• Please review our updated Terms and Rules here

IMSAI 8080 at the Rhode Island Computer Museum

We applied some Kapton tape to the gold fingers for 20 & 70 on the Cromemco 16FDC floppy controller. These signals are memory protect on the front panel and ground on the 16FDC so it inhibits the EXAMINE and DEPOSIT functions on the front panel. After the Kapton tape was applied the front panel worked normally. We can't see the contents of the RDOS boot ROM on the 16FDC. The chip is being selected, but it looks like the bus buffer chip for the data bus is not being enabled. I need to learn a little more about the S-100 bus signaling and then do some more debugging.
 
After a long break working on other systems we are back working on the IMSAI. The Cromemco ZPU and 64k Memory Merchants static RAM board seem to be working well. The Cromemco Bytesaver is really broken, but should not be difficult to fix.

We installed a Cromemco TUART board that will be the system console. We need to test it, and make a DB25 to DB9 cable so we can connect it to a laptop.

We have two Cromemco 16FDC floppy diskette controllers, and an external chassis that holds 2x 8" floppy drives. We have an 8" Cromemco RDOS diskette, but we have no idea what hardware configuration it expects.

As mentioned by others we need to isolate pins 70 and 20 on the S100 bus because we have a mix of old and new front panel, backplane, and boards.
 
After a long break working on other systems we are back working on the IMSAI. The Cromemco ZPU and 64k Memory Merchants static RAM board seem to be working well. The Cromemco Bytesaver is really broken, but should not be difficult to fix.

We installed a Cromemco TUART board that will be the system console. We need to test it, and make a DB25 to DB9 cable so we can connect it to a laptop.

We have two Cromemco 16FDC floppy diskette controllers, and an external chassis that holds 2x 8" floppy drives. We have an 8" Cromemco RDOS diskette, but we have no idea what hardware configuration it expects.

As mentioned by others we need to isolate pins 70 and 20 on the S100 bus because we have a mix of old and new front panel, backplane, and boards.
Before proceeding further with this project, I recommend you read this discussion in the Google Groups Cromemco discussion forum on the topic of using 8" drives other than the 8" Persci drives. The 16FDC was originally designed for use with Persci 8" drives, which have an interface that is not identical to the Shugart interface found on most other 8" drive models.

Another detailed article describing how to use Shugart 800 series drives with the 16FDC can be found here:

The system console was usually connected to the serial port on the 16FDC to use RDOS. RDOS is programmed into the ROM on the 16FDC, and provides both a floppy disk boot loader and a basic system monitor. If you are using the 16FDC board in a system with a 64K RAM board you need to disable RAM from C000 to D000 which is where RDOS lives as explained on page 7 of the 16FDC manual. Your Memory Merchant 64K RAM was designed for compatibility with a Cromemco system, and provides a method to configure this hole into its memory space. You should be able to boot and run Cromemco CDOS from floppy disk with just a ZPU, 64K RAM, and one of Cromemco's floppy disk controller boards - but for now I suggest you concentrate on getting the ZPU to run RDOS and communicate with the console on the 16FDC board. After that has been accomplished you can move on to getting a floppy drive working with the 16FDC so that you can boot CDOS.
 
Last edited:
hmb,

Thanks for all of the details. One of the 16FDC boards came with two 50-pin headers that were wire-wrapped together. Many of the signals were disconnected and many were moved from one pin to another. It looks like they had a short 50-pin cable from the controller to the headers, and then another 50-pin cable to the drives. Maybe the 50-pin header jumpering does some of the modifications described in Herb's documents?

So I guess our new plan is to remove the TUART and connect a console to the 16FDC and see if we can get the ZPU to boot from the EPROM on the 16FDC. The broken Bytesaver has an EPROM labeled ZM 1.4, which I assume is the Cromemco Z-80 Monitor. I haven't found a manual for it yet, so I don't know what it expects for a console device.

The Memory Merchants board has the high 4k of memory disabled to leave room for the boot EPROMs.
 
hmb,

Thanks for all of the details. One of the 16FDC boards came with two 50-pin headers that were wire-wrapped together. Many of the signals were disconnected and many were moved from one pin to another. It looks like they had a short 50-pin cable from the controller to the headers, and then another 50-pin cable to the drives. Maybe the 50-pin header jumpering does some of the modifications described in Herb's documents?

So I guess our new plan is to remove the TUART and connect a console to the 16FDC and see if we can get the ZPU to boot from the EPROM on the 16FDC. The broken Bytesaver has an EPROM labeled ZM 1.4, which I assume is the Cromemco Z-80 Monitor. I haven't found a manual for it yet, so I don't know what it expects for a console device.

The Memory Merchants board has the high 4k of memory disabled to leave room for the boot EPROMs.
Cromemco ZM 1.4 is a very bare-bones monitor - you can find the source code and manual here:
 
Last edited:
The RICM has a Cromemco System 3 with Persci 270 diskette drives. We should inventory the boards that are in that system and the spares that came with it. I think that the Z80 CPU board in this machine was replaced with a 68000 CPU board and it ran CROMIX. That system doesn't have a blinky light front panel so we through that the IMSAI and Altair systems would be easier to debug and get running.
 
After putting some Mylar tape on pins 20 & 70 on the Cromemco 16FDC Floppy Controller, the rest of the boards, Cromemco ZPU, the Memory Merchants 64k (upper 4k disabled) Static RAM, and IMSAI front panel are all working OK.

We can't read data from the RDOS ROM on the 16FDC board at 0xC000. We can see ROM data on the D side of IC51, the 74LS373 latch, but not on the Q side. That means that the 16FDC never drives the S100 data bus. The signals on pins 1 and 11 of IC51 look OK. We swapped IC51 with a 74LS737 borrowed from the second 16FDC and it didn't make any change. We will look for some newer 74LS373 parts this week.
 
We noticed that the two 16FDC boards that we have are different PCB revisions. On one IC47 is a 74LS30 and on the other it is a 74LS133. The one with the 74LS30 matches the schematics that we have found. The one with the 74LS133 does not match any schematics that Google found. This one also has a jumper wire from the Phantom signal to the 74LS133.

While we were debugging yesterday C53 on the -18VDC input decided to go up in smoke. It was really stinky.

16FDC Electrolytic Capacitor.JPG
 
Back
Top