a combination SCSI-2 (but not Fast SCSI-2) and 16-bit RAM board, much like the @{"A2091" link "A2091 Reference"}. However, the Oktagon in combination with the A4000 does not have the extremely slow transfers of the A2091 (see @{"Slow A2091 Problems" link "Slow A2091 Problems"} because the Oktagon does not use DMA; it uses interrupt-driven PIO. The Oktagon uses 1Mx4 ZIP chips, either static column or page mode, to provide 16-bit RAM. An IDE version of the board is also available (2008AT). Oktagon ROM versions of less than 6.5 may have problems with removable media devices. The recommended (current) version is 6.8. To see the current ROM version number, press F1 during powerup, or use C:Version on oktagon.device. Problems have been reported using the Oktagon with 68060 accelerator boards. The Oktagon ROM uses a MOVEP operation, which is an illegal instruction on the 68060. An A4000 with a 68060 board and the Oktagon will refuse to boot because of this problem. Board Layout _________________________________________________________________ | Jumper |___ | : :::::: :: SCSI | |LED 123456 :: Connector | | :: |# | ::::::: 63,83 :: |# External | ::::::: 23,43 :: |# SCSI | ::::::: 62,82 :: |# Port | ::::::: 22,42 :: |# (DB25) | ::::::: 61,81 :: |# | ::::::: 21,41 :: | | ::::::: 60,80 . | | ::::::: 20,40 : Power | |_________________________________________________________________| ||||||||||||||||||||||||||| | Jumpers 1: SCSI Enable/Disable (Open: SCSI enabled, Closed: SCSI disabled) (On the IDE version, this enables or disables the IDE port.) 2: Memory Enable/Disable (Open: Memory enabled, Closed: Memory disabled) 3: MS0 (see below) 4: MS1 (see below) 5: Test Mem (Open: Autoconfig, Closed: Don't autoconfig) 6: Terminator Power (Open: None, Closed: Supply +5V SCSI term power) Memory Size Configuration Memory Size MS0 MS1 Chips Added To Sockets ----------- ----- ------ ---------------------- 2M Open Open 20,21,22,23 4M Closed Open 40,41,42,43 6M Open Closed 60,61,62,63 8M Closed Closed 80,81,82,83 @endnode @node "Retina Z2 Reference" Retina Zorro II Reference by Thomas Huber ----------------------------------------------------------------------------- The Retina ZII is Zorro-II graphics card which enables the A2000, A3000, and A4000 to display screens in 8-, 16-, and 24-bit color modes. RAM The board can have 1M, 2M or 4M of RAM. With 1MB no 24-bit modes are available; with 2MB onboard you still can't reach resolutions of 1990x1426 or 1024x768-24Bit; these resolutions require an upgrade to 4MB. For 1MB or 2MB RAM configurations, 414256 chips are used (1Mbit-ZIP-RAM). For upgrading the card to its maximum size of 4MB, 4Mbit-ZIP-RAMs like 414400's must be used in sockets 1,3,5, and 7. RAM speeds of 80 ns will work, although the manual recommends the use of 70 ns, as not all 80 ns RAMs seem to be fast enough. The detection of the ramsize is done automatically by the software. Troubleshooting If the RAM size is detected wrong, or an alert claims the Retina to be completely without memory, you can remove the jumper JP. This forces the Retina to produce waitstates which might be necessary for the use with some accelerator cards. Board Layout |___ ____________________________________________________________| | | | | | | | || || || || || || || || |# Video | || || || || || || || || |# Port | || || || || || || || || |# (HDD15) | | | 8 7 6 5 4 3 2 1 | | | | JP | | .. | |____________________________________________________________| ||||||||||||||||||||||||||| | Jumpers JP: Wait States On: No wait states (Default) Off: Enable wait states Video Port Pin-Outs (HDD-15) Pin 1: Red Pin 2: Green Pin 3: Blue Pin 4: NC Pin 5: Ground Pin 6: Ground Pin 7: Ground Pin 8: Ground Pin 9: NC Pin 10: Ground Pin 11: NC Pin 12: NC Pin 13: Horizontal Sync Pin 14: Vertical Sync Pin 15: NC @endnode @node "Warp Engine Reference" Warp Engine Reference ----------------------------------------------------------------------------- The Warp Engine is a popular 68040 processor board that replaces the A3640. It includes four 72-pin SIMM sockets and a Fast SCSI-2 host adapter. Memory: Any combination of 4M, 8M, 16M, or 32M 72-pin SIMMs, either 32-bit or 36-bit wide. Add them starting with SIMM4 and working down to SIMM1. It is advised that you put your largest SIMM in the SIMM4 socket. SIMM Speed: For a 28 MHz Warp Engine, 80 ns SIMMs are adequate. A 33 MHz Warp Engine requires 70 ns SIMMs, while a 40 MHz board needs 60 ns. A wait state jumper enables the use of 70 ns SIMMs with 33 and 40 MHz boards (although there is a slight performance reduction). SIMM Types: Single or double-sided SIMMs will work, although the double- sided 16M SIMM is not recommended due to high power consumption. (This probably also applies to double-sided 32M SIMMs; the Warp Engine manual doesn't say so, perhaps because they are rare at present.) Upgrading: All that is required to convert a 28 MHz Warp Engine into a 33 MHz or 40 MHz Warp Engine is to replace the oscillator and processor (although memory SIMMs slower than 60 ns may require jumpering jumper D to enable a wait state). On some variations of the Warp Engine, the 68040 may be soldered in place, making upgrades difficult at best. Memory Setup [courtesy of Steve Kelsey, Warp Engine hardware designer] "There are two jumpers (the second and third pair of pins on JP2) that select the SIMM slot addressing. One of the jumpers controls whether or not any of the installed SIMMs are dual bank (i.e. 8 or 32 MB). The other jumper controls whether or not any of the SIMMs are 16 MB per bank (i.e. 16 or 32 MB). If the jumpers are set correctly and the SIMMs are installed in a reasonable order, most combinations of SIMMs will result in one contiguous block of memory. A few combinations will result in two or three noncontiguous blocks. The Warp Engine has no limitation on how you mix 4, 8, 16 and 32 MB SIMMs. You can put them in any order and you can set the size and dual bank jumpers any way you want. The automatic DRAM sizing routines figure it all out. But some combinations, while completely legal and functional, are not optimal. So, if you configure your Warp Engine properly, you should be able to get the best performance possible. Example: If you have one or more 16 MB SIMMs (and possibly 1 or more 4 MB's), you should set the size jumper to 16/32 (off) and the dual bank jumper to single sided (on). Then, install your 16 MB SIMMs first, followed by any 4 MB SIMMs. This will make all the 16 MB SIMMs and the first 4 MB SIMM (if present) all one contiguous block. There are many other combinations that work in a similar way. There are also some combinations that will result in two or more chunks of memory. The thing to remember is that you should propperly set the jumpers according to the types of SIMMs you have, and then install them in a reasonable order." See also: @{"Processor Board Mounting" link "Processor Board Mounting"} @{"Processor Cooling" link "Processor Cooling"} @{"External SCSI Connector" link "External SCSI Connector"} Board Layout __________________________________________________________________ | ::::::::::: JP2 |__Fast_SCSI-2__| JP1 | | | LKJHGFEDCBA o---- o---- o---- | | | SCSI Terminators | | |================================ _________ | | | SIMM1 | | | 68040 | |================================ | NCR | | | | SIMM2 | 53C710 | | | |================================ | | | | | SIMM3 |_________| | | |================================ |____________________| | +++ +++ SIMM4 | | +++ +++ | |____________ | ^ | :::::::::::::::::::::::::::::::::::::::::::::: | 28-40 MHz | :::::::::::::::::::::::::::::::::::::::::::::: | Oscillator |_____________________________________________________| Jumpers JP1: SCSI Termination Power (Off: termination power not supplied to SCSI bus) (On: termination power supplied to SCSI bus) JP2: A: Mode Select (Off: 68040 enabled, On: 68040 disabled) B: SIMM Type (Off: double-sided, On: single-sided) C: SIMM Bank Size (Off: 16M, On: 4M) D: Wait State (Off: no wait state, On: 1 wait state) E: reserved F: MMU Disable (Off: MMU enabled, On: MMU disabled) G: Cache Disable (Off: caches enabled, On: caches disabled) H: SCSI Config (see below) J: SCSI Config (see below) K: SCSI Config (see below) JP3: reserved JP4: used for A3000 version *only* (connects to pin 21 of U350) SCSI Configuration Jumpers (H, J, K on JP2) K J H (0=Open, 1=Closed) - - - 0 0 0 SCSI autoboot disabled 0 0 1 10-second delay, LUN scan, not synchronous 0 1 0 10-second delay, LUN scan, 200 ns synchronous 0 1 1 10-second delay, LUN scan, 100 ns synchronous 1 0 0 no delay, LUN scan, 200 ns synchronous 1 0 1 no delay, LUN scan, 100 ns synchronous 1 1 0 no delay, no LUN scan, 200 ns synchronous 1 1 1 (default) no delay, no LUN scan, 100 ns synchronous @endnode @node "Drives" Drives ----------------------------------------------------------------------------- @{" " link "Seagate ST3144A Reference"} Seagate ST3096A/ST3120A/ST3144A Reference @{" " link "External SCSI Connector"} External SCSI Connector @{" " link "SCSI Pin-Outs"} SCSI Pin-Outs @{" " link "SCSI Examples"} SCSI Examples @endnode @node "Seagate ST3144A Reference" Seagate ST3096A/ST3120A/ST3144A 120M IDE Hard Drive Reference ----------------------------------------------------------------------------- The Seagate ST3096A (80M), ST3120A (100M), and ST3144A (120M) are the stock drives included with most A4000s. Jumpers | 1 3 5 7 9 | --------------- | | o o o o o | | | | Front Of Drive --> | | o o o o o | | --------------- | 2 4 6 8 10 | ------------------------------------------------------- Single drive: Pins 3-4 jumpered. Two-drive master: Pins 3-4 jumpered and pins 5-6 jumpered. Two-drive slave: Pins 3-4 open and pins 5-6 open. LED connected: Pins 9-10 must be jumpered to connect to an activity LED on the controller. Without this jumper, the drive will work, but there will be no activity light (unless you connect an LED to the connector on the front of the drive itself). @endnode @node "External SCSI Connector" Building An External SCSI Connector ----------------------------------------------------------------------------- Pin connections for external SCSI-2 half-pitch connector: NOTE: Connector is VIEWED |\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/| FROM THE BACK, or inside | 1 49 | of the computer. All odd- | -------------------------- | numbered wires go to the | \\ / | top of the connector, and | ------------------------ | all even-numbered wires go | 2 50 | to the bottom. |/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\| Procedure: obtain the connector from @{"Redmond Cable" link "Redmond Cable"}. Split the 50-pin SCSI ribbon cable connectors back a couple of inches. Start with pin 1, lay it in the crimp-on pins at the upper left of the connector (again, viewed from the back of the connector). By placing the two sides of a small pair of needle-nose pliers on the wire on either side of the crimp pins, you can gently force the wire down into the V-shaped gap between them. Locate wire #2, then use the same procedure to connect it to the other side of the connector. Repeat for all 50 wires, then clip the plastic retaining clips onto the connector and give it a final squeeze to make sure all wires are making contact. You can cut a hole in the A4000's "Expansion" port cover on the back of the machine to hold this new connector (a "nibbling" tool works well for this). Don't forget proper SCSI termination: the devices at both ends of the chain should be terminated. If the internal SCSI cable leads from the controller to an internal drive, and then to this connector, the internal drive must be unterminated for this port to be functional. When used in this configuration, the external device plugged into this port needs to be terminated. Half-pitch terminators are available from @{"Dalco" link "Dalco Electronics"}; using one of these would allow you to operate the A4000 with or without an external SCSI device without having to open up the computer to change termination. @endnode @node "SCSI Pin-Outs" SCSI Pin-Outs ----------------------------------------------------------------------------- Internal SCSI Port (50-Pin Header) External SCSI Connector (High-density 50-pin) See @{"External SCSI Connector" link "External SCSI Connector"}. All odd pins are grounds, except for pin 25, which is unused. Pin 2: *Data 0 Pin 4: *Data 1 Pin 6: *Data 2 Pin 8: *Data 3 Pin 10: *Data 4 Pin 12: *Data 5 Pin 14: *Data 6 Pin 16: *Data 7 Pin 18: *Parity Pin 20: Ground Pin 22: Ground Pin 24: Ground Pin 26: Terminator Power Pin 28: Ground Pin 30: Ground Pin 32: *ATN Pin 34: Ground (Unused?) Pin 36: *BSY Pin 38: *ACK Pin 40: *RST Pin 42: *MSG Pin 44: *SEL Pin 46: *C/D Pin 48: *REQ Pin 50: *I/O External SCSI Connector (DB25 Macintosh-Type Female Pseudo-SCSI) (Note: This is by far the most common style of DB25 SCSI connector. Most [probably all] Amiga SCSI interfaces with DB25 connectors--like the A2091--use this pin-out.) Pin 1: *REQ Pin 2: *MSG Pin 3: *I/O Pin 4: *RST Pin 5: *ACK Pin 6: *BSY Pin 7: Ground Pin 8: *Data 0 Pin 9: Ground Pin 10: *Data 3 Pin 11: *Data 5 Pin 12: *Data 6 Pin 13: *Data 7 Pin 14: Ground Pin 15: *C/D Pin 16: Ground Pin 17: *ATN Pin 18: Ground Pin 19: *SEL Pin 20: *Parity Pin 21: *Data 1 Pin 22: *Data 2 Pin 23: *Data 4 Pin 24: Ground Pin 25: Terminator Power External SCSI Connector (DB25 Future Domain-Type Female Pseudo-SCSI) (Note: This is a rare alternate standard for the DB25 connector. It was only used on the Future Domain TMC-830/845 and TMC-850/860/885, and is included here in case you have a cable meant for these boards.) Pin 1: Ground Pin 2: *Data 1 Pin 3: *Data 3 Pin 4: *Data 5 Pin 5: *Data 7 Pin 6: Ground Pin 7: *SEL Pin 8: Ground Pin 9: Terminator Power Pin 10: *RST Pin 11: *C/D Pin 12: *I/O Pin 13: Ground Pin 14: *Data 0 Pin 15: *Data 2 Pin 16: *Data 4 Pin 17: *Data 6 Pin 18: *Parity Pin 19: Ground Pin 20: *ATN Pin 21: *MSG Pin 22: *ACK Pin 23: *BSY Pin 24: *REQ Pin 25: Ground External SCSI Connector (Centronics 50-Pin Female) Pins 1-12 and 14-25 are grounds. Pin 26: *Data 0 Pin 27: *Data 1 Pin 28: *Data 2 Pin 29: *Data 3 Pin 30: *Data 4 Pin 31: *Data 5 Pin 32: *Data 6 Pin 33: *Data 7 Pin 34: *Parity Pin 35: Ground Pin 36: Ground Pin 37: Ground Pin 38: Terminator Power Pin 39: Ground Pin 40: Ground Pin 41: *ATN Pin 42: Ground (Unused?) Pin 43: *BSY Pin 44: *ACK Pin 45: *RST Pin 46: *MSG Pin 47: *SEL Pin 48: *C/D Pin 49: *REQ Pin 50: *I/O 2.5-Inch Drive SCSI Connector (40-Pin Header) Pins 1, 2, 39, and 40 are +5V. Pins 3, 4, 37, and 38 are ground returns for the +5V supply. Pins 5, 7, 9, 11, 13, 15, 19, 21, 23, 27, 31, and 35 are signal grounds. Pin 17 is used as a connector key and should not be present. Pin 6: *Data 0 Pin 8: *Data 1 Pin 10: *Data 2 Pin 12: *Data 3 Pin 14: *Data 4 Pin 16: *Data 5 Pin 18: *Data 6 Pin 20: *Data 7 Pin 22: *Parity Pin 24: Terminator Power Pin 25: *ATN Pin 26: *BSY Pin 28: *ACK Pin 29: *RST Pin 30: *MSG Pin 32: *SEL Pin 33: *I/O Pin 34: *C/D Pin 36: *REQ @endnode @node "SCSI Examples" SCSI Examples ----------------------------------------------------------------------------- It seems that the SCSI bus is one of the most misunderstood aspects of connecting hard drives and other peripherals to the A4000 (or, for that matter, any other Amiga model). This section of the guide is an