Commodore PET 4032/8050/3022

[ScottishColin]

I believe it also goes to pin2 of UB13 on page 10.

Colin.

Originally Posted by SiriusHardware

OK. Good work working that out, now I wonder what other differences lie in wait for us? With the power on, again measure the voltages on both ends of R15.

With the power off, measure the in-circuit resistance of R15. It should be 1K or less, or so we hope. If it reads anything higher than 1K, temporarily unsolder one end and measure the actual resistance of the resistor only.

The UD15 element which was under suspicion appears blameless (so you did a good job taking it out in one piece).

Don't put it directly back in the board just yet even if you plan to do that later because there are five other gates in that chip and we may, in time, come back around to this chip and find that one of its other gates is stopping something from working, in which case we will be glad that it is currently in a socket. (see below!)

If I look at the circuit fragment in #116, I see R13(really R15), UD15 pin 6 and the CPU's pin 40 (RESET) on that circuit node, but there is also a connection heading off-circuit from there to circuit sheets 2 and 3.

On circuit sheet 2, it is connected to the RESET input of a 6520 (UB16), the IEEE port handler, and also to another pin on UD15, pin 13.

On circuit sheet 3, it is connected to the RESET inputs of the 6255 (UB15) and the other 6520 (UB2)

I don't see the RESET line going to anywhere else, but I would be happy to be corrected.

Altogether, there aren't many devices on this circuit node which can be keeping it low, so try the following:

Remove UD15, put pin 6 back into shape and then bend out pin 13 and re-insert UD15. Power, on, what is the standing voltage on the CPU RESET pin then?

Power off, remove UD15, put UD15 back into shape and re-insert it with all pins properly inserted.

Remove the CPU, which is the only other socketed IC on this circuit node, and leave it out. Power on - what voltage do you see on the pin 40 (RESET) receptable on the CPU socket?

[ScottishColin]

I also get a reading of 19 ohms on my meter to pin 33 of UB13 and UB14 (6502). That seems odd.

Colin.

[SiriusHardware]

Ah, I did see that someone had added a '10' to that branch going off to sheets 2 and 3, so that's why...

I also get a reading of 19 ohms on my meter to pin 33 of UB13 and UB14 (6502). That seems odd.

Sorry, can you clarify: You get a resistance of 19R between those pins, or from there to somewhere else?

Could you still remove (one at a time) pin 13 of UD15, and then the CPU, and see if the RESET line comes up to a high state with either of those removed from the line?

[SiriusHardware]

Just looked: So the device the RESET line goes to on sheet 10 is the CRTC controller, just great. So many 40-pin ICs which could be the cause of this pulldown on RESET.

Let's see what lifting the two socketed ICs does for the state of the line first.

[ScottishColin]

I'm chasing my tail on continuity tests and going slightly mad but just to clarify, I get a good continuity test with no resistance between R15 and UB16, UB15 and UB12 pin 34.

I get a 19 ohm resistance level between R15 and UB16, UB15 and UB12 pin 33.

I also get a good continuity between pins 6 and 13 on UD15.

This may well be nothing but I can't see it on the schematics.

I'll do the tests in #123 later on tonight.

Colin.

Colin.

[SiriusHardware]

6 and 13 is a valid direct connection. UD15 Pin 6 is connected to RESET on sheet 1, UD15 pin 13 is connected to RESET on sheet 2.

[SiriusHardware]

I get a good continuity test with no resistance between R15 and UB16, UB15 and UB12 pin 34.

OK, obviously those are all meant to be directly connected together, and hopefully they are also connected to the CPU pin 40?

I get a 19 ohm resistance level between R15 and UB16, UB15 and UB12 pin 33

Pin 33 of all of those devices being the D0 line, unrelated to RESET? It would certainly be unusual for RESET and DO to be so closely connected together.

One step at a time, though, remove UD15 (6) and the CPU from the RESET line to see if the voltage on that line comes up.

[SiriusHardware]

Agghh, sorry the above was written in a hurry. I meant,

remove UD15 (13) and the CPU from the RESET line to see if the voltage on that line comes up.

I'm pursuing this angle because these ICs are in sockets and so it's something which is easy to try first.

[ScottishColin]

Did you mean remove the CPU or lift the reset pin?

Colin.

[SiriusHardware]

Just take the CPU out.

[ScottishColin]

6502 removed, UD15 inserted but pin 13 lifted.

Pin 40 of the CPU reads 2.7V.

Colin.

[SiriusHardware]

That's still not as high as I would expect it to rise if either UD15 or the CPU had been the cause of the loading on the reset line. With the CPU still out do you still have that very low resistance between reset and D0?

[SiriusHardware]

If the low resistance between D0 and RESET is still present even with the CPU out it suggests that low resistance path must be through one of the (not many other) devices which are connected to both RESET and D0. Unfortunately all of those devices are large 40-pin ICs which are not in sockets.

