IMS Series 8000 (5000SX?) Project

[twmiller]

I've picked up the card cage, motherboard, and some IMS cards that I hope to use to build a multi-user Z80 system that'll be pretty much the same as the first computer I ever actually wrote a program one, which was the IMS Series 8000 that my dad had at the radio station he owned when I was a kid. I'm not entirely sure what model these came from, though I'm told that they were originally part of a computer that was integrated into a desk. As far as I can tell, one of (or perhaps the only?) difference between a Series 5000 and a Series 8000 is that the former had 5.25" floppies, and the later had 8" floppies. The controller I have is from the 8000 and has the connections for 8" drives.

I've attached pictures of the cards and the motherboard. The IMS gear I have is:

• C00451 CPU card,
• C00464 64k Memory x 2
• C00442 I/O x 2
• C00401 Floppy Controller
• A862 MPU Slave Processor x 2
• C00540 12-slot motherboard.

Additionally, I have a couple of FD50to34 adapters by NF6X, which I initially plan to use with a couple of FlashFloppy mod'd Goteks. I have Mean Well power modules to push 7.5v and 15v on the bus, and also to provide 5v/12v to floppies and hard drives.

If you look closely, you can see that the caps have been pulled from the floppy controller, and also from one of the IO cards, presumably as part of a re-cap job that was never completed.

I think the first order of business is to replace those missing capacitors, particularly on the floppy controller, but also on the second IO card. Then wire up the power supplies and a reset switch to the 9-pin molex, and also add a proper power switch. Then I should be able to power it up and check voltages on the bus. Then add the CPU card, a memory card, and an IO card and see if I get anything on a terminal. After that, try to boot CP/M. Once that works, I think I should be able to tackle getting TurboDOS running on it and hook terminals up to the MPU.

I also want to put together a nice looking case for it. What I'd *really* like to do is build a replica of one of the original IMS cases, but the pictures on the Internet are not great, so I'm not sure how close I'm going to be able to get. (If anyone has a complete IMS system in its original case, I'd love to see some good pictures of it.)

Anyway, I am a complete noob when it comes to anything S-100 and/or CP/M, so feel free to (politely, I hope) let me know if I'm talking nonsense about any of this.

 

[new_castle_j]

Wow, look at the serial numbers, five digits, those are early production boards. Here's a picture of an IMS 8000 that was integrated into a desk. You are correct that the main difference between the 5000 and 8000 are the size of the floppies. Also check to see if any of your boards use bus pin 52 which is -16v from the power supply.

 

[twmiller]

Also check to see if any of your boards use bus pin 52 which is -16v from the power supply.

So the motherboard appears to have the original 9-pin connector that seems to be wired exactly as the 1983 "Series 5000 Microcomputers" manual says it should be wired. I was planning to provide all of the voltages outlined here (but undervolt it to 7.5v and 15v/-15v because I'm told the cards will work just fine like that and the regulators are a lot happier and generate less heat). I do have a dual power supply that can provide the required negative voltage.

 

[jima]

I also want to put together a nice looking case for it. What I'd *really* like to do is build a replica of one of the original IMS cases, but the pictures on the Internet are not great, so I'm not sure how close I'm going to be able to get. (If anyone has a complete IMS system in its original case, I'd love to see some good pictures of it.)

Might these help? I wasn't sure about the model number but the second picture gives a clue.

 

[twmiller]

Might these help? I wasn't sure about the model number but the second picture gives a clue.

Those pictures definitely help in terms of seeing how the inside was arranged and for giving some clues into how big the whole unit was. I'm pretty sure that's a second generation (or later?) Series 8000. It's got the 8" drives, but they've lost the woodgrain. I've seen (slightly blurry) pictures of it in a marketing pamphlet someone had scanned. The original standalone 8000 looks more similar to the desk in the first reply, with the woodgrain and everything. My dad says that when the hard drive crashed at the radio station, they flew it out to Carson City to have the IMS guys help them do some data recovery. They were upsold a new version with a metal case to replace the original wood grained case. Don't get him started on that, though. Apparently the employees were supposed to be doing a daily backup (to floppy!) and had let it slide. Forty years later and I think he's still a little bit bitter about that! lol

Capacitors have been ordered to finish the re-cap job on the floppy and IO cards! :D

 

[twmiller]

So I managed to get some power hooked up today. First I made sure that the caps on the motherboard were good. Then I hooked up the Mean Well power supply modules I had, and made sure that the ground, 8v, 16v, and -16v rails (7.5v, 15v, and -15v in my case) were sitting at the proper places on the motherboard, and in the card slots.