attempt to provide some simple examples of proper SCSI device connections. Please note that in the following section, and in the Guide as a whole, I have used the common term "controller" when referring to disk adapter boards, although the more accurate description for both SCSI and IDE would be "host adapter." Definitions Since understanding SCSI requires a background in the jargon, a few basic definitions might be helpful: SCSI This is the original standard, now also known as SCSI-1. The maximum theoretical transfer rate is 5 megabytes per second, although most combinations of drives and controllers do much less, usually less than two megabytes per second. Total length of the SCSI bus cannot exceed six meters. SCSI-2 An extension of the SCSI command set. Most CD-ROM drives that are double- speed or faster are SCSI-2. Note that contrary to popular belief, this doesn't go any faster than good old SCSI-1. Fast SCSI-2 Here's where the speed was increased. Fast SCSI-2 has a maximum transfer rate of 10 megabytes per second, synchronous. Again, this is theoretical, and anything more than a third of that should be considered excellent. Wide SCSI, Differential SCSI, and SCSI-3 SCSI transfers data over an 8-bit wide data path. A variation called Wide SCSI uses a 16-bit wide data path via an additional cable, potentially doubling transfer rates. SCSI-3 is essentially Fast Wide SCSI-2 using only one cable. Another variation is differential SCSI, which uses differential signal cables to provide a total bus length of up to 25 meters. None of these variations will be described in any detail here, since there don't seem to be any Amiga implementations of controllers for them. Adapters are available to connect these devices to normal SCSI controllers, though, so it is possible to connect them to the Amiga. Compatibility SCSI devices are backwards-compatible. That is, you can connect a SCSI-1 or SCSI-2 hard drive to a Fast SCSI-2 controller, or you can connect a Fast SCSI-2 drive to a SCSI-1 or SCSI-2 controller. A Fast controller can't make a SCSI-2 drive go any faster than SCSI-2, but it will work. Termination SCSI bus systems require an impedance-matching terminator circuit at each end of the bus for reliable operation. Many people find termination to be complex, but the subject can be simplified a great deal by remembering one simple rule: the SCSI bus needs to be terminated at both ends, and *only* at the ends. The most common mistake in SCSI termination is assuming that the SCSI controller itself doesn't count; in fact, it does count as a device, and the termination rules apply to it just like other devices. Many Amiga controllers have the termination resistors soldered into place, under the assumption that only internal or only external SCSI devices will be attached. If both internal and external devices are to be used, it is necessary to remove these resistors. SIP sockets may be soldered in their place to provide the greatest versatility, or you can just use external terminators. Terminating resistors are usually SIP resistor packs; most are yellow, blue, or black, and there may be one, two, or three of them. External terminators look like a connector with no cable attached, and can be found in Centronics 50-pin, DB25, and high-density 50 configurations. Some devices, like newer hard drives or external CD-ROM drives, have a single switch or jumper to enable termination. All of the termination schemes described so far are known as "passive" terminators. Electronically, they connect each signal pin to +5V through a 220 ohm resistor, and to ground through a 330 ohm resistor. This voltage divider circuit provides impedance matching for the SCSI bus. The alternative to a passive terminator is an "active" terminator, which connects each of the SCSI signal pins through a 110 ohm resistor to a precision +2.85V regulator (an LT1086CT, for example) which is powered by +5V. Active terminators are superior to passive terminators simply because they are active; unlike the fixed resistors in a passive terminator, the active terminator's voltage regulator will track varying voltages and properly terminate the SCSI bus. Active terminators can cure many problems with unreliable SCSI devices; their only disadvantage is that they cost a bit more (@{"Dalco" link "Dalco Electronics"} sells them for between thirty and forty dollars). Active termination chips are made by Dallas Semiconductor and Texas Instruments. Any combination of passive and active terminators may be used, although two active terminators would be best. In practice, passive/passive or passive/ active are usually adequate. Termination Power Terminator power (+5V) is supposed to be supplied on pin 26 of the 50-pin IDC header. But SCSI devices are not required to supply this power; many have jumpers to enable or disable it. So it is possible to have a proper termination setup, but no power provided to the terminators. As you might expect, this will cause problems. Make sure that at least one device is supplying termination power to the SCSI bus, preferably the controller, since external devices may be turned off, which would deprive the rest of the bus of termination power. Cable Configurations Internal SCSI devices are usually connected with 50-conductor ribbon cable. 50-pin IDC (Insulation Displacement Connector) headers are crimped onto the cable for each device to be attached. "Stub" cables of no more than 3 centimeters off the main cable are allowed by the SCSI standard, but it's better to avoid them altogether by running the cable direct from one device to the next, with no branches off the main bus at all. External SCSI device cables can use several connectors: Centronics 50-pin, DB25, or high-density 50-pin (commonly, but imprecisely, referred to as "SCSI-2", since many Fast SCSI-2 adapters use this type of connector). Adapter cables may have any combination of these three basic types. The SCSI standard states that the total length of the SCSI bus, including internal and external cable, must not exceed six meters. For Fast SCSI-2, the limit was reduced to three meters. In practice, some devices and cable combinations may limit this severely, particularly cables with DB25 connectors (since Apple created the DB25 "pseudo-SCSI" cable by simply discarding all those "extra" grounds that helped make SCSI capable of running long distances in the first place). Conversely, some SCSI bus implementations can go farther than the standard suggests. SCSI Address Numbers Each SCSI device (including the controller) has an address between 0 and 7 assigned to it by the user. These numbers are usually set as a binary number with three jumpers. Controllers often have no jumpers, either requiring software to change their address, or simply not being able to change it at all. Standard Amiga controllers of either type default to a SCSI address of 7. The rules regulating addresses are pretty simple: each device must have a unique address. (There is no physical "order" in which the addesses must occur; you can use any order or combination of numbers, as long as there is only one device with a given address.) Since the Amiga scans the SCSI bus for bootable devices starting with address 0 and proceeding to address 7, it is advised that you assign address 0 to the boot hard drive, and set "HiID" to "On" for this drive in the Rigid Disk Block (RDB). This will prevent the system for looking for other hard drives with a higher boot priority, making for the quickest booting possible, and preventing the system from trying to boot off of a higher-numbered CD-ROM drive. (Check the Aminet disk/misc directory for RDB utility programs.) LUNs Logical Unit Numbers provide a way to access more than one device at a given SCSI address. For example, some Adaptec SCSI-to-MFM adapter boards like the 4000A could control two MFM hard drives. However, the 4000A board used only a single SCSI address; to access each drive, a secondary number-- the LUN--was used: 0 for the first drive and 1 for the second. With modern SCSI devices, LUNs are relatively rare, with the exception of CD-ROM changers. These devices often use an LUN to select which CD is loaded. Specific Troublesome Devices There are a couple of devices out there that are almost guaranteed to be troublemakers on a SCSI bus. Since this section of the Guide is also distributed as the SCSI Examples document, these are included here even though they don't necessarily apply to the A4000. 