S-100 Revival After 30+ Years

[jgrant911]

Greetings,

Just joined the forum today!

I'm in the process of reviving my S-100 system, originally put together in 1981, worked on and experimented with until IBM PCs took over in the late 80's and in storage until this week.

I vaguely remember that the event that caused me to abandon it was a disk drive power supply failure that I repaired but then I couldn't solve related problems (didn't have time). I have always planned to see if I could resurrect it and have finally gotten around to it.

Progress was actually fairly fast after I solved the problem of no longer having my trusty ADM3a CRT terminal!. I got my laptop USB to serial connection to work after discovering a burnt trace on the motherboard -12V buss. With just the CPU and RAM cards installed I got the ROM Monitor to respond. Next I installed the Floppy Controller and got the Disk Monitor ROM on that board to respond. Final step, connect the Floppy Drives and try to boot. Amazingly, after a couple of glitches with baud rate mismatches I had CP/M booted on Drive A!

So that went about as well as anyone could expect on a 40 year old batch of electronics. However I do have a problem with Drive B. If I select B: from the CP/M command line it selects drive B, the read head seeks track 0 but the system hangs there and I have to reset the system to regain control.

I will be nosing around the forum for clues to solve this problem so if anyone has any ideas, please toss them my way!

Cheers!

Jim

 

[daver2]

Welcome to VCFED!

What system have you got there? Especially the disk controller and disk drives to give us a clue where to suggest you start...

I would make a stab at looking at the track 0 signal from the drive and then check to see if you have an index pulse visible as my first port of call.

Dave

 

[jgrant911]

From my Profile:

Vintage computer is California Computer Systems (CCS) 2200 mainframe with:

- 2501 12 slot Motherboard
- 2810 Z-80 CPU Board
- ExpandoRAM II 64K Memory Board
- 2422 Floppy Disk Controller
- 2710 4-Port Serial I/O Board
- EDA TMS9918 Video/Sound Board
- Home Brew MM58167 Real Time Clock Board
- Dual Shugart 801 SSDD Floppy Disk Storage

Jim

 

[jgrant911]

OK, I put a scope on Track 00 detect: good. Then checked for the Index signal: 1 pulse every 167ms which is exactly 360 rpm. Finally I looked at the Read Data signal which shows a stream of 0.2 microsec pulses at 2 or 3 microsec spacing.

So, at least the basics are good. I have the Shugart manuals and the basic disk access code listing from the Disk Controller ROM but I'm still looking for the sequence of events that would cause an infinite wait for a response from the drive.

Jim

 

[cruff]

Double check that the drive belt on drive B isn't stretched just enough that loading the heads causes the media to stall. I had that happen on an 8" drive in an HP 9895A, while the other drive was fine.

 

[jgrant911]

Nope, the heads load OK and the media index timing pulse stream doesn't change.

The obvious symptom here is that when selecting drive B: (Typing "B: CR" at the A: prompt) even though the disk select LED lights, the heads load and the stepper seeks track 0, the system hangs forever. I would have thought that there was an Error Flag or some sort of timeout in the logic?

I am assuming that the code that manages this process is part of the ROM resident basic disk I/O but Z-80 assembly code is not my forte so I'm still struggling thru it to figure out what the Disk Controller could be waiting for that it's not receiving.

Jim

 

[daver2]

I found your profile details. We are all learning with a new revision of the board - and profiles are not even on my list to sort out yet...

Your disk controller contains a WD1793 - so it should have some form of timeout.

Since you have two (2) disk drives - and one works and the other doesn't - I would be inclined to find out whether the problem is related to one of the disk drives in particular (drive B) or when using both of them together.

I would disconnect the drive that works and configure the non-working drive to be the same as the working drive. However, I would note down EXACTLY what links are being changed. Try the non-working disk drive when configured as A: and see what happens. If it doesn't work then - we are looking at a drive fault of some description. If it works then - we are probably looking at why the two drives don't like playing together with the controller.

The WD1793 looks at the index pulses for timeout (if I remember correctly). Hence the interest in them...

Just looking at the data sheet. Yes, 5 index pulses can abort most commands. However, there are some commands where the controller can 'get stuck' - but I would rule these out initially if one drive is working.

Dave

 

[jgrant911]

Thanks for the feedback Dave. Yes, I was thinking along the same lines (swapping drives). I was just trying to talk myself into clearing enough room on my work table to be able to access the backplane circuit boards of both drives at the same time. That's my project for this afternoon. : )

To be entirely honest, I am amazed that I have made as much progress as this in just a few days. My recollection from 30 years back was that I had a serious power supply failure in the floppy drive enclosure that involve a power transistor failing in a way that allowed high voltage through to the drive circuits. Obviously it wasn't as bad as I thought because so far there has been no obvious damage except for the burned trace on the -12V rail on the S-100 backplane (which may have been completely unrelated).

