Rusty Commodore 4040 disk drive

[ScottishColin]

This one may be a fools errand as it's a delicate piece of equipment and has been sat in a barn for 15+ years, but you know me - I like a challenge.

This one doesn't boot up in any meaningful sense at all - no disk drives spinning, no LEDs lighting up but there is power. Schematics are here:

https://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/4040/index.html

I have a working 4040 next to it I can compare against.

I checked CR7, CR8 and CR9 as per post 26 on the thread below and get different readings to the working 4040.

https://www.vintage-radio.net/forum/showthread.php?t=190218&page=2

These are the from the failing 4040 using my multimeter:

CR7 - 11.97v
CR8 - 12.15v
CR9 - 0v

These are from the working 4040:

CR7 - 11.69v
CR8 - 12.11v
CR9 - 5.18v

Is this issue that CR9 that has failed or do I need to look deeper (knowing me it's the latter)?

I have tested the known working 4040 digital motherboard in the rusty 4040 case and it powers up and LEDs light up so I currently don;t suspect the PSU in the rusty 4040.

Colin.

[SiriusHardware]

The 5V regulator could have failed but there could also be a short across its output, so with the power off, measure the resistance between your test point on CR9 and 0V. Also try measuring the voltage on the 'other' end of CR9, because if you don't have any unregulated voltage on the input to the regulator, you won't get any regulated voltage out.

CR7, 8 and 9 are diodes of course, in that original thread it was suggested you measure there because it was a convenient point at which to measure the output voltages from the regulators, with those diodes being connected between the inputs and outputs of their respective regulators.

[ScottishColin]

Resistance measurement - 179.1 ohms.

Voltage on the 'other' end of CR9 is 1.9v (the working 4040 is 9.82v).

Colin.

Originally Posted by SiriusHardware

The 5V regulator could have failed but there could also be a short across its output, so with the power off, measure the resistance between your test point on CR9 and 0V. Also try measuring the voltage on the 'other' end of CR9, because if you don't have any unregulated voltage on the input to the regulator, you won't get any regulated voltage out.

CR7, 8 and 9 are diodes of course, in that original thread it was suggested you measure there because it was a convenient point at which to measure the output voltages from the regulators, with those diodes being connected between the inputs and outputs of their respective regulators.

[TonyDuell]

With the unit switched off/unplugged from the mains, what resistance do you measure between ground and the anodes (non-banded end) of CR5 and CR6?

With the unit turned on what AC voltage do you measure between ground and the anodes of CR5 and CR6?

[ScottishColin]

Switched off resistance:

CR5 : 5M ohms and rising
CR6 : 0L

Switched on voltage:

CR5 : -3.3v
CR6 : -3.2v

Colin.

Originally Posted by TonyDuell

With the unit switched off/unplugged from the mains, what resistance do you measure between ground and the anodes (non-banded end) of CR5 and CR6?

With the unit turned on what AC voltage do you measure between ground and the anodes of CR5 and CR6?

[TonyDuell]

Those readings are definitely on the non-banded ends of those diodes? And to confirm you are making AC voltage measurements.

Those diode anodes should connect to pins 1 and 2 of connector P4 -- the connector that carries the wiring to the transformer and smoothing capacitors. 'Pin 4' is keyway. With that connector still plugged into the logic board (but unit turned off/unplugged), measure the resistance between pins 1 and 2 and ground. Check for corrosion on all pins of that connector too.

[ScottishColin]

Duh. Apologies.

Switched on AC voltage:

CR5 : 6.03v
CR6 : 5.81v

I'll do the resistance checks next.

Colin.

[TonyDuell]

I'd now check for bad connections (wiring, P4 connector) between the mains transformer and those diodes.

[SiriusHardware]

Annoyingly, and not unusually for Commodore, the circuit diagram for the unit does not include that of the power supply 'block', but one thing which is noticeably absent from the diagrams we do have is the smoothing capacitor which is likely to be a huge / very large value thing mounted offboard just as it is in the computers.

One thing we have seen a few times on the computers is the failure of the connector connection which carries the output from the rectifiers off the PCB to the big smoothing capacitor, it can fail in an odd non-linear way where, measured in power off conditions, the connection appears OK / Low resistance, but when you attempt to draw any amount of current through it, there is a big voltage drop across it. My theory is that the pin / socket connection is gradually / progressively damaged by the massive switch-on inrush into the capacitor.

If you get to the point where this is suspected here, the way to find out for sure is to install a bypass wire around the suspect connector pin / socket connection.

[ScottishColin]

OK - thanks Tony. I've removed the board and cleaned up the Pin 4 connecting pins and now we're moving forwards to a single solid LED in the centre. So we have power.

This motherboard has 9 white sockets so I'm going to replace them next.

Colin.

[ortek_service]

Originally Posted by SiriusHardware

Annoyingly, and not unusually for Commodore, the circuit diagram for the unit does not include that of the power supply 'block', but one thing which is noticeably absent from the diagrams we do have is the smoothing capacitor which is likely to be a huge / very large value thing mounted offboard just as it is in the computers.

