Commodore PET 4016

[ScottishColin]

No - none of them are socketed.

4032 UE7 and UE6 pins 13 (red) and pin 1(Black) read 2k ohms.

All 4016 readings between Pullup and Init pins on the ICs listed on post 279 read 131 ohms.

Thanks.

Colin.

[SiriusHardware]

I saw that.

We have an unusually difficult problem here, as the state of the INIT and PULLUP_1 lines combined is shutting down virtually every chip that either of them are connected to, so we can't do what we would usually hope to do and logically follow signals through devices.

Therefore, your only way forward may be to start socketing and removing the devices you have identified, checking as you go to see if that low resistance between Pullup_1 and INIT has cleared, and hope the faulty one isn't the very last one you check. As you already have a replacement for UD3 on the way, I would start with that one.

If anyone can suggest a way to narrow it down without having to remove devices, I'm all ears.

[TonyDuell]

I was going to suggest concentrating on the chips that have both INIT/ and PULLUP_1 connected to them. Alas that doesn't help much, there aren't that many that only have one one of the 2 signals.

[SiriusHardware]

Colin provisionally identified the devices with both signals going to them in #279 - not a mega long list, but long enough.

[Mark1960]

I’m guessing only one ic has an output to init/ and none have an output to pullup_1.

If 5v is applied to pullup_1, can the effect of pullup_1 be tested on each of the ics its connected to.

Remove the ic driving init/ and try the same on init/.

With the ic driving init/ removed, connect 5v to init/ and 0v to pullup_1, or the other way round, and see what shows up on thermal camera. Wear safety glasses just in case something blows.

[SiriusHardware]

Actually there is no IC driving INIT, they are all inputs on that line. There is a single pullup resistor to +5V. It's one of the great PET circuit mysteries.

[SiriusHardware]

Re: Mark's idea of taking one or both of Pullup_1 and or INIT directly to 5V, that might result in something actually burning up. And not necessarily in a good way.

What does everyone think of the idea of bypassing the 1K pullup resistors with lower resistance resistors, say 100R, that might pull the lines high enough to let the working ICs work, it may also warm up the faulty IC just enough to make it stand out on Colin's thermal camera.

Probably the worst that will happen is that the resistors will cook (?).

Suggestions please?

[SiriusHardware]

I'm going to be AFK for most of the evening as I'm travelling a fair way down south - I will try to look in every now and again to see how things are going.

[Mark1960]

I was expecting the current to be limited by the faulty input. Voltage divider to give the measured 1v on the inputs from 1k pullup would suggest approx 250 ohm. Colin was measuring 130 ohm short circuit.

I wonder if 5v supply is missing on one of the ICs, this would be an easy thing to check before removing any ICs. It might be that with 5v missing the ic is trying to take power through the input bias diodes from the pullup resistors.

[Refugee]

Another way would be to make a dry joint on each of the chips that are connected to both lines. you only need to do one of the pins on each of the chips. Fully de-solder these pins.
Using a home made probe with a center punch dimple in the end flex each pin with your probe and a normal probe on the other side. The pin should flex enough to lose contact with the track. An analogue meter would help if there is one to hand. One will stay at 131 ohms.

[SiriusHardware]

I was expecting the current to be limited by the faulty input. Voltage divider to give the measured 1v on the inputs from 1k pullup would suggest approx 250 ohm. Colin was measuring 130 ohm short circuit.

Well reasoned, but I don't know whether we can expect that resistance to stay stable / constant when subjected to hard +5V. It's a faulty component so it could change when subjected to stress / hard voltage. Could stay the same, could go open-circuit, could go lower resistance or short-circuit.

I think it would be safer to at least start off with lower value pullup resistors to see if we can lift those voltages into the logic '1' area in order to enable the chips affected by these lines being low.

Refugee, this is a double sided board of course - I know Colin has quite decent desoldering gear now so we'd better let him decide whether he thinks he can desolder individual pins 'dry' enough to completely disconnect them from their associated top pad, via and lower pad.