1. The NEC CDR-36 CD-ROM drive. This is a single-speed (150K/second) external drive with a top-loading case. It may be helpful to disable termination on this drive and use an external terminator; it may not. If odd problems persist, check that pin 17 on the DB25 connector on the cable is not grounded. If it is, disconnect that pin. 2. A3000 problems. The A3000's internal SCSI controller has a few minor flaws that can be problematic. A. The Western Digital 33C93 SCSI chip itself: revision 04 of this chip has some bugs that usually show up when a CD-ROM or tape drive is attached, and revision 08 fixes them. B. The DB25 used as an external SCSI connector on the A3000 can cause problems. Use only short, high-quality SCSI cables attached to this connector, or run 50-pin ribbon cable from the internal connector. C. Termination. Various A3000s seem to have come with no terminators, soldered-in SIPs, or even sockets. Check the motherboard controller termination, and follow the guidelines laid out in the Termination section above. D. Many A3000s had a manufacturing flaw which resulted in terminator power not being supplied at the external SCSI connector. The easiest way to test this is with an external terminator with an LED indicator. Otherwise, you'll need to check pin 25 of the A3000 DB25 SCSI connector for +5V (the shield around the connector provides an easy ground test point). If no voltage is supplied on pin 25, diode D800 (or D801, this may vary depending on motherboard revision) is reversed inside the A3000. Unsolder and replace it (this should be a 1N34 type, although a 1N5817 should work and might be more suitable). The motherboard silk screen is likely to be wrong as well, so ignore it. 3. Some GVP controllers. For a while, it seemed like all the email I received was from owners of GVP controllers. For many of these people, upgrading to the Guru ROM solved their problems. Disabling termination on some of these boards is also non-trivial; Guru ROM author Ralph Babel explains: "Most GVP cards use only two 10-pin SIP terminators _plus_ two extra resistors (SMD, except for the very first revisions of the Series-II hard card) for the parity line for a total of 17 terminated lines (they leave out the RST line)." The use of SMD (Surface Mount Device) resistors complicates disabling termination on these boards. It will be simplest for many users to reorganize the SCSI bus so that the controller is on one end. GVP's current technical support number is 215-633-7711. Example SCSI Bus Setups These examples show connections to the A2091 controller (see Drives/ @{"A2091 Reference" link "A2091 Reference"}), but the connections for other controllers will follow the same standard. In Example 1, the 200M hard drive is used as the boot drive, and the "HiID" flag is set to "On" in this drive's Rigid Disk Block. (The HiID flag may be called by another name, like LastDrive or HighDrive.) For examples 2 and 3, the 540M drive is used as the boot drive, and the HiID flag is set in that drive's RDB. Example 1: 2091 controller, internal 200M SCSI-1 hard drive. Cable connections are 50-conductor ribbon. ________________ ________________ | 2091 | | 200M SCSI-1 | | Terminated |------| Terminated | | Address 7 | | Address 0 | |________________| |________________| Example 2: 2091 controller, internal 200M SCSI-1 hard drive, internal 540M Fast SCSI-2 hard drive. Cable connections are 50-conductor ribbon. The SCSI-1 drive has been renumbered as address 1, and the new Fast SCSI-2 drive is now set at address 0 and used as a boot drive, to provide better performance on the system partitions. (Even though it will only be accessed at SCSI-1 rates, it is a newer drive, and will probably have significantly better transfer rates than the older 200M drive.) Additionally, the newer drive will likely have the more desirable active termination on-board. ________________ ________________ ________________ | 2091 | | 200M SCSI-1 | |540M Fast SCSI-2| | Terminated |------| Not Terminated |------| Terminated | | Address 7 | | Address 1 | | Address 0 | |________________| |________________| |________________| Example 3: 2091 controller, internal 200M SCSI-1 hard drive, internal 540M Fast SCSI-2 hard drive, external SCSI-2 CD-ROM drive. The cable from the CD-ROM drive to the A2091 is a Centronics 50-pin to DB25 adapter cable, and the internal cables are 50-conductor ribbon. An active terminator is attached to the last available external SCSI connector on the CD-ROM drive. Note that the terminating resistors on the A2091 have been removed so that the SCSI bus is terminated only at the ends (the CD-ROM and the 540M drive). ________________ | CD-ROM SCSI-2 | | Terminated |---. External cable connected to A2091 | Address 4 | | external SCSI DB25 connector |________________| | External | | ____________|___ ________________ _______________ | 2091 | | 200M SCSI-1 | |540 Fast SCSI-2| | Not Terminated |------| Not Terminated |------| Terminated | | Address 7 | | Address 1 | | Address 0 | |________________| |________________| |_______________| @endnode @node "Monitors" Monitors ----------------------------------------------------------------------------- @{" " link "1081 Reference"} 1081 Reference @{" " link "1084 Reference"} 1084 Reference @{" " link "1085 Reference"} 1085 Reference @{" " link "1940 Reference"} 1940/1942 Reference @{" " link "1950 Reference"} 1950 Reference @{" " link "1960 Reference"} 1960 Reference @{" " link "Idek Iiyama Vision Master 17 (MF-8617) Reference"} Idek Iiyama Vision Master 17 (MF-8617) Reference @{" " link "Mitsubishi DiamondScan AUM-1381A Reference"} Mitsubishi DiamondScan AUM-1381A Reference @{" " link "NEC 3D Reference"} NEC 3D Reference @endnode @node "1081 Reference" 1081 Reference ----------------------------------------------------------------------------- The 1081 is a 14-inch monitor manufactured by Philips for the European market, and similar to the @{"1084" link "1084 Reference"}, with SCART and composite video connectors. The 1081 may have only been distributed in Europe. Common Problems A standard failure is for the monitor to "pop," then go dark. Hitting it may bring back the picture. This is often caused by cold or cracked solder joints on the flyback transformer, which resoldering should cure. Specifications Sync Frequency: 15.6 kHz Horizontal Dot Pitch: 0.39mm or 0.42 mm Sound Output: 1.0 W RMS/Channel at 5% maximum THD Input Connectors: One permanently attached HDD15 and Audio R/L Pin-Outs (SCART Euroconnector) Pin 1: Unused Pin 2: Audio Input (0.5 Vrms/ > 10Kohms) Pin 3: Unused Pin 4: Audio Ground Pin 5: Blue Ground Pin 6: Audio Input (0.5 Vrms/ > 10Kohms) Pin 7: Blue Video (0.7 Vpp/ 75 ohms) Pin 8: Unused Pin 9: Green Ground Pin 10: Unused Pin 11: Green Video (0.7 Vpp/ 75 ohms) Pin 12: Unused Pin 13: Red Ground Pin 14: Unused