Reverse Engineered non standard ram expansion board for use in Modular Circuit Technology C386-33.
Diagram (png,pdf). Interesting detail about design - it exposes two ways of driving CAS (Column Address Strobe) signals. Designer made it flexible enough to be able to work with different memory controllers/interleave configurations. One set of four CAS pins drives consecutive SIMM sockets in every Bank, the other 4 pin set drives whole banks per CAS line. Brilliant.
Potentially also compatible motherboards:
- Modular Circuit Technology MCT-C386-33/40
- Northgate Computer Systems
- NOVACOR, INC.
- QDI P386
- Iwill 80386 WRITE BACK CACHE
- Shuttle
- Digicom DIGIS-386WB
Incompatible platforms using similar slot:
- Modular Circuit Technology MCT-386-M
Grand daddy of C386-33 from 1987, 386DX 20 and IIT FPU 387 20Mhz. Looks like it might be dual layer PCB! :o - OGIVAR, INC. SYSTEM 386 unlikely as its a 386XS
- POSITIVE CORPORATION PC340 wrong slot position
- ECS 386A Same Chips & Technologies CS8230 chipset, same slot, not sure about slot position
- Micronics
Those six pictures uploaded by McBierle was the only source of information. Provided all vias, back tracks and about 25% front ones. Rest is obscured by components and had to be deduced logically using knowledge of vintage memory subsystems.
McBierle authored Vogons Story about C386-33 and me thread.
100% reconstructed schematic. 95% ready for manufactuiring, remaining tasks:
- Slot position is hard to judge from pictures picture alone due to lens distortion (perspective, pillow). I had to undistort them using ShiftN and Gimp.
- Dimensions are slightly off, clearances need real measurements.
- Hard to guess which Edge Connector power pins are 5V/Ground on the connector. Only half has tantalums with painted polarization. Better to measure and verify.
- I didnt bother with decoupling capacitors before confirming dimensions.
11/9/2024: 100% schematic. 100% electrical PCB. Woke up, looked at the CAS mess and immediatelly knew how to route it, magic of a good night's sleep! 0 ERC, 0 DRC errors. 1 DRC warning for unused Via. No idea if original Designer left it unused or somehow routed the board in different way. Personally I dont see a way to use this Via in any way :)
11/8/2024: 95%. SIMM slots and DIP14 chips obscure CAS traces. Struggling hard with routing CAS lines with all the missing information, still dont undertand purpose of OR gates instead of simple buffers
11/7/2024: 85-90% Got RAS mapping correct. 90% back
11/6/2024: 60-80%. Deduced Address/WE buss pairing, Address/Data buss Edge Slot pinout. Still ~300 ERC, ~200 ERC errors.
11/5/2024: 65%. Figured all 240/244 chips used same connections to corresponding pair of Resistor Pair Packs, Designer copy pasted it 12 times.
11/4/2024: 60%. Further fixed picture distortions wih Gimp Transform functionality (what a pain!). Allowed almost ideal component placement and drawing board outline.
10/31/2024: Very VIP. Preliminarily undistorted source pictures using ShiftN to help with more precise component placement. 10% there front10% back
10/30/2024: Project start. Looking at pictures of 386RC-16 uploaded by McBierle on vogons forum and assesing feasibility. Rougly placing components more or less by eye and tracing few lines. 1% there
Interested in having a piece of hardware reverse engineered/cloned? Drop me an email.