Big Book of Amiga Hardware

Standard Specifications

 

Case Type:	Desktop
Processor:	68000@7.14MHz (NTSC) / @7.09MHz (PAL)
MMU:	None
FPU:	None
Native Chipset:	OCS Agnus/Denise (Model A & B) ECS Agnus/OCS Denise (Model B) ECS Agnus/Denise (Model B & C*)
Native Kickstart:	V1.1 (Model A) V1.2 (Model A & B) V1.3 (Model A & B) V2.04 (Model B & C)
Bus Controller:	None (Model-A's only) Buster Rev 1 (pre-production?) Buster Rev 2 ('tower' PCB with missing logic, made for early production) Buster Rev 3 (final/common)
Expansion Slots:	5 x 100pin Zorro II slots 1 x Genlock slot (NOT Zorro-inline, Model A only) 1 x OCS/ECS Video Slot (NOT Zorro-inline, Model B & C only) 2 x inactive 8/16bit ISA slots (2x Zorro-inline) 2 x inactive 8/16bit ISA slots (NOT Zorro-inline, PCB holes for 16-bit slot extensions not populated) 1 x 86pin CPU Fast Slot (limited functionality on Model A)
Standard CHIP RAM:	512K (Rev 3 & 4 Model A's) 512K Reconfigurable to 1MB (US 3.x/4.x/5.x) 1MB (Rev 6.x)
RAM sockets:	None
Hard Drive Controllers:	None
Drive Bays:	2 x 3.5" (with faceplates) 1 x 5.25" (with faceplate)
Expansion Ports:	1 x DB25 Male Serial 1 x DB25 Female Parallel 1 x DB23 Male RGB Video 1 x DB23 Female External Floppy 2 x DB9 Male Joystick/Mouse 2 x RCA-F Audio (Left/Right) 1 x RCA-F Composite (Models B & C only) 1 x large 5pin DIN Keyboard Connector
Floppy Drive:	1 x 3.5" Internal 880K Floppy Drive
Motherboard Revisions:	Model A (German design derrived from the A1000 w/ an expansion card sidecar unit):  - Rev 3.0 (pre-production version)  - Rev 4 (limted production version, 1986) Model B (US / West Chester design partially derived from the German Rev 3/4 by Dave Haynie):  - Rev 3.8 (pre-production)  - Rev 3.9 (pre-production)  - Rev 4.0 (Rev 3.9+final FCC fixes/changes, never released due to an error)  - Rev 4.1 (1st production version, 1987)  - Rev 4.2 (undo excessive FCC-releated tweaks)  - Rev 4.3 (fixes for CSG Gary, a phased in replacment for Toshiba Gary)  - Rev 4.4 (Tick-line fixes, as some production PSU sources changed)  - Rev 4.5 (Adjustment due to buffer-part vendor tollerance variations seen with accelerators/DMA)  - Rev 5.0 (very rare, estimated 5 boards in existence, never put in production) Model C* (unofficial designation) - 'Rev 6' is printed in the PCB in two places.  Stickers identified revisions 6.1+.  - Rev 6    (ECS Agnus/1MB ChipRAM introduced.  Buggy with expansion cards, TOD clock noise.)  - Rev 6.1 (Addressed the TOD clock noise)  - Rev 6.2 (Addressed the expansion bus issues)  - Rev 6.3 (Kickstart 2.04 ROM introduced)  - Rev 6.4 (ECS Denise introduced, last 6.x revision at A2000 End Of Sale*)  - Rev 6.5 (Post-End of Sale replacement motherboards w/spare part 8375 Agnus+adjustments*) * The A2000-C/A2000C designation has conflicting origins.  The US / West Chester 'B2000' version of the A2000 is clearly identified at the front-left corner, in copper-trace, on all Rev 4.1-4.5 and Rev 6 PCBs manufacturing.  The latter Rev 6 also has it printed in white silkscreen near the ChipRAM.  1) The community loosely adoped the Rev 6 as the 'C' version, with it's (then) updated 1MB Agnus/ChipRAM and (later) other updated Kickstart 2.04 & Denise parts, which followed seperately.  This 'C' designation was never collectively identified in any C= documentation.  2) An extremely late production (more likely a replacement A2000 issued under warranty), with Rev 6.4 clearly labeled, but the post-8372A production spare part (NTSC) 8375 Agnus w/318069-16 (and required mod), which came with an A2000C sticker on the rear of the case.
Battery Backed Up Clock:	Yes. Model A used square varta batteries, Model B & C uses "barrel" shaped batteries.

