• Please review our updated Terms and Rules here

IMSAI 8080 at the Rhode Island Computer Museum

m_thompson

m_thompson

Veteran Member
Joined
Jul 8, 2014
Messages
1,990
Location
Rhode Island, USA
The RICM just received an IMSAI 8080 donation. It contains a Cromemco ZPU, two early IMS Associates 4k static RAM boards from 1975, and IMS Associates PIO board from 1975, and an IMSAI SIO board from 1976.

We need will need get three blue switch handles, but it is otherwise in good physical condition. It has a power harness for an external device that looks like it was an 8" floppy diskette drive.

We will clean it, reform the power supply capacitors, removed the boards, power it on, and see if the power supply regulates. If so, then we can install the CPU and RAM and see if the front panel is alive.

If we replace the RAM boards with something newer and larger and add floppy diskettes, it should make a nice system.

We haven't found a schematic or manual for the early 4k IMSAI RAM boards yet. Any ideas?

Pictures of everything are here: https://www.ricomputermuseum.org/collections-gallery/small-systems-at-ricm/imsai-8080
 
Very nice! I hope to find one too once in the future... You are clearly in the good part of the world to find these machines. Good luck restoring it!

Regards, Roland
 
I made up reproductions of the original IMSAI MPU-A 8080A CPU board as I was in a similar situation - an IMSAI without the original cpu board.

I'd be happy to donate a bare board if you'd like to return to the original CPU configuration. You'd need about another $60 worth of easily available parts and and an hour or two of soldering to get a close replica of the original CPU board.

You can see the board at http://www.the-planet.org and PM me if you'd like me to ship you one.

Sorry, but I couldn't locate any docs on your rather early 4K ram board in my files.

If you're looking for a simple solution to get the system up and running for usage/demonstrations you should consider Josh Bensadon's 8080A board that includes everything you'd need to get the system running on a single card.

- Gary
 
Our current plan is to add a diskette controller and 8" floppies. If the currently installed ZPU is a problem then we will likely consider your reproduction CPU.
 
We reformed the capacitors in the power supplies and powered it up. All of the voltages look OK.

The front panel lights blink, so there are some signs of life. Examine and Deposit don't always use the switch settings for the address and data. Looks like we have some debugging to do.

Does the Cromemco ZPU that we have work with the IMSAI front panel?

I was thinking of moving the three blue paddles from Examine, Reset, and Single-Step to the missing 5, 6, and 7 locations and then buying some red switches. We might also try printing some paddles.
 
I just looked at the museums website and I am extremely impressed. It looks like quite the collection and quite the team. I assume that you work/volunteer there? What is that like? Are computers often repaired as a team or is it more of an individual effort? I will have to try and make it up there some day soon.
 
m_thompson said:
We reformed the capacitors in the power supplies and powered it up. All of the voltages look OK.

The front panel lights blink, so there are some signs of life. Examine and Deposit don't always use the switch settings for the address and data. Looks like we have some debugging to do.

Does the Cromemco ZPU that we have work with the IMSAI front panel?

I was thinking of moving the three blue paddles from Examine, Reset, and Single-Step to the missing 5, 6, and 7 locations and then buying some red switches. We might also try printing some paddles.
Click to expand...

If the CPU board has a ribbon cable connector for the cable from the front panel it should work. Make sure the connector is oriented correctly. The problem you are having sounds like the cable is backwards in the connector. This happens when the cable is wired backwards ( a common issue for home built ones ).
If it is backwards, some addresses like 0000 will work but others won't.
If it is not the cable it is likely one or more of the 7406's. These are a common failure item.
Dwight
 
It really depends on what it does do. The cable from the front panel is used to jam an instruction into the processor. As an example, the EXAMINE switch jams the JMP address into the CPU. The display on the front panel is the actual address and data on the bus. The EXAMINE NEXT is just a NOP to cause the CPU to increment its address, without doing anything. Writes are similar but actually drive data to the bus instead of just changing the address.
If the cable is reversed, the JMP instruction is 0C3 or 11000011 in binary. So, if the cable is swapped, it does the JMP as the bit reverse of C3 is C3. The addresses are scrambled though. Imagine the address as two bytes with the bits swapped. Also, there is one status bit used to tell the front panel when the instruction completes. This bit being wrong would cause the processor to sometimes not halt after just one instruction.
Still, the most common problem is bad 7406 OC buffers, used to drive the processors data lines. I've seen the flip flops used to count the states fail and surprisingly rare one of the one shots fail.
The 7406's are relatively easy to check. One can connect a scope to data lines and see how many bits of the instruction and address are loaded, when the examine switch is hit, by watching the bits on the data cable.
Another problem I saw a couple year back was with a Altair, that was said to have been working once. It had 4 solder shorts on the bus lines. I didn't check this as it was supposed to be working once.
If the CPU board you have doesn't have the cable connection, only things like reset will work. These CPU boards expect to boot from ROM.
One of the coolest disk setup is the Digital Systems disk setup. It was one of the first of two different disk systems available for the S100. It had one cool advantage over most all of the other S100 disk setups. There was no wasted space on the S100 for disk drive ROMs. On system reset it would DMA the first sector of the disk into memory. The code on that sector would bootstrap the rest of the OS ( CP/M if you like ). You could have a full 64K of pure RAM. It used several I/O location to control it after boot.
Dwight
 
Dwight Elvey said:
If the cable is reversed, the JMP instruction is 0C3 or 11000011 in binary.