[SiriusHardware]

I wonder if 5v supply is missing on one of the ICs, this would be an easy thing to check before removing any ICs. It might be that with 5v missing the ic is trying to take power through the input bias diodes from the pullup resistors.

Quite a few ICs to check but it wouldn't take too long and as you say would be a non-invasive test to try before doing anything else.

[ScottishColin]

I've tested all the ICs listed in post #279 and plenty more besides - +5v everywhere in the right places.

Colin.

Originally Posted by Mark1960

I was expecting the current to be limited by the faulty input. Voltage divider to give the measured 1v on the inputs from 1k pullup would suggest approx 250 ohm. Colin was measuring 130 ohm short circuit.

I wonder if 5v supply is missing on one of the ICs, this would be an easy thing to check before removing any ICs. It might be that with 5v missing the ic is trying to take power through the input bias diodes from the pullup resistors.

[ScottishColin]

Thanks for the suggestion - I've done a lot of soldering and de-soldering of PETs in the last few years but I'm not sure I'm confident enough to do it this way.

Colin.

Originally Posted by Refugee

Another way would be to make a dry joint on each of the chips that are connected to both lines. you only need to do one of the pins on each of the chips. Fully de-solder these pins.
Using a home made probe with a center punch dimple in the end flex each pin with your probe and a normal probe on the other side. The pin should flex enough to lose contact with the track. An analogue meter would help if there is one to hand. One will stay at 131 ohms.

[Mark1960]

One more idea in case the 5v is disconnected inside one of the ICs is to try and measure resistance between some other pins on each IC between the pin and 5v. Try to pick pins that are not connected to too many other ICs. Possibly use the diode test selection on the multimeter if it has one.

[ScottishColin]

Will do. What's a 'good' or 'bad' measurement?

Colin.

Originally Posted by Mark1960

One more idea in case the 5v is disconnected inside one of the ICs is to try and measure resistance between some other pins on each IC between the pin and 5v. Try to pick pins that are not connected to too many other ICs. Possibly use the diode test selection on the multimeter if it has one.

[Mark1960]

It probably depends on your multimeter. Diode test should be about 600mv with positive on the input pin and negative on the 5v pin. You would be measuring the internal parasitic diode of the input pin.

[ortek_service]

Although it is mostly CMOS IC's that have ESD-protection diodes on their inputs to supply & ground.

However, TTL inputs are into emitters of a bipolar transistor (which will be effectively two diodes from base to collector and emitter(s)).
So diode-test on these will be effectively testing part of the input transistors (plus internal resistors).
Also, 74(L)Snn types also have Schottky diodes added, to make them faster, as detailed here: https://en.wikipedia.org/wiki/Schottky_transistor

[SiriusHardware]

I'm now as confused as Colin probably is over whether this measurement from pins to +5V will yield useful results - partly because of what Owen said just now.

Unless anyone violently disagrees I'm going to suggest tack-soldering a lowish value resistor, perhaps 100R, across each of the two 1K pull-up resistors R2 and R28, with the aims of:

-Enabling the ICs which are currently disabled by the logic-0 state on the pullup_1 line and the INIT line, allowing most of them to work and hopefully making it easier to see which ones still aren't working

-Passing a greater, but still limited, current through the faulty IC may warm it up enough to make it stand out on Colin's thermal camera.

The first task after fitting these resistors would be to measure the voltages on UD3 pins 1 and 13 to see what, if any, change fitting the resistors has made. If we are lucky both voltages will have risen up to above the logic '1' threshold.

If you find that is the case, then look for 1MHz on pin 12 of UD3 and we'll see whether that IC is in fact working.

[Mark1960]

The measurement of pins to 5v is probably not going to work, I might experiment with some 74ls or 74 series myself just to see what the result is.

I think increasing the pull up current looks like the best option, but see if you can notice the traces warm up on the thermal camera too.