The A2000 was launched in 1986.  It was aimed at the professional productivity market. The initial version was designed in Germany.  It was based on the A1000.  Later that year, Commodore US was already working on the cost-reduced A500, and expanded from that effort for their version of the A2000 (referred to as A2000-B or B2000).  In 1987, this became the high-end Amiga model for which the A2000 is widely known as. There are probably more variations and revisions of this model than any other Amiga. Not only were at least 15 separate motherboard revisions produced, the model spans several versions of Kickstart and two chipset generations. Several models with additional peripheral options as standard offerings were also produced (A1500, A2000HD and A2500).

It was initially possible to tell the difference between the OCS and ECS versions of the A2000 by the colour of the front panel LEDs. The OCS versions had a red power LED and a green HD LED, while the ECS versions have a green power LED and an orange or yellow HD LED. The change in LED colors coincided with a collective electronics industry agreement at the time for the use of certain LED colors for certain purposes.  The easiest way to separate an A2000-A (the A1000-based German version) from the later, A500-based models, is to see if there is a composite connector beside the 2x RCA sound connectors at the back. If there is none, it's the original version.

All Amiga 2000s were sold with 1MB of RAM, but how this was implemented varied over motherboard revisions. The German (model A) designs have 512K chip RAM on the motherboard and 512K Fast RAM on an expansion card fitted to the CPU Fast slot. The US 4.x (model B) motherboards had 1MB of RAM soldered to the motherboard with 512K configured as CHIP RAM and 512K as Fast. The Fast RAM on these models can be reconfigured into Chip RAM by upgrading to the ECS Agnus and two jumper/pad modifications. Rev 6.x ("model C") boards have 1MB of RAM soldered to the motherboard, all of which is configured as CHIP RAM.

Early Rev 4.1 A2000's were fitted with a Buster Tower which sits in the Buster socket. This is because the Buster had a logic mistake.  It was missing a term signal in one of the backplane buffer control signals. The tower contained an additional PAL which fixed this problem.  An updated Buster soon negated the need for the the tower.

Motherboard revisions according to Dave Haynie (Commodore Engineer)

"The A2000 Model A was a German design, like an A1000 in a different box. 512K CHIP and non-standard processor and video slots. Yes, The orginal A2000 was designed in Germany. It was based on an integration of the A1000 motherboard design and the example Zorro II backplane from "Schematics and Expansion Specifications", the A1000 hardware manual. It used the thin Agnus, which handled only 512K of DRAM. They added a "Genlock" slot, which was essentially just the 23-bit video signals on an internal connector, and the "MMU" slot, which was essentially just the A1000 external edge connector on an internal slot. The machine shipped with 512K of Fast RAM in this slot, though the case was slow and had some reliability problems. About 60,000 of these machines were made, so I've heard.

A2000-B Redesigning The German Model:

This was the original A2000-CR (Cost Reduced) that I designed. I took the A500 chip set and integrated most of the control logic for the expansion bus into a gate array called Buster (the original thin Buster). This board was code-named the "B2000", not out of any special letter-sequencing, but because it was made from some A500 parts -- the A500's code-name was the "B-52".

To the German design, I added a second video connector, which brought out the remaining 8-bits of digital the video signal and some parallel port lines (just in case someone needed a way to control something out there). I also designed a "coprocessor interface" for the internal CPU slot, which allowed processor cards to be added without the need to remove the 68000. And of course, given the much higher integration (thanks due to Buster, Gary, Fat Agnus, etc), I was able to fit 1MB of RAM on the motherboard, even though I went to a 2-layer board (management's demands for cost, not my idea, though I did make it work OK.)

REV 3.9:

About 200 pre-production B2000s were released, as "Dealer Demo" units, never meant to be sold. That's labelled Rev 3.9, of course some dealers sold them, since the A2000 was hot. These may have problems with CPU slot devices since, like the orginal A2000, there was no additional buffering on the clocks that drove that slot (and timing is very critical in many CPU slot applications).

REV 4.x Series:

The Rev 4 A2000 was around most of the system's life. This was really just the Rev 3.9 with all the finalized changes. Extra buffering splits Agnus clocks between internal, expansion, and CPU slots. Rev 4.1 immediately corrected one missing signal on the secondary video slot (no plain Rev 4.0 boards were released). Rev 4.2 was some kind of production fix, I think it removed some overzealous FCC hacks, like the extra filter capacitors on the keyboard connector. Rev 4.3 tweaked a few pullup resistor values, to adjust for differences in the new CSG-made Gary chip (a full custom part) versus the Toshiba-made Gary chip (a gate array). After some 150,000 or so A500/A2000s, the CSG Gary was phased in. Rev 4.4 I believe made some changes to filtering of the TICK line, to account for some new power supplies that had been particularly noisy on that line. Rev 4.5 increases the value of the pullup resistor on the BAS* line of the expansion bus; there was a strange problem in conjunction with expansion bus DMA turnaround (what happens after a DMA device gives the bus back), BAS*, certain vendors of the 74ALS245 buffer that sits between BAS* and the local bus's AS*, and accelerator cards that could cause crashes. Add a 1K or so pullup to pins 11 and 20 of U605 on pre-4.5 boards to fix this problem (this was the only severe problem addressed by any subrevision).

