• Please review our updated Terms and Rules here

Repairing a Commodore 2040 IEEE-488 drive

Hm, interesting thought that a DOS 1 board might not report anything meaningful on startup. VICE confirms this as PRINT DS$ with 2040 emulation results in a load of @ signs at the top of the screen. Perhaps I should try to access floppy disks until I find a working combo, or burn my own DOS 2.x chips. 2532 should be fine, right?

I am working on the document labelled "2040 SERIES SERVICE MANUAL" but I must disappoint you it is very non-technical. At the very least, you need a special floppy disk with diagnostic software (that I don't have - perhaps it is on Zimmers or elsewhere?) and *drum roll* a mostly working 2040 drive. You can load the diagnostic software from tape though, if your drive doesn't work.

BINGO!
Previously I had only tried to read the error channel at boot-up and then access DOS 2.x floppies, two operations doomed to fail on a DOS 1 floppy drive. Thus I assumed them all to be mostly broken. Now I accessed an unknown floppy, managed to PRINT DS$ to which it reported READ ERROR and then reformat it as DOS 1. The drive mech sounds like a 19th century submarine, but it appears to work.

Thus I will assemble drives as I have parts to. In order to upgrade to DOS 2.x, I'm afraid I need different 6530 RIOT chips so it might not be as easy as replacing ROM with EPROM?
 
Last edited:
I am working on the document labelled "2040 SERIES SERVICE MANUAL" but I must disappoint you it is very non-technical. At the very least, you need a special floppy disk with diagnostic software (that I don't have - perhaps it is on Zimmers or elsewhere?) and *drum roll* a mostly working 2040 drive. You can load the diagnostic software from tape though, if your drive doesn't work.
Anders,
I have the 4040 diskette that came with the drives, but alas the Performance program is corrupt. I have made two 4040 performance tests (one for drive 0 and one drive 1) by modifing a 1541 performance program if you want them.

You are right. I'd better copy those tests to cassette in case the drive starts faultering and can not read the diagnostic floppy. Good point!

In order to upgrade to DOS 2.x, I'm afraid I need different 6530 RIOT chips so it might not be as easy as replacing ROM with EPROM?

Oh, you are right :(
The UK3 RIOT chip in my box has been upgraded to a 901466-04. I did not notice that before. You'll need to find someone with a broken 4040 that can give you the chip as it is mask programmed.

It would be possible to inhibit the 1K ROM internal to the 6530 and use a small EPROM with the -04 code, but that would be quite a kludge and look awful.
-Dave
 
I asked on the mailing list, and Martin H-V pointed me to the daughter board found in the 8250LP, SFD-1001, 8296-D and alike. I still own a working 8250LP so I could look inside and compare with the schematics. As far as I am concerned, kludges are not a problem as long as they improve functionality, everything is socketed and possible to restore to original condition. Never mind these were drives I was about to toss due to failed troubleshooting.
 
Please post the 2040 manual I have been searching for it.
Here is an early draft. So far I looked through two copies, but both are missing appendices C - Component Cross Reference and D - SA390 Parts List. Those might also be floating on the 'net, but I should have a look through the other four copies as well. I also found some schematics, but I better check if those already were scanned.

http://www.cbm.sfks.se/files/cbm2040-servicemanual.pdf
 
Here is an early draft. So far I looked through two copies, but both are missing appendices C - Component Cross Reference and D - SA390 Parts List. Those might also be floating on the 'net, but I should have a look through the other four copies as well. I also found some schematics, but I better check if those already were scanned.

http://www.cbm.sfks.se/files/cbm2040-servicemanual.pdf

Anders,
Thanks so much. There are very good things in this manual.

Zimmers has the SA390 parts list at:
http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/shugart-1.gif

The 2040/4040 parts list is at:
http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/parts.txt

The schematics are at:
http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/index.html

-Dave
 
Aha! I found like 5-6 copies of the 2040 Service Manual and another 2-3 copies of 8050 service manuals. :shock:

Time to start scanning, as I see these are neither on Zimmers, Bombjack nor Commodore.ca, the three places I usually look for these kind of docs.

Anders,
When you get time, you should upload the 2040 manual to Zimmers at:
ftp.zimmers.net/incoming

And maybe send Bo an email and suggest that he place the manual at either
http://zimmers.net/anonftp/pub/cbm/documents/disk-drives/index.html
although so far this directory only has 1541 info.

or at:
http://zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/index.html
which has the most 2040/3040/4040 stuff.
 
Good news from my horizon: After realizing these drives have DOS version 1, I managed to put together four mostly working units. The fifth one has a broken power supply, floppy drive mechs out of alignment and an untested digital logic board missing 6502 and 6522 so that is the "spare parts" unit from which I'll fetch levers for gsteemo. Well, perhaps he'll be interested in even more parts, at least the digital and analog logic boards.

On the mailing list, I got advice how to make or possibly obtain daughter boards consisting of a 6530 RIOT and 2716 EPROM in order to upgrade the working drives with DOS version 2 (plus actual ROM EPROMs of course). That might be a project for the near future, although I managed to read 1541 formatted disks in these drives without a DOS upgrade.
 
...so that is the "spare parts" unit from which I'll fetch levers for gsteemo.

Anders, I could use a plastic latch also. On my drive 0, I have to hold the broken latch with my thumb when I read from that drive.

On the mailing list, I got advice how to make or possibly obtain daughter boards consisting of a 6530 RIOT and 2716 EPROM in order to upgrade the working drives with DOS version 2 (plus actual ROM EPROMs of course).

