• Please review our updated Terms and Rules here

Repairing a Commodore 2040 IEEE-488 drive

gsteemso

Experienced Member
Joined
Sep 28, 2008
Messages
80
Location
north of Seattle, USA
I recently acquired one of these units, and was not at all shocked to get the famous "blinking LEDs" error advisory when I turned it on. Sources online seem divided about the precise interpretation I should attach to the code thus reported (7 flashes, pause, repeat), though the general consensus is the fault probably lies in the second common RAM range. (I have seen conflicting reports indicating that the 2040 may or may not share a board design with the 4040. Did they use the same chip identifiers between the two boards? UD5 in one drive is not necessarily the same thing in a different model.) Alas, the only chip (and I checked them all) that seemed to get notably hot was a perfectly ordinary 7404 with no chip identifier printed next to it. It was right below what looks like an oscillator can, so for all I know, it's supposed to do that as part of the clock generation.

Any ideas?

I'd also be very interested in hearing any techniques people might have for cleaning the grime out of the actual drive units. With the computer circuits not yet working I can't really test the old full-height floppy mechs (which use late-1970s technology that is several years more primitive than anything else in my collection and thus is not compatible with anything else I have available), but I observe a fairly consistent layer of what looks like heavy, mysteriously-adhered (is it magnetized?), black dust throughout the whole 2040 unit's insides. Don't get me wrong, it actually looks pretty good for a 29-year-old piece of retrotech that has probably been in a storage locker or worse for most of its existence, but it's a long way from being clean enough that I'd feel safe trusting elderly, fragile magnetic media to it.

Other than these issues, the thing is in beautiful shape, with only minor cosmetic issues -- and a sticker in the lid indicates it's been upgraded to DOS 2.0, too. Definitely worth the $20, IMO.

Does anyone know what order the two mechanisms are supposed to sit in the 2040 case? According to the drive mechanism bezels, reading from left to right it goes {drive 1}, then {drive 0}. Totally unimportant, I know, it just seems a bit eccentric, and I can tell someone's had the thing apart at some point because the bezels and so on are sitting very slightly crooked in the case.

Thanks to anyone who can advise,
G.
 
Last edited:
Yes. Bad RAM. I *think* you'd want to replace D4/5 RAM. I also *think* that the 2040 and 4040 boards have the same RAM markings.
Let me know how it goes. There may be other problems too, but get some new 2114's and see how that goes first. I don't think a hot chip in this case is a problem.
 
Gah! Ninja'd during my edit!

Thanks for that much, anyway, Bill. :) I lack the skill or experience to dare attack the thing with a soldering iron myself -- does anyone know what the going rate to have someone do the chip-swapping at an electronics shop is? I probably can't afford it anyway, but there has to be some solution.

G.
 
-- and a sticker in the lid indicates it's been upgraded to DOS 2.0, too. Definitely worth the $20, IMO. Does anyone know what order the two mechanisms are supposed to sit in the 2040 case?

$20 is a good gamble. I paid $50 plus $50 shipping for my 2040 with only the standard notice of "it was working 25 years ago when I put it in the closet". I was lucky and it worked. Drive 1 is on the left and drive 0 is on the right. With DOS 2 ROMs, it is functionally equivalent to the 4040.

Check zimmers for info on the Shugart 390 drives.
http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/index.html

This is the closest to a manual. It is for the single drive 2031:
http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/2031/index.html

Let me know if I can help with anything. What PET will you hook up to?
-Dave
 
Last edited:
I'm sure you could find a volunteer to help. where are you located? I have the 2040 user guide and some tech info. Tech references interchange the 2040 and 4040 often, they must be similar as far as the control boards go.

first blow it out to clear out the dust and debris, then clean with soap and water (lightly), then clean with isopropyl alcohol. It may seem like a daunting project but if you keep at it it's worth the experience.



bd
 
If you get to the point that you have the correct lights, contact me or others here for software/testing tips. remember that this is a DOS 2 system so you have to use the old commands.
 