Before putting any cards in the motherboard, I checked out the CPU card, the memory cards, and the IO cards to make sure they were electrically sound. One of the IO cards had some open capacitors that I don't have on hand, and one of the 64k memory cards seems to have a 12v to ground issue when the memory is populated. Someone did a re-cap on that memory card and it seems like they might have been more used to soldering water pipes than electronics. I'm (relatively) confident that it's not destroyed beyond the point of no return, though. Anyway, I checked all of the caps that the voltage regulator feeding 12 volts sends and it looked good, and didn't see anything else that was obviously broken (to me, anyway).

After checking out the CPU card, the non-borked memory card, and the 'good' IO card as much as I could with continuity tests, I started adding them to the system, one at a time, and checking the voltage coming out of the regulators. Using the schematics for each card, I made sure that I was getting 5v, 12v, -12v, and -5v when / where I should be seeing it.

I was hoping at this point I could hook up a null'd serial to USB modem up to the IO card on the serial out header and see some sort of indication of life, but no joy. I'm not sure whether it needs to bootstrap a floppy or a hard drive before it'll start pushing data across the serial line, or whether I screwed something up. I was at least expecting to see an error message telling me it couldn't find a proper TurboDOS and/or CP/M disk to boot off of, but there didn't seem to be anything coming across at all. I *believe* I have the three cards (The C00451, the C00464, and the C00442) set up correctly. I went for the 'default shunt settings' for all of them.

I also tried swapping out the ROM on the other IO card when I was grasping at straws trying to make something happen. Both ROMs have a sticker that says '7297' on them. Interestingly, in the June 16th, 1980 Series 5000 / 8000 User Manual, Appendix C on the IPL has a subtitle of "16-bit checksum = 7297 Hex)" I think that means the ROM I have is *probably* the ROM from 1980. Unfortunately, I have an EMP-20 and an EMP-31, and NEITHER will read a 2708! (/me shakes fist at sky!)

Tomorrow I'll pull them all out and go over those settings again and make sure I didn't miss anything. I might also be able to get the floppy controller re-capped and see if having it populated changes anything. (I think I have all of the stuff I think I need to hook it up to the Gotek except a known good floppy image in a format the Gotek can read, which is a pretty big 'except'.) If worst comes to worst, I think I can probably steal the 5.25" drive from the AT clone and see if I can't get some sort of image on a real disk.

Anyway, I'm going to take 'proper voltages' and 'nothing popped, cracked, smoked, or caught on fire' as wins for today, though. . I think I need a port 80 card for the S-100 bus.

 

[new_castle_j]

The OEM ROMs from IMS had no output messages to the terminal, they were very simple for the generation A boards that you are working with. Later IMS boards came with a built in monitor program that you could boot to after switching a jumper on the board. There is a homebrew Monitor ROM that is out there in the wild and was written specifically for the hardware you are working on. See here: https://www.retrotechnology.com/herbs_stuff/thwaites.html#res if you want to compile it and burn it to an EPROM, you should be able to boot to a rudimentary menu prompt and check out the basic functions of your boards.

 

[twmiller]

The OEM ROMs from IMS had no output messages to the terminal, they were very simple for the generation A boards that you are working with.

Yeah, that had become my assumption, because I'm pretty sure I have everything shunted correctly and everything *seems* to be working (as near as I can tell given the setup I have). At least the power is good and some rudimentary poking around the CPU with a scope shows that it's alive and kicking.