REV 6.x Series:

The Rev 6 A2000 changed the DRAM array from 256K x 1 parts to 256K x 4 parts. Some layout and FCC changes were made, I don't know just what, I was kind of out of the picture at the time (Rev 5 was just Rev 4.5 with the DRAM changes, nothing else. Rev 6 did more, I wasn't really involved). The original Rev 6 board has some noise problems on the expansion bus, possibly due to a new production of 68000s from Motorola that happed to be used at the same time. Rev 6 also had some extra noise on its time-of-day clock. Rev 6.1 corrects the clock noise, Rev 6.2 addresses the expansion bus noise.

The Rev 6 A2000 was never called the A2000C, except maybe by geeks on usenet :-).

(~) 07/05/2018 - BBoAH user Christian Stich has discovered an Amiga 2000 - with a A2000C sticker on the back (see photo section). The machine is a rev 6.4 motherboard with the NTSC 8375 VBB Agnus (the version that is pin compatible to the 8372A, other than the Test/NTSC/PAL pin).

A2000 Rev 5

While there are many Amiga 2000s in the world, the Rev 5 is extremely rare. As far as I know, only about five boards were made. The purpose of the Rev 5 board was simply to add higher density memory to the Amiga 2000. The original design used sixteen memories for a bank of memory, this design used four (there are two banks). This was the last Amiga 2000 motherboard I had all that much to do with..

At the time, the Rev 4.x was in production, and the new board was desired for new production, as the price crossover from the 256K x 1 to the 256K x 4 memories had been reached. However, the Rev 5 board was completed ahead of the need for it. So the PCB guys did some additional cleanup work on the design, and I think the FCC people got involved, too. Regardless, the resulting Rev 6 motherboards were plagued with problems (in fairness, some were due to the switchover to a 68000 with faster signals, that was also noisier). - Dave Haynie

Additional

"This first motherboard is a Rev 4 German Amiga 2000 motherboard. This is the one based on the original A1000 logic, not the A500 logic. It supports only 512K of on-board memory, and the CPU slot is limited in function. This board appears to be in good physical order, though it hasn't been tested. There is a note attached:

Works Good!! Thanks FYI: Board has same 1 <-> 31 swap as early A500's -Bryce

Apparently a note from Bryce Nesbitt. The "1 <->31 swap" was for pin numbers on the ROM. Commodore has apparently received bogus information about which pin would supply the next address line, as these ROMs grew from 256K to 512K. The German A2000 and some early A500s had this problem, as the note says. The correct information prevented this issue in US Amiga 2000s.

The board on the top right is a US "B2000" Rev 3.8 (preproduction) Amiga 2000. This has been stripped of its ROM and its 8520s, but it does include the Buster "Tower". That tower's my fault, lemme 'splain. When I took over the Amiga 2000 project, in the fall of 1986 (well, after the shock wore off), I set out to implement the first "Buster" (Bus controller) chip. This did with the expansion logic much of what Gary and the Fat Agnus did to the original Amiga 1000 logic: shrunk it in price and in size. Since I was, in essence, copying the German design (they _do_ look similar, eh), I took the PAL equations from the Amiga 2000 on the left as my basis for the Buster chip.

Whoops. There was a bug in the German PALs. There are a set of buffers between the A2000's local bus (directly on the 68000) and the Zorro bus. When a Zorro bus master talks to a Zorro bus target, these buffers are supposed to point away from Zorro, toward the local bus. During pre-production testing at the Commodore plant in Taiwan, I discovered that wasn't the case with the German A2000, or my shiny new Buster chip -- motherboard buffers fought with anything driving the expansion bus in this case. Whoops.

In Germany, they fixed it with a new PAL. The Buster "tower" fixed it here, until a corrected Buster chip could be produced.

The board on the bottom left is another "B2000" Rev 3.8 board. This one still has those precious 8520s, but it has been stripped of the 68000, the ROM, and any sort of Buster chip. All A2000 Busters do work in all motherboards, you don't need a towered Buster for these older boards. However, the Rev 3.8 board was questionable with CPU cards -- the primary reasons for Rev 4 were FCC issues and beefing up the whole coprocessor mechanism with additional buffers.

