Re: Non-working Commodore PET 3016
 

Refugee, it appears to be a zero-ohm short.

If so the only thing which will get warm is the 22R resistor as the short has no apparent resistance for any watts to appear across. I'm also not keen on the idea of connecting a low impedance supply to a line which has an IC output, the CLK1 output from UG5, on it. If the PSU trick manages to blow the short open circuit, the voltage on CLK1 will fly up to +5V and that voltage will be applied to the output of UG5 and the inputs of 7 otherwise unpowered ICs.

I sometimes do use a heavy current PSU to try to blast open an invisible track short but I always remove every active component from the line in question first, just in case I do succeed in blowing the short open.

Re: Non-working Commodore PET 3016
 

We have not yet found out what the lowest value the ohm meter can realistically see.
If a very sensitive volt meter could be found or made with an OP-amp and basic meter the drop along the tracks could be measured.
There is plenty that can be done before risking a broken via hole.

Re: Non-working Commodore PET 3016
 

I did say I really don't favour the track cutting approach, and especially not on antique equipment. I think that with some care and patience it should be possible to desolder and electrically separate only the IC pins of interest from their associated pads and through-holes.

Re: Non-working Commodore PET 3016
 

OK - here's a thing. I've spotted from the schematics that CLK1 also goes to UH3 pin1. Continuity testing is giving me an intermittent connection here. If I press a bit on the connections on the board (I'm sure they have a proper name) I can get a signal, but without pressing, I can't.

I'm assuming this isn't helping overall, but might it affect the frequency problem?

Re: Non-working Commodore PET 3016
 

Yes, I missed that one. Hidden in plain sight. Your circuit diagram reading skills are a lot better than you originally thought. Better than mine, in fact. ;)

That brings us up to 8 ICs (discounting the CPU which you have already shown is not the source of the short) which could be where your short is.

Continuity between CLK1-out on UG5 and all of the other IC pins which that signal should go to will be one to look at later, but unfortunately the priority remains to find and clear that short between CLK1 and 0V / GND.

Once you've cleared that short then - (if UG5 has not been damaged by having its output shorted to 0V or if indeed UG5 is not the actual source of the short) - then you should be able to go to each of the other IC pins which CLK1 goes to and you should see 1MHz on each of those pins.

I've noticed that your new meter only seems to have one resistance / ohms range on it so obviously it must auto-range, does the old meter have manually selectable ohms ranges - maybe 200 Ohm / 2 K / 200K / 2M, that sort of thing?

Going back to what refugee mentioned, it would be good if we had a 'second opinion' about the actual resistance value of that short - if possible could you measure between UG5 pins 7-8 and say what the resistance is according to your old meter (on its lowest Ohms range, if it has more than one Ohms range).

If there is any significant resistance at all (more than 1 or 2 Ohms, say) that would make it much more likely to be a short inside a chip, but if the resistance is as ultra-low as your newer meter says it is then it remains more likely to be a physical short.

Re: Non-working Commodore PET 3016
 

You could just desolder the one pin on UG5 to prove it is generating the 1Mhz as a start.

Re: Non-working Commodore PET 3016
 

Yes, that might be progress of a sort. If Colin could manage to desolder and disconnect pin 7 of UG5 from its associated pads and through-hole then we could see whether UG5 is the source of the short / is working when disconnected from the short.

If we eventually get to the point of suspecting that one of the eight, no, nine ICs soldered to the CLK1 line is faulty then I would probably choose UG5 to rule out first, as the output on pin 7 is driving the inputs of all the other ICs connected to that CLK1 line. I think Mark suggested much the same thing.

Colin, you haven't really said what you have in the way of desoldering gear, if any?

Re: Non-working Commodore PET 3016
 

I'll check out the resistance as suggested. I found the connection by continuity checking every single pin from UG5 pin 7, so I've done that and found no 'extra' pins that shouldn't have a connection. I thought I was onto a winner with UH3 Pin1 until I looked at the diagram...

In terms of desoldering gear - I have none. What would you suggest?

Thanks.

Re: Non-working Commodore PET 3016
 

How deep are your pockets?

At the bottom of the pile is the humble manual desolder pump which has a cylindrical metal body, a spring loaded plunger at the top end, a tapered teflon nozzle at the bottom end and a trigger button on the body. You melt the solder joint with your soldering iron and then while the solder is 'wet' whisk away the iron, apply the tip of the button to the still molten solder and press the button. They are cheap but crude and you usually have to do each pin several times until you get it reasonably clear.

Next level up from that is something very similar to the above but in place of the teflon tip it has an electrically heated hollow metal nozzle which you place over the pin / pad, allow the solder to melt and then press the button.

Far better than either of these is a proper desolder station which incorporates an electric pump which provides constant suction through the hollow metal tip. As you might guess these are a lot more expensive, in fact until quite recently a complete desoldering gun station from a 'respectable' manufacturer like Weller would set you back quite a few hundred pounds. Nowadays I would expect you could get something functionally equivalent from China for in the mid tens of pounds upwards - however I have no experience with these so I'll invite everyone else to chip in with their recommendations.

As well as all of the above there is something called 'desolder wick' or 'desolder braid' which is flat copper braid soaked with flux. You can press this onto a soldered area with the flat tip of a soldering iron and it will soak up quite a lot of solder. It's not a substitute for any of the above but there are certain jobs where a bit of desolder braid will work better than a desoldering iron, so really you need a desolder pump / desolder iron and some desoldering braid in your desoldering armoury.