There is a homebrew Monitor ROM that is out there in the wild and was written specifically for the hardware you are working on.

I had found that page you linked earlier and had gone over it a couple of times. It looks like the card set he has/had is pretty much identical to mine, except that has the 5" C00431 floppy controller instead of the C00401 8" controller that I have. I have that second 442 card, so my plan is to compile and burn that Monitor ROM and stick it on the second card, then just move between that and the IPL-ALL ROM on the first IO card by using the ROM enable/disable jumpers on the two cards. Right now I'm just waiting for some 2732s that I ordered to show up.

The good caps for the floppy controller and the second IO card are out for delivery this morning, so I should be able to get some work done on the floppy controller. I think grabbing the 5.25 from the AT and using that for step one is probably going to be easier than trying to coax the Gotek into working with it. (That is to say: my inclination is to get a known working 5.25" drive working with the FD50to43 convertor before I try to get the Gotek working with it, so I don't have to debug Gotek config issues on top of figuring out the floppy controller issues.)

I need to start tracking down some diskette images for all of this, too.

 

[twmiller]

Okay, so I didn't get *quite* as much done as I was expecting to get done over the long weekend, but I *did* manage to get the second IO card and the C00401 floppy controller re-capped. I pulled a 5.25" drive out of the 386, and hooked it up to the FD50to34 card, and then used a 50-pin SCSI cable to hook the 50-pin end of the fd50to24 up to the controller. I didn't see any signs of trying to read the drive, though. The drive's getting good juice, because the motor spins on power up, but there doesn't seem to be any attempt by the IMS system to read the floppy disk. Hitting the reset line doesn't seem to coax any signs of life out of the drive, either.

I *think* I'm probably dead in the water until I get a known EPROM in the C00442, because right now I'm kind of flying blind. LEDs and speakers hadn't been invented when this computer was made, so there are no beep or flashing lights to tell me what's going on. :D Anyway, all of the cards seem to be good from a power perspective except one of the 64k RAM cards, which as I mentioned earlier has 12v going to ground when the chips are populated. I have a parts for a chip tester on order, as well as a bunch of 2732s and a burner than can handle 2708s (so I can read the existing IPL ROMs).

 

[twmiller]

So related to this project, I pulled out the two ST-251-1 drives I have that I can use with this project and stuck them in the Tandon 286 and did a low level format and surface scan on them to make sure that they were still working. Both drives are in remarkably good condition for their age and showed no signs of impending failure. Speedstor showed one drive had 5 defects, and the other had 14 defects after a low level format and scan. Not bad in the grand scheme of things. I'll probably check them out a little bit more tomorrow, but at this point, I'm pretty sure that one or both of them will be good options for a future TurboDOS install.

 

[new_castle_j]

Yes, those drives should work, just be aware that the way IMS wrote their formatter and drivers for HDDs with TurboDOS was not fully customizable, they supported a number of drives that you will be prompted to select from on a list, If your drive isn't listed you can still proceed by choosing one from their list that has an equal or lesser CHS value.

 

[twmiller]

Yes, those drives should work, just be aware that the way IMS wrote their formatter and drivers for HDDs with TurboDOS was not fully customizable, they supported a number of drives that you will be prompted to select from on a list, If your drive isn't listed you can still proceed by choosing one from their list that has an equal or lesser CHS value.

Do you know if there's a list of the supported drives, by any chance? I can rummage through some boxes and see if I have any other MFM drives around. (I'm pretty sure I don't have others beyond the 10 MB Conner that's in the PC XT, but there are a lot of boxes around here filled with misc. computer parts, so who knows?). It looks like the 2732s I ordered off of eBay have disappeared between there and here. They were last spotted about a week ago. I'm waiting on those so I can burn the IPL-ALL ROM.

 

[new_castle_j]

Here's the list, but there may be others depending on the version of WINFMT that you use, some could have been added later than the documentation I have.

 

[twmiller]

Okay, so after spending a few weeks waiting for parts (that is to say, 2732A chips) and taking last week off for a spring break vacation with the family, I'm (slowly) back on this.