This final board is a "B2000" Rev 4.3 motherboard, often considered the first "golden" version of the Amiga 2000." - Dave Haynie

-------------------------------------
Work In Progress (to resolve motherboard revision identification ambiguidy)

The German 'A2000' motherboard is defined by the following features:

 - It is a 4-layer motherboard, with solid Ground/VCC layers on the inside.
 - Production motherboards are simply labeled 'Rev 4' (no sub-revision '.x' designations or revision labels)
 - It has 512K of ChipRAM on the motherboard using 16x 256x1 DRAM parts.
 - They use the 512K DIP Agnus chip from the A1000.
 - It does not have the 'Buster' logic controller.  Discreet 74-series parts and PALs make up the control logic.
 - It does not have the 'Gary' logic chip.  Discreet 74-series parts and PALs make up the control logic.
 - It used a black rectangular battery, and the Oki clock chip is to the right of the Kickstart socket
 - It has only the single slot connector on the video slot.
 - It does not have the Composite Out RCA jack.
 - It's internal 100-pin expansion bus is effectviely the proposed A1000 Expansion Bus cage design.
 - Accelerator use require removal of the 68000 CPU.  The 86-pin slot lacks the _BOSS signal/coprocessor-logic.

The US/West Chester  'B2000' became the standard A2000 motherboard, and their revisions were as follows:

Rev 3.8, Rev 3.9, Rev 4.0 - Pre-production/Dealer Demo/Developer - Dave Haynie has documented what transpired across these early revisions, including the cost-reductions implemented.  This helped to produce stability in the expansion bus, and upgrade the CPU/Video slot functionality, over the German A2000/A1000-era design, and led to the Rev 4.x production series.

Rev 4.1-4.5 - Production Release.  The modification information between revisions can be distilled from the C= Tech Topics / Field Service Bulletins.  The 4.x series gained the following observable features:

 - The Agnus becomes an 84-pin PLCC chip (Fat Lady label on the silkscreen)
 - The motherboard uses 32x 256x1 chips (in 2 banks) for 1MB of RAM
 - 512K is configured as ChipRAM (w/an 8370/8371 Agnus), and 512K is configured as 'slow' FastRAM at $C00000
 - It is a cost-reduced 2-layer motherboard.
 - The Buster chip (318075) appears at the rear of the Zorro expansion slots, absorbing several 74-series logic chips and PALs, to provide bus arbitration and expansion slot stability.
 - The Gary chip (from the A500 cost reduction) is appars in the middle of the motherboard, absorbing several 74-series logic chips and PALs, improving system stability.
 - A barrel battery (Varta or other) and the Oki clock chip are located close together.
 - The video slot gains an additional connector to provide more video-related signals, and providing access to the parallel port signals for communications to the video card.
 - A Composite Out (b/w-only) RCA jack appears next to the L/R Audio jacks.
 - The silkscreen has smaller embedded numbers (1-5) for the revision markings (the .# part) on all in the series.  Paper labels were also added for factory and/or (less often) dealer/field service center updates.

Notable chip-swap changes from the C= TechTopics include replacement of a 'towered' buster with the revised chip, and replacement of the Toshiba-made Gary with the MOS-made Gary.  These would only apply to the earliest 4.x revisions.  All other modifications will generally require soldering efforts to perform.

The A2000 4.x (and earlier) motherboards will have problems with programed 27C400 (and equivlent) EPROMs.  The ROM sockets are not wired for one pin that is present on an EPROM, but not on a mask-ROM that might have been provided from C=.  The pin controls whether the EPROM is in 8-bit or 16-bit data mode.  A connection will be needed between pins [connection detail needed here] to resolve the issue.

Rev 5 - Non-Production Release.  Dave Haynie expected 256x4 DRAM parts would eventually become cheaper than 256x1, and so provided an updated revision at the end of the Rev 4.x series that would use them.  This revision was the basis for the Rev 6 series.

Rev 6 Series - Production Release.  This series gained the following obvious features:

 - The 1MB of onboard RAM was derived from 8x 256x4 DRAM chips (vs 32x 256x1)
 - The Agnus chip was the 'ECS' 8372A, and the onboard memory was configured natively for 1MB ChipRAM

The Rev 6(.0) & 6.1 revisions had noise/expansion bus problems that were resolved as described in the C= TechTopics bulletins.  Dave Haynie has documented some of the issues that he knew about.  Simply translated, they amount to a change by (M)otorola in some CPU mask electrical properties of the (then) production MC68000, and some over-zealous FCC-compliance adjustments since he did the Rev 5. 