This would be a perfect solution, a small printed circuit board, instead of the kludge with 40 wires that I was thinking of. Is this a Commodore board or did someone make them?
-Dave
 
Perhaps I was unclear above, but instead of producing new custom 6530 chips for the 8250LP, SFD-1001 and 8296-D (not sure about the full height 8250), Commodore installed this kind of daughter board into the drives which supposedly would allow use of an older 6530 combined with an EPROM with the latest controller code. So yes, the existing spare boards would be CBM stuff, given that they work at all with a 35 track drive.
 
Perhaps I was unclear above, but instead of producing new custom 6530
chips for the 8250LP, SFD-1001 and 8296-D (not sure about the full height 8250), Commodore installed
this kind of daughter board into the drives which supposedly would allow use of an older 6530
combined with an EPROM with the latest controller code. So yes, the existing spare boards would be
CBM stuff, given that they work at all with a 35 track drive.

Anders,
Yes, the daughterboard should work just fine to turn your 2040 into a 4040. It does not matter what is programmed into the 6530 RIOT chip. The 1K ROM is bypassed and as long as you put the 901466-04 code in the companion EPROM (2716?), it will support DOS 2. Existing code for the 8250, etc that may be in the daughter board EPROM must be replaced with 901466-04 code.

But don't forget to check the part number on the RIOT chip in the daughter board as you may get lucky and find it is a 901466-04 (6530-034) chip. Then you are done; simply put it into the main board and forget about the daughter board.

As you know, you also need to upgrade the following ROM slots with 2532 EPROMS for DOS 2.1:
UJ1 901468-14
UL1 901468-15
UH1 901468-16
 
Last edited:
As you know, you also need to upgrade the following ROM slots with 2532 EPROMS for DOS 2.1:
UJ1 901468-14
UL1 901468-15
UH1 901468-16

Anders,
I am willing to swap you a set of three 2532s programmed with DOS 2.1 for one of the plastic levers from your 'parts' drive.

gsteemso will need the other one if he ever fixes his 2040 drive. As I recall he was going to replace his RAM.
-Dave

-Dave
 
No worries, I have every intention of carrying out that repair. I will be overjoyed to acquire a latch from Anders if it turns out the mechanism my broken one is attached to still works, but I've had to put the whole business on hold for a while due to a death in the family and the concomitant massive hole in our budget. I am following the discussion here quite avidly, but have not had anything to add.
 
You have our condolences, gsteemso. Of course your family comes first.

One good thing about old computers, they have been around for so long another month or two does not matter.
-Dave
 
Or year for that matter, as long as the chips don't break down.

Dave, I'll return to your offer shortly. However I was told and confirmed through own research that a 2532 EPROM might not be a slot-in replacement for 2332 ROM. They are similar but a few patches as regarding to /OE and /CS lines need to be done. Still, I only have a few 2532's left and might find use of them in some other application.
 
However I was told and confirmed through own research that a 2532 EPROM might not be a slot-in replacement for 2332 ROM. They are similar but a few patches as regarding to /OE and /CS lines need to be done. Still, I only have a few 2532's left and might find use of them in some other application.
Anders,
Good point. I'll check the pinouts for the 2332 also. Perhaps they are the same as the 2732??
-Dave
 
Anders, a quick look indicates that the 2332 and the 2532 are compatible. However, it looks that the chip is only enabled during the Phase 1 clock (not for the full 1 uS period), so that means the fastest access time 2532s are needed; at least a -35 or -30. I think I have that type.
-Dave
 
Last edited:
Anders, a quick look indicates that the 2332 and the 2532 are compatible.
Anders,
OK, I burned a set of 2532 EPROMs for DOS 2 and replaced the 2332 ROMs in my 4040 drive and they did not work.

The problem may be pin 21. On the 2332 ROM, it is a positive going chip enable while on the 2532, it is Vpp and should normally be wired to +5V during read operations. So a true level signal may not be good enough for it to work if it must source a little current.

On the early PETs pin 21 is tied to +5V and on the later PETs, pin 21 is tied to +5V through a 100 ohm resistor. Both of these configurations work, but apparently a TTL true on pin 21 of the 4040 design is not good enough for the 2532. I will try to lift pin 21 and tie it to +5V and see if that cures the problem. I'll let you know.
-Dave
 
Last edited:
Anders,
OK, I burned a set of 2532 EPROMs for DOS 2 and replaced the 2332 ROMs in my 4040 drive and they did not work.

The problem may be pin 21. On the 2332 ROM, it is a positive going chip enable while on the 2532, it is Vpp and should normally be wired to +5V during read operations. So a true level signal may not be good enough for it to work.

On the early PETs pin 21 is tied to +5V and on the later PETs, pin 21 is tied to +5V through a 100 ohm resistor. Both of these configurations work, but apparently a TTL true on pin 21 of the 4040 design is not good enough for the 2532. I will try to lift pin 21 and tie it to +5V and see if that cures the problem. I'll let you know.
-Dave
Probably not the issue here, but have you tried to read the 2332 as a 2532 and checked that pin 21 is really +true? Although they usually were positive true, AFAIK the 2332s could be programmed either way.

FYI, have a look at the schematic of the old 2008 with 2316 ROMs to see how it worked. Each pair is completely in parallel with pin 18 (CE2) programmed as CE on one chip and /CE on the other; makes chip selecting a little easier, effectively combining two chips into one.

http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320081-2.gif
 
Back
Top