Non-working Commodore PET 3016

[SiriusHardware]

We have what look like good clock / clear input signals to both halves of UH1 but no apparent output from UH1, either part.

Could you just verify that UH1 pins 2 and 12 both have steady high states on them? Also look for any activity on UH1 pins 5 and 9.

[Mark1960]

From #1119 inputs to UH1 you should have activity on UH1-8 and UH1-6.

To be a little more certain please confirm pins 2, 4, 10 and 12 of UH1 are constant high.

Then See if there is any activity on UH1-5 and UH1-9.

[ScottishColin]

UH1/2 and UH1/12 both high - 5.3V.

UH1/5 - slight uptick measured as 0.6V on a meter, but no waves. UH1/9 goes high - 5.4V on the meter.

Colin.

Originally Posted by SiriusHardware

We have what look like good clock / clear input signals to both halves of UH1 but no apparent output from UH1, either part.

Could you just verify that UH1 pins 2 and 12 both have steady high states on them? Also look for any activity on UH1 pins 5 and 9.

[ScottishColin]

UH1 2/4/10/12 all go high and stay there.

Colin.

Originally Posted by Mark1960

From #1119 inputs to UH1 you should have activity on UH1-8 and UH1-6.

To be a little more certain please confirm pins 2, 4, 10 and 12 of UH1 are constant high.

Then See if there is any activity on UH1-5 and UH1-9.

[Mark1960]

I think its time to get the solder sucker going again.

Do you have a 74S74? It might also work with a 7474 or 74LS74, though the timing will be slightly different it would at lease confirm the 74S74 was faulty.

[SiriusHardware]

Are we thinking dud UH1? An awkward one, 74S74. Are they available? Must be a specific need for that type as they used 74LS74 elsewhere. Crossed with Mark.

I think Colin may have a 74LS74 already, but Commodore surely used an 'S' for a specific reason.

[ScottishColin]

I bought one from post 807.

It's marked as SN74LS74AN.

That be ok?

Colin.

Originally Posted by Mark1960

I think its time to get the solder sucker going again.

Do you have a 74S74? It might also work with a 7474 or 74LS74, though the timing will be slightly different it would at lease confirm the 74S74 was faulty.

[Mark1960]

I would try the 74LS74AN, maybe try and get a 74S74 later. We’ll need to keep it in mind in case of any strange issues later.

Make sure its socketed.

[ScottishColin]

Cricklewood have the 74S74 if I need that specifically?

Colin.

Originally Posted by SiriusHardware

Are we thinking dud UH1? An awkward one, 74S74. Are they available? Must be a specific need for that type as they used 74LS74 elsewhere. Crossed with Mark.

I think Colin may have a 74LS74 already, but Commodore surely used an 'S' for a specific reason.

[ScottishColin]

So - what's the difference between a 74S74 and a 74Ls74 then? What does the L mean?

Colin.

[SiriusHardware]

It should probably show different behaviour from the apparently dud original UH7, but best to try to source a proper replacement if possible. I'll have a look around and see if there are any historic instances of this having been replaced with other flavours. Bit of a long shot.

What's so exotic about the 'S' series? Slothie will know.

If Cricklewood have the 74S74 I am beginning to think they must be PET parts specialists, so yes, if you can get one please do.

[SiriusHardware]

Wikipedia description of the 74LS series which suggests that the difference between 74S and 74LS is lower switching speed in 74LS, traded for lower power consumption.

74LS – Low-power Schottky. Implemented using the same technology as 74S but with reduced power consumption and switching speed, due to larger resistors. Typical 10 ns gate delay, remarkable (for the time) 2 mW dissipation, 4.75–5.25 V.

[ScottishColin]

Ordered. Is there likely to be a problem if I pop a socket in and try the 74LS74 first?

Colin.

Originally Posted by SiriusHardware

It should probably show different behaviour from the apparently dud original UH7, but best to try to source a proper replacement if possible. I'll have a look around and see if there are any historic instances of this having been replaced with other flavours. Bit of a long shot.

What's so exotic about the 'S' series? Slothie will know.

If Cricklewood have the 74S74 I am beginning to think they must be PET parts specialists, so yes, if you can get one please do.

[SiriusHardware]

As in a chain reaction in which all the other ICs burst into flames? Possibly not. Since the right IC is available, I personally would be Mr. super-patient and wait until it arrives, but you are free to do as you wish.

Ironically, when I gave you that short list of ICs to order just in case, I deliberately sidestepped that 74S74, hoping we wouldn't end up needing to find one.

A while ago I joked that the alternative repair method would be just to re-chip the whole PCB. It's almost heading that way. I can't imagine how this could have ended up in this state with such a high and seemingly almost random chip mortality rate.

[ScottishColin]

I'll get the socket in tomorrow. Probably Tuesday before the new IC turns up.

I hope none of the duff chips have been due to my ham-fisted testing. I'm still determined though.

Colin.

Originally Posted by SiriusHardware

As in a chain reaction in which all the other ICs burst into flames? Possibly not. Since the right IC is available, I personally would be Mr. super-patient and wait until it arrives, but you are free to do as you wish.

Ironically, when I gave you that short list of ICs to order just in case, I deliberately sidestepped that 74S74, hoping we wouldn't end up needing to find one.

A while ago I joked that the alternative repair method would be just to re-chip the whole PCB. It's almost heading that way. I can't imagine how this could have ended up in this state with such a high and seemingly almost random chip mortality rate.

[SiriusHardware]

I don't think you are responsible for anything which has happened here. The only thing we can definitely say is that the machine lived in a very damp environment for a long time, but I can't really imagine how that would cause such large scale chip death, I would usually expect that the hardware (tracks, vias) would be the first to suffer under those circumstances.

[Ambientnoise]

The S series was 3 times as fast as LS (3nS delay vs 10nS, I think) but had 10 times the power consumption. There were other later series with different trade offs, like ALS (Advanced Low power Schottky). Maybe the high speed of S was needed in a specific location but ALS or AS are reasonable fits too if the LS doesn’t work.

[SiriusHardware]

Thanks for the info. They obviously needed that quickness in this specific location. As the ever dependable Cricklewood seem to have the 74S, I think we may as well fit the real thing.

[Mark1960]

I suspect this pet had some overvoltage or static discharge that stopped it working before it went into storage.

For the 74LS74, I don’t think we need to worry about setup time as D is held high and there looks like plenty of recovery time from reset to clock. Propagation time from clock or reset to /Q could affect timing between RAS and CAS, which could be a reliability issue for the dynamic ram.

I don’t think there is a risk of the 74LS74 causing any further damage. I’d be tempted to try it, but its not my pet.

[SiriusHardware]

For the avoidance of doubt / fear, I don't think it will cause damage but it may cause misleading behaviour which could send us off on a wild goose chase, unnecessarily so when the right IC is actually available. If you want to try it, I'm quite happy to observe from a safe distance.

I did wonder if possibly the PSU has a mind of its own and likes to lash out every now and again - if possible maybe eventually put some high wattage 'just above rail voltage' zeners on the supply rails so that if the regulators do go doo-lally it won't do too much damage.