I have modified the two 442 I/O boards to accept 2732 chips in addition to the 2708 chips they came with (and also 2716) chips. I've burned the IPLALL.BIN to a 2732A, but didn't see any floppy activity, and nothing on the serial port (though I'm not sure whether I should see anything on the serial port with that particular ROM). I managed to build minimon9.com correctly (I *think*) and burned that to an EPROM, too, but also didn't see anything over the serial port when that was installed. I'm using a minimal card loadout with just the 451 CPU card, the 464 memory card, and the 442 I/O card. I have all of the boards jumped to their default configuration.

Since I have two 442 IO cards and two memory cards, I've also tried multiple combinations to see if maybe there was a bad card in the mix, but I've seen nothing at all. I haven't done much in the way of debugging the CPU board, but at one point I did stick a scope on it and verified that I was getting a good clock.

It's time to step back for a second and ponder the next move.

 

[new_castle_j]

The IPLALL ROM won't output anything to the serial ports upon boot, but it does poll the floppy drives. I've never built the minimon ROM before, but if you post it, I can try it in my system and verify that it built correctly.

 

[twmiller]

@new_castle_j here are the two compiled files if you want to mess with them: Minimonx.zip. I tried both just burning the minimonx.com file to an eprom, and then I did a 'copy /b monlift.com + minimonx.com minimon9.com' to create a combined file for the high memory which I burned to a second eprom. neither seemed to work. I compiled both of them using the compiler referenced in the readme on an Agon Light, which is the only Z80 I have. (Technically, I guess the Commodore 128 has a Z80 in it, but...)

I've been going through the Debugging for beginners page at s100computers.com and discovered that the reset pin on the Z80 was not coming up high. after a bit of poking around with a logic probe, I replaced a 27LS240N and now have high signals for wait, int, imi, reset, and busreq. now that the CPU card is (I think?) working, I'm seeing a 'parity error' light on both of the 464 memory cards. I'm not *entirely* convinced that the CPU is working correctly, though, and this comment on Richard Thwaite's adventures with his Series 5000 reinforces this feeling: "The parity LED lightup was, in the end, a red herring." It's possible that both 464 boards have bad memory, I guess, but I suspect there's another bad buffer chip somewhere.

 

[new_castle_j]

I wouldn't worry about the parity error light at this time, on my system the parity error light will illuminate during power up and remain until some code in the ROM resets the parity error circuit.

 

[new_castle_j]

I burned a 2732 with minimonx from your download and put it into my system, it does nothing. I even re-compiled it from the source and got the same binary that you did.

 

[twmiller]

I burned a 2732 with minimonx from your download and put it into my system, it does nothing. I even re-compiled it from the source and got the same binary that you did.

Out of curiosity, did you try running it (minimonx.com) from the command line (assuming you used your IMS system to build it)? The readme seems to indicate that it can be run that way, too. I tried it on the Agon and it worked about as well as you'd expect it to (i.e. it just printed some gibberish on the screen), but I don't think that's too surprising because the Agon is a substantially different hardware environment than the IMS system the monitor is supposed to work on.

I burned a fresh copy of your IPLALL ROM to a 2732A. I tried multiple combinations of my IO and memory cards and a) never saw the memory parity error led go out and b) never saw any activity from the floppy drive. I think there's something happening that's preventing the ROM from loading.

 

[new_castle_j]

I didn't try running as a .com file, I compiled it on the altair emulator from my laptop. Now that you mention that it can run as a .com file though, that would mean it would have to be ORG'd at Hex 0100. So I looked at the code again and sure enough, line 146 shows an ORG statement of 0100, and line 150 shows an ORG statement of C600. If we are going to run this program from ROM, it would need to be ORG'd at Hex 000 in the IMS machine. So I changed line 150 to read "MONB EQU 000H ; location for EPROM" and re-compiled, then burned into a 2732, when I powered up, I then saw the Monitor prompt on the screen, typing "M" brings up the menu of options. You can download my compiled version below or try the modification to the code yourself.