It's great to see that there is still a community of enthusiasts enjoying this old tech.

Cheers!

Jim

 

[jgrant911]

OK, problem solved (or more correctly No Problem!).

In the 30+ years since I have gazed upon this miracle of 1980's technology, my organic FIFO stack had eliminated any knowledge or awareness of soft sector vs. hard sector disks. It turns out that I have a mixture of these types of floppies in my disk collection. Also, I had conveniently forgotten what the little toggle switch on the front of each of my Shugart drives was for...

It turns out that I used to be smarter than am now and when I bought my drives I purchased them with the optional front panel switch that selects Shugart 800 vs. 801 mode: soft sector or hard sector disks! I had some recollection that these switches should normally be in the right hand position, which happens to be for hard sectored disks. This worked fine for my CP/M boot disk but by chance every disk I put in the B: drive was soft sectored causing the "it does not compute" error!

Now that I have basic system functionality I will begin re-installing my optional boards (4 port serial board, colour video graphics board and real time clock with programmed shutdown and restart)

Pictures of my now working system and the confusion causing toggle switch. I think I'll put a label on these!

Cheers!

Jim

 

[daver2]

>>> I think I'll put a label on these!

It sounds like a plan to me !

Good to see you are on the road to getting it running. Oh yes, we enjoy vintage tech. here!

I had a similar problem with my Cromemco system with the disk drive regulators. Fortunately, I checked them out without having them connected to the floppy drive - so the shorted regulator didn't do any damage. Sometimes circuitry survives - sometimes it doesn't...

Let us know if you need anymore help. It will be good to see the results when it is finally working.

Dave

 

[MikeS]

Apologies for the OT intrusion but I'm curious: Dave, which Cromemco(s) have you got?

 

[daver2]

A Cromemco Z-2D (the big black box). I am still looking for a System-3 though (with the 8” disks).

Dave

 

[jgrant911]

Just a follow up on the floppy drives.

First item:

I have both drives working, but with intermittent problems. I thought I had things going quite well after my initial glitch (above) but both drives are now giving me occasional BDOS Errors (B: drive quite consistently now).

I am working from about five 8" Control Data boot floppies and a dozen or so miscellaneous file floppies that I created about 40 years ago so I am not completely surprised to have read/write problems. It was actually during the process of trying to create new boot disks on un-used (but still very old) floppies that the BDOS errors became much worse.

I am hoping for some insight from the forum as to possible causes for the increased errors. My guesses so far include:
- weak or failing drive or controller electronics (read/write degradation, track/sector sense issues...)
- dirty heads (maybe aggravated by aging disk surfaces)
- ???

Second item:

I am obviously concerned that if my 8" drives fail and all I have is 8" media, I will have a problem. For this reason I am looking for the easiest way to use the CCS 2422 Multimode Floppy Disk Controller to connect to some alternate drive (5.25", 3.5" or ???) that I can buy new media for.

I have seen several mentions of using other drives on the 2422 controller, which was designed for both 8" (50 pin header ) and 5.25 " (34 pin header) drives. I haven't found anything that amounts to a DIY on this though.

I have a 5.25 floppy drive from a IBM clone but from what I have found in my research is there are some differences in the cable pin assignments:



If anyone has any pointers to share (or direct me to) that might help me to migrate to different drives, please let me know!

Cheers!

Jim

 

[DeltaDon]

I believe, not an expert but from my own experiences, you need to find 720K 5.25" drives (Or quad density - 1Mb unformatted) and not the later PC 1.2MB drives. That is unless you modify the PC drives to run at 300 RPM. I went through this with a Ampro Little Board Z-80 system recently. There's some NOS Teac FD55 drives being offered on eBay if the price isn't a problem.

 

[MikeS]

A Cromemco Z-2D (the big black box). I am still looking for a System-3 though (with the 8” disks).

Dave

A shame you're on the wrong side of the big pond; I've got a CS-3 (type II) that I'd love to get rid of and will quite possibly scrap

 

[daver2]

A shame you're on the wrong side of the big pond; I've got a CS-3 (type II) that I'd love to get rid of and will quite possibly scrap

Mike,

Can I ask around before you do anything hasty (like scrap it) :-(...

I will see if I can find a 'solution'...

