1978 Commodore PET project

[SiriusHardware]

And just to be sure, you know that by 470R I meant 470 Ohms? 'R' is a commonly used abbreviation for 'Ohms' because it is usually quite a chore to find a 'Ω' symbol on a keyboard...

OK, your 1K resistors will do. Get one with its leads pointing straight up and down away from the resistor body, the way they come out of the tray / box. Lay the body of another one alongside the first one and wrap its leads around the straight leads of the first one.

Get your meter and measure your new home made resistor which should now show about 500 ohms, pretty close to the 470R we were looking for. (When you connect two resistors of the same value together side by side, the resistance is halved). Be careful not to include your fingers / your body resistance in the measurement.

Let us know when you are at that point.

[BrackenFix]

ok, have done that and got 493 ohms

[Refugee]

Have a look for a resistor that is yellow, violet. brown.

[SiriusHardware]

OK, experiment number one, as per Mark's earlier suggestion.

Try placing your newly made resistor between UF7 pin 14 and UF7 pin 9 for a second or two and see if that makes any difference to the display.

Experiment number Two:-

Connect the resistor between UF7 pin 14 and UF7 pin 9 before turning the machine on, turn it on and observe the effect on the screen, if any.
With the machine still on, remove the resistor and report the effect, if any.

Take your time, be careful not to skid or slip or short any adjacent pins together.

[BrackenFix]

That was so hard!! lol

I managed to connect it while off and turn on and it improved!!

Taking it off while running brought back all the text on the screen

[BrackenFix]

Getting there

[SiriusHardware]

OK, so what we seem to have here is that one of the four IC pins connected to this particular circuit node is faulty and keeping that line a bit higher than it should be.

By inserting a resistor to pull the line back down a bit you managed to make the display look a bit more normal.

There are four pins from three different ICs connected to that circuit node - display RAM IC UF7, Display latch IC UF9, and data bus buffer UE7 which has two of its pins connected to that node.

If you want to definitely fix it without thinking too hard, replace those three ICs, either one at a time (and stop replacing ICs when it starts working) or all three at once. If you remove the existing ICs by cutting the pins off them, removing the bodies and desoldering the pins one by one, that will have the least impact on the PCB.

Taking the one at a time approach, I would try it in the order UE7, UF7 and UF9.

If you have the confidence to desolder UF7 and UF8 intact, fit sockets and swap UF7 and UF8 over that will allow you to rule in or rule out UF7 as the cause without spending any more money on parts. If you have the patience to wait for your new desoldering gun and practice using it on something else first, you may consider doing that.

Otherwise, I would have to say go with snip and replace UE7, UF7 and UF9. (Any suggestions to narrow down the field would be welcomed).

[Timbucus]

It is a hard call as the RAM (UF7) and the Buffer IC (UE7) are both as likely to have a fault.

We could just snip the pin on the RAM chip and bend it out - that will isolate the RAM chip or the Latch/Buffer as the issue (the latch should stop the buffer input floating). Easy enough to resolder a single pin then - not much we can do about the buffer.

[BrackenFix]

Ok, I just bought all 3 chips for ease

[Timbucus]

Originally Posted by BrackenFix

Ok, I just bought all 3 chips for ease

Cool - hopefully you have some suitable sockets as well no point in not making the board future proof. See what the others think but, it may be worth the snip of the affected pin on the RAM as that is likely the way it will be removed - does not hurt having those spares they are pretty common failure points.

[Mark1960]

Originally Posted by Timbucus

It is a hard call as the RAM (UF7) and the Buffer IC (UE7) are both as likely to have a fault.

We could just snip the pin on the RAM chip and bend it out - that will isolate the RAM chip or the Latch/Buffer as the issue (the latch should stop the buffer input floating). Easy enough to resolder a single pin then - not much we can do about the buffer.

I would suggest cutting the pin on the buffer. Then the pin on the ram could be held high or low during power on with a resistor to verify the ram holds high or low. This could then give some confidence that only the 74LS244 needs replacing or if the 2114 should be replaced first.

I’d also suggest leaving the 74LS373 until after the 74LS244 and 2114 have both been replaced, it seems to behave as expected with the input pulled down.

Should the keyboard be working if this is only the display fault? Just wondering if we should try and figure that out and get some spares for the keyboard circuit in the same order.

[Mark1960]

Just to add that cutting the pin on the 2114 would leave that data line floating as the 2114 is going to be in read mode for most of the time. The 74ls373 input would float high so the symptoms would not change.

[BrackenFix]

I have discovered with the keyboard that a few more keys work if I press harder so will take it apart and clean etc

[SiriusHardware]

I would suggest cutting the pin on the buffer.

It needs to be intact or easily shortable at the point of switch on, as the initial content of the screen is loaded into the screen RAM via that buffer. (And incidentally, which pin? There are two pins of the 74LS244 connected to this node).

When we say 'cut a pin' in this context for diagnostic purposes, we mean cut the pin so that there are equal amounts of pin left attached to the IC and to the PCB so that we can make connections only to the pin, only to the PCB, or easily rejoin the two cut halves of the pin. This is difficult to do without a teeny tiny pair of very sharp cutters (the aforementioned 'cuticle nippers' may be suitable).

When we are talking about cutting pins off an IC prior to removing it, then we mean that the IC pin should be cut as high up and as close to the body of the IC as possble, as that poses the least risk of damage to the PCB tracks on the top side of the PCB.

[SiriusHardware]

I have discovered with the keyboard that a few more keys work if I press harder so will take it apart and clean etc

Have a look at this video by the 8-bit guy from about 2:46 onwards, lots there about how the keyboard works and how to troubleshoot it. ScottishColin also had to fix quite a few of the keys on his, not just dead keys but in some cases open circuit PCB tracks on the keyboard PCB.

https://www.youtube.com/watch?v=uHbhH7ISL_Y&t=38s

[Mark1960]

I think UE7 pin 3 should be cut, thats output of the buffer when writing to SD4, enabled from pin 19, but always good to have someone else confirm I didn’t make a mistake.

Pin 2 is an input, so can probably stay connected, though that does depend on the failure mode of UE7 if that is the faulty chip.

It probably doesn’t matter if SD4 floats during power up init of the display, thats probably what its doing anyway.

[ajgriff]

Originally Posted by SiriusHardware

Have a look at this video by the 8-bit guy from about 2:46 onwards, lots there about how the keyboard works and how to troubleshoot it. ScottishColin also had to fix quite a few of the keys on his, not just dead keys but in some cases open circuit PCB tracks on the keyboard PCB.

https://www.youtube.com/watch?v=uHbhH7ISL_Y&t=38s

Please remember this a chiclet keyboard and note my warning in post #71.

Alan

[SiriusHardware]

Fair point, AJ.

[ScottishColin]

The chiclet plungers can be bought from here if necessary.

https://www.shapeways.com/product/P6...et-key-plunger

They don't seem all that cheap - I'm happy to design/print them if needed but that will obviously take time.

Colin.

[BrackenFix]

My plan is to take the keyboard apart and assess it this week while waiting for the ICs to arrive