Still, the most common problem is bad 7406 OC buffers, used to drive the processors data lines.

If the CPU board you have doesn't have the cable connection, only things like reset will work.
Click to expand...

I will try reversing the cable to see if it behaves better. If not, I will look for bad chips. I had to do similar front panel repairs on one of the Museum's Altairs. I will also try the memory boards in one of the Ailtairs to make sure they work, and if not find a replacement in our collection.
 
boxerbomb said:
I assume that you work/volunteer there? What is that like? Are computers often repaired as a team or is it more of an individual effort?
Click to expand...

I have been keeping the DEC PDP-8, PDP-9, PDP-12, and Teletypes running for more than a decade, and also recently revived the Altairs and the Wang LOCI-II. Lately I have help from new volunteers. Zhao, Emilio, and I just revived our SOL-20. Emilio and I revived some Intel MCS-85 boards. Lou and I revived the Intel MCS-800 system. We are always looking for new volunteers to help with projects. We could really use some help turning the now running machines into interesting interactive displays.
 
Yesterday we tried reversing the ribbon cable between the front panel and the ZPU board. That didn't improve the behavior so we put it back.

We looked at the schematic for the ZPU and found that 4x 74367 ICs drive the ribbon cable to the front panel. The legs on these socketed ICs were black, so we cleaned them. The front panel behaves better now, and for a little while we could examine and deposit RAM. We will likely remove and clean all of the socketed ICs on the ZPU, put some Deoxit on the sockets, and wiggle the ICs back into the sockets. We had to to the same to get the SOL-20 to run.
 
Sounds like they were those silver plated leads. They are effected by sulfur oxides. Areas ,that are close to coal fired power plants oxidize, effect silver faster. Car fuels used to be really bad but the newer regulations keep it to a minimum.
Anyway, it sounds like you are doing good work. I wait to hear it is booting CP/M soon.
Dwight
 
We removed all of the ICs on the Cromemco ZPU board, cleaned the black leads, and reinstalled the ICs. It is really broken now. Examine and Deposit don't work, but the problem is probably in the front panel and not in the CPU. We tested the two IMS Associates 4k RAM boards in an Altair, and the both work OK. Next time we will disassemble the front panel and start working on that.
 
After removing the front panel from the backplane, cleaning the gold fingers, and reinstalling the front panel, it looks like things are working now. I can toggle in and run little programs. We can read the Programmed Input switches on the front panel and control the Programmed Output LEDs on the front panel and on the IMSAI parallel I/O board.

We borrowed boards from a rack mount Cromemco system in our warehouse. We removed the two IMS 4k RAM boards and replaced them with a Memory Merchants 64k static RAM board. We installed a Cromemco 8k Bytesaver that has a single EPROM containing the Cromemco monitor. We will disable the upper 8k of RAM so it doesn't conflict with the monitor. The Cromemco Monitor wants a Cromemco TU-ART serial I/O board which we got from the warehouse.

We got a Cromemco Floppy Diskette controller from the warehouse and a Lobo dual 8" diskette chassis and some 5 1/4" floppies. We also found an 8" CDOS diskette in one of the drives.

We should be able to get this running as an IMSAI chassis with Cromemco boards inside. Then when we have more confidence that it is working we can swap some of the Cromemco boards for IMSAI boards.
 
We have a Cromemco rack mounted system in the warehouse that is not in very good condition. I borrowed all of the boards from it to see what works and what we could put in the IMSAI chassis. It had the same ZPU that came in our IMSAI, and it even works. The Memory Merchants 64k static RAM board works nicely. We started working on the Cromemco 8k Bytesaver because it contains an EPROM for the Cromemco Z-80 Monitor. That board has lots of corroded ICs, so it will take some work to get it working. We got a Cromemco TU-ART that we can use for the serial console for the Cromemco Z-80 Monitor. The Zebec SASI host controller won't do us any good unless we can find a SASI-MFM board to go with it. The Cromemco 16FDC floppy/serial controller should work nicely. We have both 8" and 5 1/4" diskette drives to go with it.

Once we get the Cromemco boards to work in the IMSAI chassis we will start replacing the Cromemco boards with the original IMSAI boards.
 
We are working on the 8k Bytesaver board so we can run the Cromemco Z-80 Monitor. The board is a mess of corroded chip leads and sockets. Our work yesterday shows that the 2708 EPROM might be empty. Oh well.

We tried the Cromemco 16FDC in the chassis to see if we could see the boot ROM on the board. Unfortunately with either of two 16FDC boards installed the front panel won't do EXAMINE or DEPOSIT.
 
You might want to isolate the panel from 70 and 20 on the S100 bus. Offhand I can't remember why other than it conflicts with many boards, but all my front panels have insulated tape that cover 20 and 70 (they are back to back so that makes it easy to wrap tape from pin 20 on the front to pin 70 on the back). This could be the problem why the 16FDC brings the system down.

My disk system for the IMSAI is a Morrow's DJ 2D/B and it works very well for the 8080 card, the California Computer Systems CPU (properly configured...see file at Deramp) \ as well as the ZPU.
 
You must log in or register to reply here.
Back
Top