The corrections resulted in the Rev 6.2, which is stable.  It is considered the standard for the Rev 6 series, and was last technical revision.  A pair of termination resistor packs soldered in on the expansion-bus side of the U603/U604 27LS245N bus buffers is the most notable adjustment.

Rev 6.3 was the production change from Kickstart 1.3 to Kickstart 2.04.  This update (Kickstart ROM) was also sold as an upgrade part in the field through dealers (applied to both Rev 4.x and Rev 6.x).

Rev 6.4 was the production change from the 8362 OCS Denise to the A3000's 8373 ECS/Super Denise.  An 8372A ECS Agnus was required. (mentioned only for the benefit of the 4.x series, which may still have the 512K 8370/8371 Agnus).

All motherboards of this series will have 'Rev 6' on the silkscreen.  All revisions above the base Rev 6 are supposed to be 'paper label' marked when updated.  This didn't always happen.  Sometimes the revision labels were put at the back corner or somewhere under the PSU/Drive cage area.  Many have probably fallen off altogether by now.

There was a period of production for both the A500 Rev 6, and the A2000 Rev 6, where C= had purchased a large quantity of the wrong type of 256x4 memory chips.  These chips were of the type 'static column'.  The parts typically had a '258' vs '256' in the maker's part number.  A small upside-down PCB appears in the position of the U540/U541 74F245 buffer chips (between the Kickstart ROM/ChipRAM on the A2000) as a solution to allow use of these memory parts.  The net effect is no changes to the system's behavior.

After the sale of the A2000 ended, motherboards were still available for warranty replacement (1yr) and for sale as replacement for a time after.  The 8372A 1MB Agnus part had also run out of production during this time.  When spare parts were requested by the dealers/service centers, C= had to produce Agnus parts that were pin-compatible to the older 8372A's socket.  This spare is the 318069-16/-17 (PAL/NTSC) 'VBB' Agnus parts.  As the new parts are (again) PAL or NTSC-defaulted on power up (unlike the 8372A, which used a pin to select the mode), a modification is needed on that pin depending on the use in a PAL or NTSC machine.  This revision - for the use of this Agnus part - was described as 'Rev 6.5' with the parts made available for purchase or a warranty exchange.  It did not appear in any C= TechTopic bulletins, and the documented mod may only have been included with the part when ordered.

 

Jumpers

These setting are based on the manual for an Amiga 2000 Revision 4.x and is compliant with Rev 6.x Amiga 2000 boards as well.

J101	This jumper determines the high order address bit for Fat Agnus. In it's normal position, the high-order bit is A23; in it's other position, this bit is A19. The current OCS Fat Agnus chip requires the A23 signal for proper management of the memory at $C00000. The later ECS Fat Agnus chips addressing 1MB do things differently; this keeps the current board compatible with simple enhancements.
J102	This jumper, if available, configures the A2000 for NTSC, or PAL operation, when used with an 8372A 1MB Agnus (any part marked 318069-02) to support switching between NTSC and PAL at power-on. Native 4.x motherboards may have this jumper pre-set for the targeted PAL/NTSC market the unit was sold in, but it does not alter the power-on mode of 8370 (NTSC)/8371(PAL) these 512K Agnus chips.  All motherboards which are revised for the 8375 1MB Agnus (318069-16 PAL, or 318069-17 NTSC) spare part, by default, no longer support switching via hardware at power on.  They are, however, software-switchable between modes in the OS.
J200	This jumper determines the time base used for the 50/60Hz CIA timer chip. In the normal posistion, the 50/60 TICK clock, based on AC line frequency, is used as a time base. In the alternate position, the verticle sync pulse from the video section is used. The system will not operate properly without one of these clocks.  (If the PSU caps become aged, and/or the AC frequency clock 'tick' becomes 'dirty', moving this jumper to the video section position may improve system stability, but could also affect Genlock operations).
J301	This jumper is closed to add a second internal floppy drive, open to leave the second floppy external to main unit box. The Amiga expects an ID bit stream from each floppy drive; this lets it determine the drive type. External floppies have this ID circuitry on board, but as it's not an industry standard capability, it has to be implemented on the A2000 motherboard to save cost on internal drives. Leaving the jumper open prevents the Amiga from seeing the ID sequence.
J500	This jumper is used to enable the 512k of RAM at $C00000 and the real-time clock. It is normally closed; opening it will disable both this extra RAM and the real-time clock.