Pin 15: Red Video (0.7 Vpp/ 75 ohms) Pin 16: Fast Blanking Pin 17: CVBS Ground Pin 18: Fast Blanking Ground Pin 19: Unused Pin 20: CVBS Input (1 Vpp/ 75 ohms, sync for linear RGB input) Pin 21: Screening Plug Pin-Outs (Digital RGB 8-Pin DIN) Pin 1: Status Computer (?) Pin 2: Red Pin 3: Green Pin 4: Blue Pin 5: Intensity Pin 6: Ground Pin 7: Horizontal Sync or Composite Sync Pin 8: Vertical Sync @endnode @node "1084 Reference" 1084 Reference ----------------------------------------------------------------------------- The 1084 and its variants (1084S, 1084S-P, 1084-P, 1084S-P2, 1084-D, 1084S-D, and 2080) are all 15.75 kHz monitors. They do not handle AGA "double" screenmodes, nor will they display the deinterlaced output from the A2320 Amber board or the motherboard deinterlaced output on an A3000. However, they will show all normal 15.75 kHz displays, and many (most? all?) of the 1084 versions have a separate input for composite video. The 1084 is a usually a variation of the Philips CM8833 monitor; the 1084S-D was made by Daewoo (as was the 1084D, probably). The display tubes used in these monitors were made by Orion, Toshiba, Hitachi, and Samsung. At one time or another, every possible permutation of connectors and video capabilities on the 1084 seems to have been reached, so don't be surprised if your 1084 has some bizarre combination of connectors and specifications. Common Problems * The door covering the front-panel controls is typically broken off. * The attached cables or connectors on some models tended to fail, causing loss of color or other problems. Resolder the pins. You can glue connectors in place to provide support. * Failed or insufficient insulation may cause arcing. * The power switch may partially fail, causing separate parts of the monitor to power down. * A phone caller suggested that internal connectors could tarnish with age, and disconnecting and cleaning them may help clear up some problems. * Repeatedly blown fuses can indicate a bad power supply. * A loud whistling noise indicates a bad flyback transformer (also known as an LOPT). Some 1084 monitors also have digital RGB (PC clone CGA) inputs, and there was such a profusion of minor or major variations that any 1084 might have any combination of analog RGB, digital RGB, and composite inputs. Other Notes * The 2080 is a long-persistance phosphor, .39 mm dot pitch. Specifications Sync Frequency 15.75 kHz (NTSC; 15.6 kHz PAL) Pin-Outs (Analog RGB, 6-Pin DIN) Pin 1: Green Pin 2: Horizontal Sync Pin 3: Ground Pin 4: Red Pin 5: Blue Pin 6: Vertical Sync Pin-Outs (Digital RGBI, 8-Pin DIN) Pin 1: Unused Pin 2: Red Pin 3: Green Pin 4: Blue Pin 5: Intensity Pin 6: Ground Pin 7: Horizontal Sync Pin 8: Vertical Sync Pin-Outs (1084S DB9) Pin 1: Ground Pin 2: Ground Pin 3: Red Pin 4: Green Pin 5: Blue Pin 6: Unused Pin 7: Composite Sync Pin 8: Horizontal Sync Pin 9: Vertical Sync 2080 Pin-Outs (SCART Euroconnector) Pin 1: Unused Pin 2: Audio Input (0.5 Vrms/ > 10Kohms) Pin 3: Unused Pin 4: Audio Ground Pin 5: Blue Ground Pin 6: Audio Input (0.5 Vrms/ > 10Kohms) Pin 7: Blue Video (0.7 Vpp/ 75 ohms) Pin 8: Unused Pin 9: Green Ground Pin 10: Unused Pin 11: Green Video (0.7 Vpp/ 75 ohms) Pin 12: Unused Pin 13: Red Ground Pin 14: Unused Pin 15: Red Video (0.7 Vpp/ 75 ohms) Pin 16: Fast Blanking Pin 17: CVBS Ground Pin 18: Fast Blanking Ground Pin 19: Unused Pin 20: CVBS Input (1 Vpp/ 75 ohms, sync signal for linear RGB input) Pin 21: Screening Plug 2080 Pin-Outs (Digital RGB 8-Pin DIN) Pin 1: Status Computer (?) Pin 2: Red Pin 3: Green Pin 4: Blue Pin 5: Intensity Pin 6: Ground Pin 7: Horizontal Sync or Composite Sync Pin 8: Vertical Sync @endnode @node "1085 Reference" 1085 Reference ----------------------------------------------------------------------------- The 1085(S) is a cost-reduced version of the 1084, with lower resolution (.52 mm dot pitch) and no non-glare screen treatment. Like the 1084, the 1085 is a fixed-frequency 15.75 kHz monitor, and is not compatible with most AGA screen modes. Pin-Outs (DB9) Pin 1: Ground Pin 2: Ground Pin 3: Red Pin 4: Green Pin 5: Blue Pin 6: Unused Pin 7: Composite Sync Pin 8: Unused Pin 9: Unused @endnode @node "1940 Reference" 1940/1942 Reference ----------------------------------------------------------------------------- This monitor is a 13-inch bisync (not true multisync) monitor built by Samsung for Commodore. It has two ranges of sync frequencies to match both normal (15.75 kHz) and doubled screen modes. The 1942 differs from the 1940 only in that it has a smaller dot pitch. MonEd may be useful in getting the picture to fill the entire screen (the 1942 came with a set of patched monitor drivers). Both variations feature stereo speakers. Specifications Sync Frequency: 15.6-15.8 kHz and 27.3-31.5 kHz Horizontal 47 Hz to 75 Hz Vertical Dot Pitch: 0.39 mm (1940) 0.28 mm (1942) Sound Output: 1.0 W RMS/Channel at 5% maximum THD Input Connectors: One permanently attached HDD15 and Audio R/L Pin-Outs (HDD15) Pin 1: Red Pin 2: Green Pin 3: Blue Pin 4: Unused Pin 5: Unused Pin 6: Red Ground Pin 7: Green Ground Pin 8: Blue Ground Pin 9: Unused Pin 10: Digital Ground Pin 11: Digital Ground Pin 12: Unused Pin 13: Horizontal Sync Pin 14: Vertical Sync Pin 15: Unused @endnode @node "1950 Reference" 1950 Reference ----------------------------------------------------------------------------- The 1950 monitor was actually produced by a company called AOC. Parts may still be available direct (although there have been conflicting reports). The AOC model of the monitor was known as the AOC CM314. The tube is made by Hitachi. Common Problems A sync problem may be caused by the monitor detecting sync on the Green input and then disabling the horizontal and vertical sync inputs. Removing the 10K R854 resistor or the 10uf C812 capacitor on the small vertical board may fix this. It may also cause problems of its own, so be warned. The analog/TTL switch appears to be prone to failure. However, if switching the switch brings back the picture, it may actually be that the 74LS123 (IC805) on the same board is failing. The suggested course of action is to replace (or at least resolder) this chip first, since it's a commonly available part. A defect in the way the 74LS123 is mounted may be present; there should be +5V on pin 3, but the way the chip is mounted or the board is manufactured, it may be intermittent. There should be a PCB trace between pin 3 and pin 16, however, pin 3 is not soldered to this trace, but only friction-fit (non-plated-through holes?). When the analog switch is moved, it causes intermittent contact between pin 3 and the trace. A soldered jumper to pin 16 is an easy way to fix this, or you may be able to solder to the trace already present. Another common problem is the failure of a multifunction sync chip. Replacements should be available from Sony. The Toronto ABUG user group confirms that this chip is available from Sony Canada, on Gordon Baker Road (presumably in Toronto, although they don't say). The high-voltage boards may crack; this can be one cause of the monitor that starts working when you hit it. A 1 megohm resistor in the second power supply's startup circuit goes bad, causing the monitor to remain dark. Replacing the resistor with a higher wattage one may help prevent the problem in the future. Manufacturer AOC International 311 Sinclair Frontage Road Milpitas CA 95035 (408) 956-1070 Specifications Sync Frequency: 15 kHz to 35 kHz Horizontal 50 Hz to 80 Hz Vertical Dot Pitch: 0.31 mm Input Connectors: One permanently attached HDD15 (AOC CM314 also has an attached DB9 for digital RGB.) Pin-Outs (HDD15) Pin 1: Red