Project update:

The board is indeed the one at Bo Zimmerman’s site, but this one appears to have been modified. That does tend to explain why the one chip was getting hot — it wasn’t intended to be there at the factory. If you look at the upper left corner of that board layout, there are two unpopulated chip positions between C40-C39 and C39-C2. The first of those has a 7406 (not, as I erroneously reported above, a 7404) occupying the lower 14 of the 16 solder pads at that position; this is the only chip on the board that gets noticeably hot. The one on the right has a normal-sized resistor with three red bands and one gold, soldered across the bottom two solder pads of that position. If memory serves, that means it is good for 220 Ω ±5% with a 1/8 watt power rating, correct? I am completely at sea as to the purpose of this two-component modification. Has anyone seen such a thing elsewhere?
 
Project update:

The one on the right has a normal-sized resistor with three red bands and one gold, soldered across the bottom two solder pads of that position. If memory serves, that means it is good for 220 Ω ±5% with a 1/8 watt power rating, correct? I am completely at sea as to the purpose of this two-component modification. Has anyone seen such a thing elsewhere?

This is not much of a mod, an open collector hex inverter chip and an associated pull-up resistor of 2.2K ohms. If all the inputs were open, then all the 6 outputs would be drawing current so it might get a little warm but nothing to worry about.

There could be any number of reasons for this. Possibly someone was thinking of adding a new LED indicator or maybe someone bypassed a circuit that had a shorted trace as a repair, etc.

Are any of the outputs wired to something? You may have to trace the connections.

Don't worry about this mod unless there is a problem after you replace the RAMs.
 
Last edited:
I lack the skill or experience to dare attack the thing with a soldering iron myself --
Sure you can. Here are the steps.
1. Buy several 2114s in case one is bad, also buy two 18 pin IC sockets and some solder wick.
2. Using small diagonal cutters (dikes), cut off each leg of RAM chip close to the body. This will mimimize possibility of overheating the board when unsoldering legs.
3. Remove each leg using mimimum time as possible. Don't worry about the solder that is left.
4. Use solder wick to clean holes. http://www.ehow.com/video_4435740_use-solder-wick-remove-solder.html
5. Solder in the socket
6. Insert chip
 
Last edited:
Thanks Dave, it does sound simple when you put it that way.

I can’t do anything about the 2114s until I get paid, but I looked at the back of the board to see where the extra components are connected and... They’re not! The chip is connected to power and ground rails at the usual points, pin 4 is tied high at one leg of the resistor, and neither the other end of the resistor nor any of the 7406’s other pins are connected to anything. Very odd. Maybe someone got interrupted midway through a project?

Scrounging around for datasheets reveals pin 4 of a 7406 is the output of open-collector inverting buffer #2. I think this was intended as a pull-up resistor but whoever installed it got turned around in going from the front side to the back side of the board, and soldered the wire to the wrong leg of the resistor. No wonder the chip gets so hot, so quickly — only whatever internal resistance it has is keeping it from being an outright short circuit.

D’you think the excess current being thus diverted could be disrupting the RAM’s operation? The 7406 and UD5 are almost next to each other on the board, connected in parallel by about 1.5 inches of power and ground rails. (UC5 and some decoupling capacitors, also connected in parallel, lie between them.)

Investigations continue,
G.
 
Scrounging around for datasheets reveals pin 4 of a 7406 is the output of open-collector inverting buffer #2. I think this was intended as a pull-up resistor but whoever installed it got turned around in going from the front side to the back side of the board, and soldered the wire to the wrong leg of the resistor. No wonder the chip gets so hot, so quickly — only whatever internal resistance it has is keeping it from being an outright short circuit.

Yikes, that's not good for the 7406. I hope you unhooked the +5v from that circuit. It should not have much effect on the rest of the board assuming the +5V power supply is good, but who knows? If the chip was super hot maybe it affected adjacent components.
 
Oh, I chopped out the wire to the wrong leg of the resistor, and the one from the 7406 to the ground rail, as soon as I realized what the implications of that circuit were. Couldn’t readily cut the +5V connection because the chip is soldered directly to the rail (side effect of the soldering points’ position on the board). Still, the whole abused 7406 is now floating at +5V with no other connections, so apart from adding a teeny bit of stray capacitance it shouldn’t really affect anything now.

Naturally, the drive’s startup error is unaffected by this change. Ah well, it was always a long shot.
 