One thing we have seen a few times on the computers is the failure of the connector connection which carries the output from the rectifiers off the PCB to the big smoothing capacitor, it can fail in an odd non-linear way where, measured in power off conditions, the connection appears OK / Low resistance, but when you attempt to draw any amount of current through it, there is a big voltage drop across it. My theory is that the pin / socket connection is gradually / progressively damaged by the massive switch-on inrush into the capacitor.
>>
>>

It may also be worth checking that large Electrolytic can-type capacitor for the right capacitance at least - Many DMM's will do this, although their capacitance range doesn't always go high-enough.
And ideally ESR & leakage which Colin's 'Universal Tester' may do (ESR at least) - although these may also be limited to measuring upto 1000uF like many DMM's capacitance ranges?

Whilst larger low-voltage Electrolytics are normally fairly-reliable, lasting many decades and not usually needing 'reforming' that higher-voltages ones can need - although if not used in decades maybe still a risk of breakdown with a high inrush current suddenly appearing before oxide-layer can reform.
Or the liquid-electrolyte could maybe dry-out completely, if badly stored.
- Chris recently found the (rather non E6/E12 value) 23,000uF 16V? one in his 'Barn find' PET only measured a few nF's so was basically open-circuit.
Although it was so-large that there would be no trouble substituting it with a modern one around that value (22,000uF would no-doubt be OK, given the large tolerance on Electrolytic's capacitance), or maybe 're-stuff' the original can with a few paralleled 4700uF etc ones to make the required value).

[ScottishColin]

Partway through and three of the removed chips came out with one leg missing due to rust so I think there's going to be more damage as we go.

Luckily one is a 6502 (got a spare), one is a ROM chip (I can burn one of them) and one is a 6532 which I have spare as well.

I'll keep going.

Colin.

[ScottishColin]

Getting there, but I'm out of 24 pin sockets now so I'm waiting for the postie, along with a couple of 2532's on order as I have finally used all my stock of them too.

Another ROM lost a leg - I may try soldering legs back onto inidividual chips to see if I can revive them while I wait.

Also the other 6532 was faulty (tested on a working 4040) but I have spares there.

Colin.

[ortek_service]

Originally Posted by ScottishColin

Getting there, but I'm out of 24 pin sockets now so I'm waiting for the postie
>>
Another ROM lost a leg - I may try soldering legs back onto inidividual chips to see if I can revive them while I wait.
>>
Colin.

Yes, It's well worth trying to retain the original ROM's if possible to keep it as authentic as possible.


I presume the ROM IC has only lost its lower leg 'pin', with the upper 'shoulder' half still there ? In which case:

If you still had any DIL24W sockets left, then these can be useful to solder the remains of the IC's upper leg 'shoulder' to or to insert a bit of Tinned Copper Wire (e.g. a leaded resistor off-cut) into to hold it in place whilst soldering onto the IC's remains of the IC's leg 'shoulder'.
And keeping an with repaired / fragile legs in a socket, does protect these from possible future damage, even though you'd be 'double-stacking' sockets with a socket already on the PCB.

You could also try cutting-off the whole leg upper 'shoulder' from a cheap / faulty IC, and grafting that back onto the ROM IC - But 'surgery' like this can be a little fiddly and needs a quite steady hand.

If the upper 'shoulder' has gone as well, snapping-off right at the package exit, then have to start trying to drilling into the epoxy-plastic case around it, to get enough to solder a wire onto.

[ScottishColin]

6502 resurrected ok and tested by soldering a donated leg onto the shoulder that remained.

6530 next which would be really handy to fix as they're not easy to come by.

Colin.

[ScottishColin]

Nearly back in the game but I was wondering about the markings on the underside of the motherboard (see attached photo).

They look like heat to me so I suspect they may not be good news. They're mostly capacitors - C4, C15 and C36 are examples but not the only ones.

Should I pop one or two of these out to test them?

Colin.

[SiriusHardware]

That's an interesting looking phenomenon but I don't think it is heat. For a capacitor to heat up enough to blemish the PCB, it would have to have failed fairly low resistance, causing little or no power to the affected rail(s). In the short time I spent looking it looked as though the capacitors you specifically mentioned are ceramics, not electrolytics or tantalums?

By all means desolder and check one or two of them - does your component tester 'do' capacitors, I can't remember? Important: When checking capacitors with any kind of smart tester always discharge the capacitor before you connect it to the tester. In this case the all the chips on the rails should already have discharged the capacitors, but get in the habit of doing it anyway.

I don't know if this PCB lives component side up or component side down but either way I think this may be where condensation has formed, coalescing into droplets which later evaporated, leaving behind deposits of minerals which were dissolved in the liquid.

[Refugee]

It looks like the caps were hand soldered not cleaned after the chips were soldered and cleaned. It will be flux residue that has gone sticky and attracted dust.

[SiriusHardware]

That's also a possibility - I guess the real question to ask is - does it clean off if you try some IPA or similar on it?

[ScottishColin]

The motherboard is mounted component side down. I have tried cleaning with IPA but this doesn't remove the markings.

I have removed C41 and tested - photos attached.

I am going to put this motherboard in the known working 4040 chassis, with known working drives, analog board, cables and PSU. Also all removable chips have been tested one by one in the working 4040 motherboard so I know they're good.

Colin.