Pin 2: Green Pin 3: Blue Pin 4: Unused Pin 5: Test Pin 6: Ground Pin 7: Ground Pin 8: Ground Pin 9: Unused Pin 10: Ground Pin 11: Ground Pin 12: Unused Pin 13: Horizontal Sync Pin 14: Vertical Sync Pin 15: Unused @endnode @node "1960 Reference" 1960 Reference ----------------------------------------------------------------------------- The 1960 may have been made by Daewoo (Korea) or a Taiwan company. The tube is made by Hitachi. There has been some debate over whether it is a true multisync or a trisync monitor. Reports have been provided that indicate it can handle Super72 screen modes at about 23 kHz, and the manual says it can sync up to 38 kHz. It may be a sort of hybrid, with a wide "window" in the 15.75 kHz to 31.5 kHz range. Common Problems One common problem involves a component that is insulated with eletrical tape (inadequately) from the factory, resulting in arcing. Replacing this insulation can cure the problem. Typical failures also result from cold solder joints on the 1960 boards, which can be repaired by resoldering. Additionally, some solder joints (such as those on the flyback transformer) tend to go bad with age. Resoldering them may cure arcing problems. The screen size adjustment pots may be prone to failure, making adjustments difficult. Specifications Sync Frequency: 15 kHz to 38 kHz Horizontal 50 Hz to 87 Hz Vertical Dot Pitch: 0.29 mm or .31 mm Input Connectors: HDD15 (Analog RGB), DB9 (Digital RGB; some 1960s may not have this connector.) Pin-Outs (HDD15 Analog RGB) Pin 1: Red Pin 2: Green Pin 3: Blue Pin 4: Monitor Sense, Ground to Pin 10 Pin 5: Ground Pin 6: Red Ground Pin 7: Green Ground Pin 8: Blue Ground Pin 9: Unused Pin 10: Digital Ground Pin 11: Jumper to Pin 10 Pin 12: Unused Pin 13: Horizontal Sync Pin 14: Vertical Sync Pin 15: Jumper to Pin 10 Pin-Outs (DB9 Digital RGB) (EGA?) Pin 1: Ground Pin 2: Red Prime Pin 3: Red Video Pin 4: Green Video Pin 5: Blue Video Pin 6: Green Prime Pin 7: Blue Prime Pin 8: Horizontal Sync Pin 9: Vertical Sync @endnode @node "Idek Iiyama Vision Master 17 (MF-8617) Reference" Idek Iiyama Vision Master 17 (MF-8617) Reference ----------------------------------------------------------------------------- This is a fairly popular monitor for use with the Amiga, since it is a high quality, relatively inexpensive 17-inch monitor that can sync down to about 23.5 kHz, and therefore works with most (all?) AGA "double" screenmodes. All presets and controls are digital, set through three front-panel buttons and an LCD display. The image can easily be expanded to fill the screen in all modes. (See Boards/@{"A2320 'Amber' Reference" link "A2320 Reference"} for information on the A2320 'Amber' board that may be used with this monitor.) Specifications Sync Frequency: 23.5 kHz to 86.0 kHz Horizontal 50 Hz to 120 Hz Vertical Resolution: Maximum 1280 x 1024 at 80 Hz Input Connectors: Five BNC connectors and one DB15 (not high density) (A cable is included to connect a HDD15 VGA-type connector to the DB15 connector on the monitor.) Pin-Outs DB15: Pin 1: Red Pin 2: Red Ground Pin 3: Green Pin 4: Green Ground Pin 5: Blue Pin 6: Blue Ground Pin 7: Ground Pin 8: NC Pin 9: NC Pin 10: NC Pin 11: NC Pin 12: NC Pin 13: NC Pin 14: Horizontal or HV Sync Pin 15: Vertical Sync @endnode @node "Mitsubishi DiamondScan AUM-1381A Reference" Mitsubishi DiamondScan AUM-1381A Reference ----------------------------------------------------------------------------- The DiamondScan is one of the few VGA-type multisync monitors that has a composite video input, and that made it relatively common for use on the Amiga (although I believe that Mitsubishi no longer makes this model). The official scan rates cover the range from 15.6 kHz to 36 kHz, so the DiamondScan should work with all normal Amiga video modes. User controls are standard knobs and buttons, and there are no digital memory features, so using it with the Amiga means that you have to juggle the monitor's picture location settings along with the Amiga overscan and screen position settings. The DiamondScan works fine with the Amber board (see Boards/ @{"A2320 'Amber' Reference" link "A2320 Reference"}). One feature of the DiamondScan is particularly applicable to the video production uses of the Amiga: the "Composite/RGB Select" (pin 22) on the DB25 input. Connect this pin through a switch to ground, and then a flip of the switch will select composite video or analog RGB display without reaching for the switches on the back of the monitor. Specifications Sync Frequency: 15.6 kHz to 36 kHz Horizontal 45 Hz to 90 Hz Vertical Resolution: Maximum 800 x 560 (Rated...normally considered to be an 800x600 monitor.) Input Connectors: BNC (Composite Video) DB9 (EGA/CGA/Mono TTL) (DB9-to-DB9 cable was included.) DB25 (Analog RGB) Pin-Outs DB9: (For TTL 16-Color CGA) Pin 1: Ground Pin 2: Unused Pin 3: Red Video Pin 4: Green Video Pin 5: Blue Video Pin 6: Intensity Pin 7: Unused Pin 8: Horizontal Sync Pin 9: Vertical Sync DB9: (For TTL 64-Color EGA) Pin 1: Ground Pin 2: Secondary Red Video Pin 3: Primary Red Video Pin 4: Primary Green Video Pin 5: Primary Blue Video Pin 6: Secondary Green Video/Intensity Pin 7: Secondary Blue Video Pin 8: Horizontal Sync Pin 9: Vertical Sync DB9: (For TTL Mono) Pin 1: Ground Pin 2: Unused Pin 3: Unused Pin 4: Unused Pin 5: Unused Pin 6: High Intensity Pin 7: Video Pin 8: Horizontal Sync Pin 9: Vertical Sync DB25: Pin 1: Sync Ground Pin 2: Red Video Pin 3: Red Video Ground Pin 4: Green Video Pin 5: Green Video Ground Pin 6: Superimpose Control (YS) Pin 7: Superimpose Ground Pin 8: Video Input Select (AV) Pin 9: Composite Video Input Pin 10: Composite Video Ground Pin 11: Composite Video Out Pin 12: Composite Video Ground Pin 13: PGA Mode Control Pin 14: Blue Video Pin 15: Blue Video Ground Pin 16: Horizontal/Composite Sync Pin 17: Vertical Sync Pin 18: Unused Pin 19: Unused Pin 20: Unused Pin 21: INT (+5V ???) Pin 22: Composite/RGB Select (TTL level: Low for RGB, high or open for composite.) Pin 23: Analog/TTL Select (TTL level: Low for TTL, high or open for analog.) Pin 24: Remote (TTL level: Low to disable Mode Switch.) Pin 25: Shield Ground @endnode @node "NEC 3D Reference" NEC 3D Reference ----------------------------------------------------------------------------- The NEC 3D is a popular monitor for use with the Amiga. I've never had one, so the following information is somewhat sparse (submissions are welcome). A common complaint is that it is impossible to adjust the picture width to fully eliminate the large black borders on the left and right of the image. Specifications Sync Frequency: 15.5 kHz to 38 kHz Horizontal 50 Hz to 90 Hz Vertical Dot Pitch: 0.28mm Trio Resolution: 1024 Horizontal 768 Vertical (Interlaced) Input Connectors: DB9 or HDD15 (maybe earlier models had the DB9) Pin-Outs (HDD15 Analog RGB) Pin 1: Red Video Pin 2: Green Video Pin 3: Blue Video Pin 4: Ground Pin 5: Ground Pin 6: Red Ground Pin 7: Green Ground Pin 8: Blue Ground Pin 9: Unused Pin 10: Ground Pin 11: Ground Pin 12: Unused Pin 13: Horizontal Sync Pin 14: Vertical Sync Pin 15: Unused @endnode @node "Sources" Sources For Components ----------------------------------------------------------------------------- Opinions in this section are strictly those of the @{"Editor" link "Editor"}. This list includes sources for suppliers of parts and accessories like cables and connectors. Suggestions: * For general or custom cables and connectors: Dalco or Redmond Cable. * For general board-level parts (not custom Amiga): Digi-Key and JDR. * For custom Amiga parts: Unknown at this point. Suggestions welcome! ----------------------------------------------------------------------------- @{" " link "Altex Electronics"} Altex Electronics @{" " link "Benetech Electronic Supply"} Benetech Electronic Supply @{" " link "Chip Merchant"} Chip Merchant @{" " link "Dalco Electronics"} Dalco Electronics @{" " link "Digi-Key Corporation"} Digi-Key Corporation @{" " link "Hosfelt Electronics"} Hosfelt Electronics @{" " link "JameCo Electronic Components"} JameCo Electronic Components @{" " link "JDR Microdevices"} JDR Microdevices @{" " link "Marlin P. Jones & Associates"} Marlin P. Jones & Associates @{" " link "MCM Electronics"} MCM Electronics @{" " link "Memory World"} Memory World @{" " link "Parts Express"} Parts Express @{" " link "Redmond Cable"} Redmond Cable @endnode @node "Altex Electronics" Altex Electronics 11342 N IH 35 San Antonio TX 78233-9903 (800) 531-5369 (210) 637-3264 Fax I've not had a lot of experience with Altex, although they seem okay, and have a pretty good selection of connectors and components at good prices. @endnode @node "Benetech Electronic Supply" Benetech Electronic Supply Route 1, Box 247 Canton TX 75103 (800) 733-4705 (903) 848-0435 (800) 783-5312 Fax (903) 848-0632 Fax Just when you thought you'd never find a source for DB23 connectors...this place has them. And not just solder cup style, but even straight and right- angle board mount! They also carry many other connectors and parts. @endnode @node "Chip Merchant" Chip Merchant 4870 Viewridge Avenue San Diego CA 92123 (800) 426-6375 (619) 268-4774 (619) 268-0874 Fax So far, I've had limited experience the Chip Merchant, but it's all been good. They have very low prices on SIMMs and common chips like processors. @endnode @node "Dalco Electronics" Dalco Electronics 275 Pioneer Boulevard Springboro OH 45066 (800) 445-5342 (513) 743-8042 (513) 743-9251 Fax (513) 743-2244 BBS Extremely large selection of connectors, cables (including the relatively rare SCSI-2 and 2.5-inch IDE hard disk varieties), and pretty much everything in the way of computer assemblies. Oriented towards computer end-users. They will custom-build cables. Service is good, prices are excellent, and their catalog is filled with basically neat stuff. @endnode @node "Digi-Key Corporation" Digi-Key Corporation 701 Brooks Ave. South PO Box 677 Thief River Falls MN 56701-0677 (800) 344-4539 (218) 681-3380 Huge assortment of electronic components, including chips, heat sinks, cables, connectors, fans, and every other electronic part you can think of except DB23s. Prices tend to be a little higher, which is offset somewhat by the fact that they have such a large selection. Their catalog can be considered a reference work. Oriented towards electronics designers and experimenters. @endnode @node "Hosfelt Electronics" Hosfelt Electronics 2700 Sunset Boulevard Steubenville OH 43952-1158 (800) 524-6464 (800) 524-5414 Fax Source for Panasonic replacement fans and other parts. I've been very happy with this company: they have things in stock, ship quickly, and have extremely low prices. @endnode @node "JameCo Electronic Components" JameCo Electronic Components 1355 Shoreway Road Belmont CA 94002-4100 (800) 831-4242 (415) 592-2503 Fax Large selection of chips, power supplies, and other electronic components, including some that can be extremely difficult to find elsewhere. @endnode @node "JDR Microdevices" JDR Microdevices 1850 South 10th Street San Jose CA 95112-4108 (800) 538-5000 Orders (24-Hour) (800) 538-5005 Fax (800) 538-5002 Tech Support (408) 494-1430 BBS Chips, cables, hard drives, generic computer parts. Oriented towards the end-user, quick to deliver, inexpensive, and nice on the phone. @endnode @node "Marlin P. Jones & Associates" Marlin P. Jones & Associates PO Box 12685 Lake Park FL 33403-0685 (407) 848-8236 (407) 844-8764 Fax Chips, connectors, electronics and computer parts, much of which is surplus. They sometimes have parts unavailable elsewhere, like blue LEDs. Oriented towards electronics experimenters and designers. @endnode @node "MCM Electronics" MCM Electronics 650 Congress Park Drive Centerville OH 45459-4072 (800) 543-4330 (513) 434-6959 Fax Large assortment of parts. Oriented towards electronic repair shops. @endnode @node "Memory World" Memory World 3392 Progress Drive, Suite B Bensalem PA 19020-5899 (215) 244-7930 (215) 244-7932 Fax Source for SIMMs, ZIPs, other memory, Motorola processors. And they even know what an Amiga is! Prices tend to be excellent. @endnode @node "Parts Express" Parts Express 340 E. First Street Dayton OH 45402-1257 (800) 338-0531 (513) 222-4644 Fax Chips, cables, other parts. Oriented towards electronic repair shops. @endnode @node "Redmond Cable" Redmond Cable (206) 882-2009 (206) 883-1430 Fax (615) 478-5760 East Coast (615) 472-3647 East Coast Fax Excellent source of very unusual cables and connectors. They will custom- build cables or just sell the parts. They had SCSI-2 panel mount female connectors, which I was unable to locate anywhere else. @endnode @node "Editor" Editor And Compiler Of The A4000 Hardware Guide ----------------------------------------------------------------------------- Warren Block 602 St. James Rapid City SD 57701-3658 (605) 342-1632 (voice) wblock@rapidnet.com Music that may have contributed to the mood of this guide, and that has definitely influenced me: Mark Knopfler: Golden Heart Pink Floyd: Meddle, Obscured By Clouds Dire Straits: Dire Straits, Communique, and pretty much all the others @endnode @node "Credits" Credits ----------------------------------------------------------------------------- People who have contributed information to this document, either directly or by posting public Usenet or BBS messages that have revealed information that was incorporated into this document: Antony Alonso Kerry Gray Patrik Nordvall Bruce Abbott Francois Groleau Daniel Oberlin Ralph Babel Dave Haynie John Palmer Chuck Baker Gregory Helleren Michael Perbix Volker Barthelmann Gene Heskett Kenneth Perto Gary Bates Scott Hood Troy Pladson Bryan Beecher David Houlden Dave Platt Rainer Benda Thomas Huber Kent Polk Warren Block Kjell Irgens Thomas Radtke Martin Blom Randell Jesup Mike Redrobe Keith Burns Brian Jones Sean Riddle Tom Conlin Dan Karlsson Rhett Rodewald Randy Consemulder Oliver Kastl Greg Scott Steve Crippen Larry Keller 0laf 'Rhialto' Seibert John Crookshank John Kelly C. Deforrest Smith Dale Currie Steve Kelsey Stephen Smith Richard Davey Mario Kemper Jeroen Steenblik Joachim Deussen Dr. Peter Kittel Ben Sutter Ethan Dicks Paul Kolenbrander Derek Taylor Joanne Dow Jeff Koons Barry Tigner Stephen Dowdy Randy Kruszka Mitch Thompson Jim Drew Christopher Laprise Calum Tsang Peter Ducker Erik Lindberg Sami Waulu Niklas Edmundsson Don Maddox Doug Warner Jacob Ellis Scott Marlowe Matt Weatherford Michael van Elst David Martin Thomas Weeks Bob Emery Michael Martin Ulrich Weise Bernd Ernesti Peter McGavin Lothar Werzinger Jeff Gill Francesco Meani Phil Wright Scott Goffman Gerry Murphy Heinz Wrobel Denny Goodrich David A. Newman I'd like to thank everyone for their graciousness in sharing this very valuable information with the world, and in putting up with my seemingly endless questions on the Amiga 4000. Thank you all! Finally, a special note of thanks to Urban Müller and Fred Fish for the vital services they provide. @endnode @node "What's New With This Version" What's New With This Version ----------------------------------------------------------------------------- Several kind people have submitted new reference sections for boards and other information previously not covered. You'll notice that these new sections have a byline for the author. As before, sections without