I still have a few broken 3040 drives in the basement, items I've offered for shipping costs or pick-up but nobody yet wanted to touch. I have been meaning to dump them (!) for almost a year now, but they remain in the basement. I doubt I will have any time or incentive to have another round of troubleshooting with them, but perhaps I should keep them piled in the basement for as long as I can afford the space in case one needs parts later on. At least the circuit boards contain valuable chips (if they can be tested good) and perhaps also drive mechs, case and power supplies can be of use.
 
Anders, I’d offer to buy parts off you if we weren’t so expensively far apart. :¬) It seems there is no real difference other than firmware between 2040, 3040 and 4040 drives; not sure about the 8040. Seems Commodore just rebadged them to match whatever the current PET lineup was called. Can anyone confirm or deny?

Actually, Anders, if I pay the shipping plus a bit, could you mail me the plastic latch off one of the drive mechanisms? One of mine is broken off — the business end still works but there’s next to no handle left on it. It’s dead easy to remove; just open the lid on the unit and push inward on the latch, and it pops right off into the inside of the case. (Naturally, I did not figure that out until I had already unscrewed the little rail thingy it clips to. :¬\ )

I don’t know, maybe I should wait until I know if the affected mech even works… Not sure. I get paid Friday, so with any luck I’ll have enough to order some 2114s and try that repair mentioned above by next week. Found a cheap reseller of them at www.arcadechips.com, but he doesn’t specify access speed and his minimum order of $15 means it’s actually cheaper to pay half again as much at Jameco.
 
I've never heard of a 8040. Maybe you are thinking of the 100 tpi (77 tracks) 8050 drive? Sure I can look for a plastic latch. There may be a few different designs, so I'll PM you about my findings and which model fits your drive the best.
 
Gerkh. My mistake. I got a bit muddled there, thinking that if they rebadged the same 2040 drive for the 3000 series and 4000 series PET/CBM lineups, they must logically have done the same for the 8000 series. Logic and the Commodore people in the same sentence? Silly me… *wince*
 
Today I brought up all those drives for one final round of debugging.



I have five 3040/4040 drives which all have the same main boards, assy 320820. I don't get any error blinking, just the usual on-off on the floppy drives. However neither drive spins up, but I'm not sure these drives should spin up at power on.

When I read the error channel with PRINT DS$ on a CBM 4032, three of the boards give me an infinite number of characters scrolling in the upper few lines of the screen. One board currently lacks 6502 and 6522 so I didn't test it recently, but I know it won't work out of the box. The fifth board simply reports "F" followed by a few graphic characters, in total five or six. The behavior is consistent.

Now I have a few questions I haven't been able to answer myself:

1. Should the digital board boot properly even if no drives were attached to it, or will the boot fail and any readings be random?

2. If I get any kind of action from the drive, even garbage output on the command channel, can I assume the 6522 is OK or should I begin swapping those around?

3. All boards have an empty, but also unlabeled slot between UL1 and UH1. Is that as intended?

4. On all five board, UL1 = MPS 901468-06 and UH1 = 901468-07. These are marked as 4040 ROMs, DOS version 1 on Zimmers. Am I likely to be able to read those in my EPROM programmer, if configured as a 2532 EPROM? How likely is it that ROMs in all five drives have gone bad?

5. I have a pile of huge (!) Commodore PET binders in my basement. Do you believe any of them may contain a service manual for these drives? (Oh well, that is a hard question for you to answer, I'd better check myself)

If nothing, at least I will pick some spare parts for people like gsteemo and alike.
 
5. [..] Do you believe any of them may contain a service manual for these drives?
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.
 
I have five 3040/4040 drives which all have the same main boards, assy 320820. I don't get any error blinking, just the usual on-off on the floppy drives. However neither drive spins up, but I'm not sure these drives should spin up at power on.
On power up, the drives make a momentary rattling sound on my 2040 with DOS 2.1

When I read the error channel with PRINT DS$ on a CBM 4032,

Mine shows 73,cbm dos v2,00,00. But your DOS is so old, check the 2040 manual, it may not have that command. Maybe you should upgrade to DOS 2.1

3. All boards have an empty, but also unlabeled slot between UL1 and UH1. Is that as intended?
I think that is normal for the DOS 1 ROMs as you only have code in E000-FFFF. My DOS 2.1 has 901468-14 in that slot for D000.

Please post the 2040 manual I have been searching for it.
-Dave
 
Last edited:
Back
Top