Dave

 

[MikeS]

Well, it hasn't moved in 20+ years so I guess there's no rush. This is the type II which has a single Tandon TM-848 and a hard disk, as opposed to the original with one or two Persci drives. There's also a later CS-300 and several CS-1/1Hs awaiting their fate...

Keep in mind I'm only talking about scrapping the chassis; I'm hoping to get a few $$s for the cards and drive. I've also still got a few Z-2 parts left over from systems I sold a few years ago if you need anything that I happen to still have.

I'll PM you; let's take it off-list instead of appropriating this thread.

 

[jgrant911]

The saga continues!

My two SA801 drives have gradually become unable to read data from my disks. The current status is that my CCS computer is only able to run the MOSS Monitor (basic utilities). This means the most I can do with the floppy drives is basic Seek and Read tests.

It looks like my CCS 2244 Disk Controller is doing all the right things. If I tell it to seek and read any track on either drive it will seek the correct track but not read any data (10 tries then back to the command line with:

DSK ERR: U 00 T 00 S01 C 88 E 10 when trying to read Unit 0, Track 0, Sector 1 and
DSK ERR: U 00 T 10 S01 C 1E E 32 when trying to read Unit 0 any Track other than 0, Sector 1.

I haven't made any sense of these error messages (which are consistent over many tests). They are supposed to give the last contents of the WD1793's Command and Status registers. I would have thought that issuing 2 READ commands, only differing by track number, would have relatively consistent error returns but I can't make sense of these results when checked against the WD1793 data sheet.

More interesting, in scoping the various connector pins I found that there is a continuous random bit stream out of the READ DATA line from both of the drives even when there is no floppy in the drive and no command has been issued to access the drive (disconnected the 50 pin cable from each drive and tested individually).

The READ DATA signal comes via an enable gate (disabled only when WRITE active) directly from the Read Amp chip, Shugart 16278-0. I am wondering if it is likely that something age related is happening to the READ circuitry causing this problem.

Looking at the partial schematic for this circuit below, the READ AMP is in the lower right corner. The output pin (10) is a steady stream of random data. Measuring the primary inputs to the chip on one of the drives:

- DIFF IN on pins 12 and 13 (~250 mV on TP1 and 120mV noise signal on TP2)
- DIFF COMP on pins 14 and 15(~200mV noise signal on each input)

I won't pretend to know what exactly is happening inside this chip but for some reason the inputs it is getting are causing random bit output on READ DATA. The only adjustment that I see is the 10K trim pot on the DIFF COMP circuit.

If anyone on the forum has any input that might help solve this mystery I would love to hear from you.

Cheers!

Jim

 

[Hugo Holden]

As a test, if it is easy, check the head for continuity. In theory at least the differential amplifier should only amplify a voltage difference between its two inputs and reject common mode noise.

If the head continuity (its internal coil) is ok, there could be a stray field inducing voltage in the coil. So you could try shorting the head out to see if the random pulses go away.

There was an interesting thread a while back about a disk drive where the head amplifier had lost both sensitivity and had increased noise pickup. It was due to one of the pre-amp IC's inputs failing, this causes the gain to drop in half and the common mode noise rejection to fail.

Otherwise if all is well with the actual head and preamp section of IC 4D, there could be a defect in the remaining part of IC 4D, but before blaming the IC, it would require the IC's power supplies were clean & checked and the components around the IC checked.

One suspicious feature is that the noise you measured on TP 1 & 2 was not identical, it should be more or less equal & opposite.(you said TP1 & 2 were diff comp pin 12 &13, but TP1 & 2 feed diff in pins14 & 15)

I once had a disk drive that produced constant read data(random bit stream), it stopped the other drive it was being used with from working on the same shared cable. But the faulty drive doing this worked ok when it was used. So it was a case of a fault in one drive, inhibiting another, making it look like the other drive was faulty, but the fault didn't affect the actual drive it was in. It had me running around in circles trying to figure out what was going on. It turned out to be a faulty 74LS logic IC on the drive board. I just found the old thread I started about this:

https://www.vcfed.org/forum/forum/te...ish-disk-drive

 

[jgrant911]

Thanks Hugo,

I have removed the power supply, an old Power One Disk Drive Series CP 206, that was part of the dual 8" drive package (purchased from Jade Computer for $1015 on Dec 26, 1980. I still have all of my original receipts!).

Depending on what I've found once I am finished checking out the power supply I will go through the read head circuitry per your suggestion to see what's there. I will also check my measurements to see where I noted the wrong pins.

BTW I also have an old YD-380B that I have been playing with but that's on hold until I can boot up CP/M again. ; (

Thanks for the input.

Jim