a byline can only be blamed on your humble @{"Editor" link "Editor"}. New Sections @{"A2065 Reference" link "A2065 Reference"} Reference for the A2065 Ethernet board. @{"A4091 Reference" link "A4091 Reference"} Reference for the A4091 SCSI controller. @{"Benetech Electronic Supply" link "Benetech Electronic Supply"} A source for DB23 connectors and parts. @{"Card Guide Problems" link "Card Guide Problems"} Possible problem with A4000T expansion. @{"DKB 3128 Reference" link "DKB 3128 Reference"} Reference for the DKB 3128 RAM board. @{"Expansion Cards Not Recognized" link "Expansion Cards Not Recognized"} More common problems. @{"FastLane Reference" link "FastLane Reference"} Reference for the FastLane controller. @{"Floppy Drive Cable Problems" link "Floppy Drive Cable Problems"} Problems with internal floppy cables. @{"Hydra Reference" link "Hydra Reference"} Reference for the Hydra Ethernet board. @{"Memory SIMM Problems" link "Memory SIMM Problems"} More common problems. @{"MultiFace and FastLane Problems" link "MultiFace and FastLane Problems"} Using the MultiFace III and FastLane together. @{"MultiFaceCard III Reference" link "MultiFaceCard III Reference"} Reference for the MultiFaceCard III. @{"NEC 3D Reference" link "NEC 3D Reference"} Reference for the NEC 3D monitor. @{"Power Supply Pin-Outs" link "Power Supply Pin-Outs"} More pin-out definitions. @{"Retina Z2 Reference" link "Retina Z2 Reference"} Reference for the Retina Z2 graphics board. @{"Speeding Up IDE Boot-Up" link "Speeding Up IDE Boot-Up"} IDE hardware patch. Enhanced Or Revised Sections @{"1084 Reference" link "1084 Reference"} @{"1940/1942 Reference" link "1940 Reference"} @{"1950 Reference" link "1950 Reference"} @{"1960 Reference" link "1960 Reference"} @{"A2091 Reference" link "A2091 Reference"} @{"A2320 Reference" link "A2320 Reference"} @{"A3640 Reference" link "A3640 Reference"} @{"Common Questions" link "Common Questions"} @{"Credits" link "Credits"} @{"Dead Machine Problems" link "Dead Machine Problems"} @{"Editor" link "Editor"} @{"Emplant Reference" link "Emplant Reference"} @{"IDE Drive Problems" link "IDE Drive Problems"} @{"Internal IDE Hard Disk Connector Pin-Outs" link "Internal IDE Hard Disk Connector Pin-Outs"} @{"Motherboard Jumpers" link "Motherboard Jumpers"} @{"Oktagon Reference" link "Octagon Reference"} @{"Power-Up Self Test" link "Power-Up Self Test"} @{"Redmond Cable" link "Redmond Cable"} @{"SCSI Examples" link "SCSI Examples"} @{"SCSI Pin-Outs" link "SCSI Pin-Outs"} @{"SCSI Reselect Problems" link "SCSI Reselect Problems"} @{"VGA Monitor Pin-Outs" link "VGA Monitor Pin-Outs"} @{"Warp Engine Reference" link "Warp Engine Reference"} @endnode @node IndexNode "Index" Index ----------------------------------------------------------------------------- @{"-5V Power Problems" link "-5V Power Problems"} @{"1081 Reference" link "1081 Reference"} @{"1084 Reference" link "1084 Reference"} @{"1085 Reference" link "1085 Reference"} @{"1940 Reference" link "1940/1942 Reference"} @{"1950 Reference" link "1950 Reference"} @{"1960 Reference" link "1960 Reference"} @{"68030 Processor Board Reference" link "68030 Processor Board Reference"} @{"A2060 Reference" link "A2060 Reference"} @{"A2065 Reference" link "A2065 Reference"} @{"A2091 Reference" link "A2091 Reference"} @{"A2320 Reference" link "A2320 Reference"} @{"A3640 Reference" link "A3640 Reference"} @{"A4091 Reference" link "A4091 Reference"} @{"Altex Electronics" link "Altex Electronics"} @{"Ariadne Reference" link "Ariadne Reference"} @{"Backplane Problems" link "Backplane Problems"} @{"Battery Problems" link "Battery Problems"} @{"Benetech Electronic Supply" link "Benetech Electronic Supply"} @{"Boards" link "Boards"} @{"Cable Routing Problems" link "Cable Routing Problems"} @{"Card Guide Problems" link "Card Guide Problems"} @{"Chip Merchant" link "Chip Merchant"} @{"Common Problems" link "Common Problems"} @{"Common Questions" link "Common Questions"} @{"Connecting VGA Monitors" link "Connecting VGA Monitors"} @{"Connector Pin-Outs" link "Connector Pin-Outs"} @{"Credits" link "Credits"} @{"Dalco Electronics" link "Dalco Electronics"} @{"Dead Machine Problems" link "Dead Machine Problems"} @{"Definitive Buster" link "Definitive Buster"} @{"Digi-Key Corporation" link "Digi-Key Corporation"} @{"DKB 3128 Reference" link "DKB 3128 Reference"} @{"Drives" link "Drives"} @{"Editor" link "Editor"} @{"Emplant Reference" link "Emplant Reference"} @{"Expansion Cards Not Recognized" link "Expansion Cards Not Recognized"} @{"External Floppy Port Pin-Outs" link "External Floppy Port Pin-Outs"} @{"External SCSI Connector" link "External SCSI Connector"} @{"Fan Problems" link "Fan Problems"} @{"FastLane Reference" link "FastLane Reference"} @{"Floppy Drive Cable Problems" link "Floppy Drive Cable Problems"} @{"Green Display Problems" link "Green Display Problems"} @{"Hosfelt Electronics" link "Hosfelt Electronics"} @{"Hydra Reference" link "Hydra Reference"} @{"IDE Drive Problems" link "IDE Drive Problems"} @{"Idek Iiyama Vision Master 17 (MF-8617) Reference" link "Idek Iiyama Vision Master 17 (MF-8617) Reference"} @{"Internal Floppy Connector Pin-Outs" link "Internal Floppy Connector Pin-Outs"} @{"Internal IDE Hard Disk Connector Pin-Outs" link "Internal IDE Hard Disk Connector Pin-Outs"} @{"Internals" link "Internals"} @{"Introduction" link "Introduction"} @{"JameCo Electronic Components" link "JameCo Electronic Components"} @{"JDR Microdevices" link "JDR Microdevices"} @{"Joystick Port Pin-Outs" link "Joystick Port Pin-Outs"} @{"Keyboard Port Pin-Outs" link "Keyboard Port Pin-Outs"} @{"Keyboard Self-Test" link "Keyboard Self-Test"} @{"Main" link "Main"} @{"Marlin P. Jones & Associates" link "Marlin P. Jones & Associates"} @{"MCM Electronics" link "MCM Electronics"} @{"Memory SIMM Problems" link "Memory SIMM Problems"} @{"Memory World" link "Memory World"} @{"Mitsubishi DiamondScan AUM-1381A Reference" link "Mitsubishi DiamondScan AUM-1381A Reference"} @{"Monitors" link "Monitors"} @{"Motherboard Jumpers" link "Motherboard Jumpers"} @{"MultiFace and FastLane Problems" link "MultiFace and FastLane Problems"} @{"MultiFaceCard III Reference" link "MultiFaceCard III Reference"} @{"NEC 3D Reference" link "NEC 3D Reference"} @{"Oktagon Reference" link "Octagon Reference"} @{"Other Video Problems" link "Other Video Problems"} @{"Parallel Port Pin-Outs" link "Parallel Port Pin-Outs"} @{"Parts Express" link "Parts Express"} @{"Power Supply Pin-Outs" link "Power Supply Pin-Outs"} @{"Power-Up Self-Test" link "Power-Up Self-Test"} @{"Processor Board Mounting" link "Processor Board Mounting"} @{"Processor Cooling" link "Processor Cooling"} @{"Redmond Cable" link "Redmond Cable"} @{"Retina Z2 Reference" link "Retina Z2 Reference"} @{"SCSI Drive Problems" link "SCSI Drive Problems"} @{"SCSI Examples" link "SCSI Examples"} @{"SCSI Pin-Outs" link "SCSI Pin-Outs"} @{"SCSI Reselect Problems" link "SCSI Reselect Problems"} @{"Seagate ST3096A/ST3120A/ST3144A Reference" link "Seagate ST3144A Reference"} @{"Serial Port Pin-Outs" link "Serial Port Pin-Outs"} @{"Slow A2091 Problems" link "Slow A2091 Problems"} @{"Sources" link "Sources"} @{"Speeding Up IDE Boot-Up" link "Speeding Up IDE Boot-Up"} @{"Tips" link "Tips"} @{"VGA Monitor Pin-Outs" link "VGA Monitor Pin-Outs"} @{"Video Banding Modification" link "Video Banding Modification"} @{"Video Banding Problems" link "Video Banding Problems"} @{"Video Port Pin-Outs" link "Video Port Pin-Outs"} @{"Warp Engine Reference" link "Warp Engine Reference"} @{"What's New With This Version" link "What's New With This Version"} @{"Zorro-III Problems" link "Zorro-III Problems"} @endnode