Try measuring the resistance between RESET and D1, D2, D3, D4, D5, D6, D6, D7, do any of those other lines show a similarly low resistance?

On your other PET where all of the large devices are sensibly socketed the next step would be to remove those one by one until that low resistance clears, but it is not so easy on this one.

One thing you could try, because you are equipped to do so, is to put everything back where it goes and then power the machine on for a few minutes and check to see if one of the large ICs is warming up much faster than the others.

[ScottishColin]

Morning.

With the CPU removed, I get a 15.2 ohms connection between pins 33 and 40 of the socket.

Colin.

Originally Posted by SiriusHardware

That's still not as high as I would expect it
to rise if either UD15 or the CPU had been the cause of the loading on the reset line. With the CPU still out do you still have that very low resistance between reset and D0?

[SiriusHardware]

OK, with everything fitted back in can you do a bit of thermal observation with that handy thermal camera of yours to see if any of the larger ICs is heating up more / faster than expected when the machine is powered for a while?

[ScottishColin]

Reset (40) to other pins:

DA0 (33) 15.3 ohms
DA1 (32) 28.3 ohms steady
DA2 (31) 11 ohms rising
DA3 (30) 21 ohms rising
DA4 (29) 12 ohms rising
DA5 (28) 19 ohms rising
DA6 (27) 50 ohms rising to infinity
DA5 (26) 60 ohms rising to infinity

A second reading gave me infinite readings on DA2->DA7 but the readings on DA0 and DA1 stayed steady.

Colin.

Originally Posted by SiriusHardware

If the low resistance between D0 and RESET is still present even with the CPU out it suggests that low resistance path must be through one of the (not many other) devices which are connected to both RESET and D0. Unfortunately all of those devices are large 40-pin ICs which are not in sockets.

Try measuring the resistance between RESET and D1, D2, D3, D4, D5, D6, D6, D7, do any of those other lines show a similarly low resistance?

On your other PET where all of the large devices are sensibly socketed the next step would be to remove those one by one until that low resistance clears, but it is not so easy on this one.

One thing you could try, because you are equipped to do so, is to put everything back where it goes and then power the machine on for a few minutes and check to see if one of the large ICs is warming up much faster than the others.

[ScottishColin]

Thermal imaging pictures attached after 15 minutes powered on with all chips and legs inserted.

The 6502 was the warmest (by a little) followed by the 6345. The ROM chips got warm too but not as much as the 6502.

Colin.

Originally Posted by SiriusHardware

If the low resistance between D0 and RESET is still present even with the CPU out it suggests that low resistance path must be through one of the (not many other) devices which are connected to both RESET and D0. Unfortunately all of those devices are large 40-pin ICs which are not in sockets.

Try measuring the resistance between RESET and D1, D2, D3, D4, D5, D6, D6, D7, do any of those other lines show a similarly low resistance?

On your other PET where all of the large devices are sensibly socketed the next step would be to remove those one by one until that low resistance clears, but it is not so easy on this one.

One thing you could try, because you are equipped to do so, is to put everything back where it goes and then power the machine on for a few minutes and check to see if one of the large ICs is warming up much faster than the others.

[ScottishColin]

And pictures 1 hour 30 minutes after power on. The 6502 is still the warmest but nothing is massively standing out to me. Unless it shouldn't be warm because it isn't actually doing anything?

Colin.

[SiriusHardware]

The thermal image approach was worth trying because it may have revealed one IC to be drawing significantly more that the rest but it seems not to have been particularly conclusive.

Make a few more resistance measurements, all with the black meter probe on 0V/GND and the red meter pin on the following points:-

From RESET down to 0V
From D0 through to D7 down to 0V
From A0 through to A15 down to 0V

My feeling is that you are going to need to socket all of the large ICs which are not already socketed so that we can add / remove them for diagnostic purposes, but try the above measurements first so we get a bit more of a picture of what is going on.

[ScottishColin]

Nothing to be read on all pins (ie infinity) except pins 1 and 21.

I was thinking the same. Not very much looking forward to this, but I'll start with a 6520 as I have one to spare of them.

Do you think I should replace the 6502 socket?

Colin.

Originally Posted by SiriusHardware

The thermal image approach was worth trying because it may have revealed one IC to be drawing significantly more that the rest but it seems not to have been particularly conclusive.

Make a few more resistance measurements, all with the black meter probe on 0V/GND and the red meter pin on the following points:-

From RESET down to 0V
From D0 through to D7 down to 0V
From A0 through to A15 down to 0V

My feeling is that you are going to need to socket all of the large ICs which are not already socketed so that we can add / remove them for diagnostic purposes, but try the above measurements first so we get a bit more of a picture of what is going on.