Re: Non-working Commodore PET 3016
 

Here's an example of a low end 'proper' desoldering gun from CPC. Not a recommendation by me of this particular product as I have not used one, so I can't say if they are any good - it's just an illustration of the type of thing I mean.

https://cpc.farnell.com/duratool/d00...lug/dp/SD01384

Re: Non-working Commodore PET 3016
 

Would you look at what I found on my electric-bits shelf.

This do to start off with?

Also I attach a photo of the bottom of the board - this is one of the DRAMs which looks brown to me. Is that looking like it'll have to be replaced? if so I can start practicing de-soldering that one if it's already broken.

Are they easy to come by if I need to buy one (or more)?

Thanks.

Re: Non-working Commodore PET 3016
 

There used to be desoldering blocks the would fit in a soldering iron in place of the bit that would heat all the pins of an IC so the it could be pulled from the board. These were useful for rescuing ICs if the pcb was less important than the IC, as there was a serious risk of pulling tracks or pads off the board. Definitely not recommended for the PET.

You can also use a heat gun, paint stripper type, but again pulling the IC is serious risk of damage to the pcb and the uncontrolled heat can burn the pcb beyond repair.

Just listing these to recommend not using these methods.

Manual solder sucker is probably possible with a lot of care and patience.

I wouldn’t trust the soldering iron with plunger type, i think the shock to the pcb pads when the spring is released while the metal bit is in contact is likely to remove the pads.

The motorized suckers would probably be best, though I haven’t tried one of the cheap types. In the factories we used weller brand, but cost far too much for me to consider for hobby use. I think the cheap motorized suckers are probably also a bit expensive for hobby use, and not sure how well these work.

For a one off repair either try a manual solder sucker or find someone local with professional rework tools.

Re: Non-working Commodore PET 3016
 

That would probably do the job.

That DRAM looks like it has been replaced already, the brown is probably old flux that has not been cleaned. I would say leave it alone for now.

Do you have anything not on the PET that you could use for practice?

Re: Non-working Commodore PET 3016
 

I suspect the browning just means that the chip has been separately reworked at some point, the only other reason it would go brown like that is if the chip was running very hot for a very long time.

If you've never used those desoldering tools I suggest you do not start on the PET but practice on a scrap board, ideally a double sided board with the ICs soldered into it like the one in the PET.

The problem with that manual type of solder sucker is you usually have to retry each joint a few times, putting more solder back on the pad each time so the sucker has a decent amount of solder to go at.

The more heat you apply and the more often you apply it, the greater the chance of damaging the pad, the PCB, or breaking down the bond between the pad and the PCB. Knowing how much heat you can get away with for how long is a matter of experience, that's why I suggest you practice on something non-critical before you try to tackle anything on the PET.

Edit: Crossed with Mark.

Re: Non-working Commodore PET 3016
 

Forgot to mention: I don't know what sort of solder you are using but I would absolutely avoid Lead-free solder especially for this job where all the original solder will be Lead-Tin solder anyway. Unleaded solder is so dire when compared to the real thing that I would not wish it on anyone.

Re: Non-working Commodore PET 3016
 

Yes, if you do use lead-free solder you have to make sure you've removed all the original solder, as they don't generally mix very well.
Plus any lead present (In tin-lead plating etc) can stop lead-free solder from making a good joint (although that my be more of an issue with solder pastes and auto-assembly machines).

You also need the soldering Iron temperature around 30degC higher for many Lead-free types, as with original tin-lead types, the lead lowered the melting point of the alloy compared to that of Tin alone.
So some older fixed-temperature iron types, might not quite get hot enough to fully-melt lead-free solder (Although the cheapest ones, with no thermal feedback, just overheat with no thermal load on them making the tips / elements not last too long!)

Re: Non-working Commodore PET 3016
 

Hi all - apologies for dropping off for a few days - a few issues at home. I intend to go through this whole thread and re-do what I have done as I am starting to get an intermittent 1.00Mhz reading on UG5 pin7, but it is intermittent.

I will do some more careful cleaning as I cannot see any obvious shorts. I fear I may have a dry joint somewhere that may be beyond me.

I will go through the thread and hopefully report back tomorrow or Tuesday.

Re: Non-working Commodore PET 3016
 

Hi Colin,
That might be progress, maybe your short to ground has become an intermittent short to ground. That would start to point more at two pins or component leads bent together, or possibly a loose cutting floating about under one of the sockets, that may have been loosened by the investigation so far.

Try clipping your meter from clock to ground so your hands are free and poking about at the areas where the clock is connected to anything.

Re: Non-working Commodore PET 3016
 

With a > 1 layer thru-plated PCB, Dry-joints are very unlikely. And it's also much less likely to have joints crack with time / heat-cycles etc. with these.
So any intermittent connections are more likely to be with connector contacts - especially if not gold-plated or that tight, but usually re-mating a few times sorts it out (at least in the short-term).

The could be a cracked passive component, but much more common with surface-mount, and quite rare on a fairly-rigid PCB with leaded parts.

I have seen crystals become a bit intermittent (especially if leads get pulled), but you have to be careful that 'scope probe capacitance doesn't stop them oscillating (probe on x1 will usually cause this)

Re: Non-working Commodore PET 3016
 

Don't forget there was originally a definite short between the 1MHz clock line and 0V, so I would agree with Mark that the short has maybe gone intermittent - which is annoying because that will make it even harder to find.

I wonder what happened to the pin which fell off the PROM IC? (I mean, I wonder where it went).