Perth PET

[ScottishColin]

Let me clarify my previous statement.

With power off, there is continuity from Pin 12 on UD6,7 and 8, but not on UD9.

And I will - ta.

Colin.

Originally Posted by ajgriff

IC decoupling capacitors are quite often omitted from complex schematics. Their presence is just assumed. I think you'll find that there is continuity between Pin 12 and earth even though you can't see the connection to the PCB ground plane. Have a good day tomorrow Colin!

Alan

[SiriusHardware]

Love is: Making a 140 mile round trip to repair your partner's car key.

Anyway, this development concerning no 0V to UD9 sounds exciting as that is a critical PROM, the one which the computer jumps into when it first starts up. No 0V to that socket would prevent the Slothie EPROM from working, would prevent the original UD9 from working as well.

If you are absolutely sure the connection from UD9 pin 12 to 0V is missing, then as has been suggested, try soldering a wire from the IC socket pin/pad on the underside to a known 0V point nearby.

[ajgriff]

Before going too far with this can we just check whether or not that the discontinuity arises between the pin & the socket or the socket & the board?

The other thing I’m scratching my head over is how earth discontinuity at this pin might affect the operation of the Tynemouth board with its on-board RAM and ROM fully enabled?

Originally Posted by SiriusHardware

Love is: Making a 140 mile round trip to repair your partner's car key.

As we already know you truly have the patience of the proverbial saint.

Alan

[SiriusHardware]

Before going too far with this can we just check whether or not that the discontinuity arises between the pin & the socket or the socket & the board?

Yes, good point. That socket has already been replaced once, as I recall.

The TB uses its own internal data connections to its onboard ROM and RAM and only tries to use the actual (motherboard) databus when it wants to communicate with elements which are not on the TB.

If there is a dead / unpowered EPROM sitting on the motherboard databus, that could affect the TB's ability to write to and read from elements on the mainboard.

I suppose the simple thing to try there is to remove whatever chip is in the UD9 socket and see if the TB can then run the system any better, but if there really is no 0V to UD9 then it would be best to resolve that known fault anyway.

[SiriusHardware]

I missed this:

1) what voltage should the pins be showing when powered on with the NOP device fitted?

That's an interesting question, depending on how you are measuring it. With the NOP test running (which generates a roughly 50/50 squarewave on each address pin), and using a meter to measure the voltages on the address pins, you would expect to see a voltage about half way between logic '1' level and logic '0' level because the meter is not fast enough to track the rapid changes from logic 0 to logic 1 and back.

The scope will obviously show you the real 'on' and 'off' voltage values.

[ajgriff]

Originally Posted by SiriusHardware

That socket has already been replaced once, as I recall.

.... plus PROMs have been removed and replaced in it on countless occasions. Wouldn't be a great surprise if it were getting a little stressed.

Alan

[ScottishColin]

OK - a few answers with a bit of progress I think.

Pin 12 (both from the bottom of the socket and the top of any chip inserted into the socket) has continuity to one leg of the capacitor which is through a trace on the underside of the motherboard.

The other leg of the capacitor has no traces so I am not sure where that is supposed to go to?

If I insert a wire between pin 12 socket of UD9 and pin 12 socket of UD6, I get the attached screen.

Although the keyboard works, it does strange things - you can see on the second photo some random characters being inserted even though all I have done is press the return key.

Installing the Tynemouth board with the Pin 12 bridge wire in place also works better - see third photo, but again, pressing the return key gives me strange results and extra characters.

Is there any way with my equipment that I can test the capacitor in situ?

Colin.

[ScottishColin]

Just spotted this. I will try again from the capacitor to another pin 12.

Colin.

Originally Posted by Mark1960

Then you might have found a fault. Add a wire link from the capacitor pin that connects to UD9 pin 12 and pin 12 of one of the other proms.

[ajgriff]

Originally Posted by ScottishColin

Pin 12 (both from the bottom of the socket and the top of any chip inserted into the socket) has continuity to one leg of the capacitor which is through a trace on the underside of the motherboard.

The other leg of the capacitor has no traces so I am not sure where that is supposed to go to?

If I insert a wire between pin 12 socket of UD9 and pin 12 socket of UD6, I get the attached screen.

Is there any way with my equipment that I can test the capacitor in situ?

Taking these points in turn:

- Good as this means that the socket is ok.

- The other leg should connect to pin 24 ie, the 5V supply line. Check continuity (PET off) between 'the other leg' and pin 24 and/or check the voltage (PET on) with respect to earth on 'the other leg' which should be 5V.

- Check continuity between the wire and any convenient earth.

- It's extremely unlikely that this capacitor has failed. Best to perform the above checks before considering the possibility.

Otherwise there are signs of positive progress here.

Alan

[ajgriff]

Originally Posted by ScottishColin

Just spotted this. I will try again from the capacitor to another pin 12.

Colin.

I thought you'd just done that? Connect the wrong end of the capacitor and you might short the 5V line - fireworks!

Alan

[ScottishColin]

So I popped a wire in the UD9 and UD6 pin 12 sockets - will that do the job?

Don't really want to break anything else.....

Colin.

Originally Posted by ajgriff

I thought you'd just done that? Connect the wrong end of the capacitor and you might short the 5V line - fireworks!

Alan

[SiriusHardware]

As AJ says, one end of the capacitor should show continuity to 0V, the other end of the capacitor should show continuity to +5V (PROM pin 24) even if you can't see how it gets there.

Pin 12 of the PROM and one end of the capacitor should both be connected to 0V. You need to make sure this is definitely the case, by adding a repair link if necessary, before moving on to find other problems - but well done spotting that problem.

There is a 'rule' which says that the first thing anyone should ever do when starting work on something is to check all supply voltages - we all know it, but we all too often skip that step or at least minimise it. In our defence there is also the balance of probability - why would a physical IC power connection which was OK before suddenly not be OK? It obviously must have been hanging by a thread and could explain why the machine worked / didn't / worked / didn't. Thankfully, it stayed faulty (or 'mittent', as a colleague of mine used to say), for long enough for you to find it.

If we had asked you to check the supplies to all of the ICs and not just the RAMs earlier on, you would have been down on that fault a lot quicker, although it would have seemed like a lot of unnecessary, tedious slog at the time.

After you have fixed the 0V fault local to UD9 I think it might be as well to check continuity of UD9 pin 12 to the 0V pins of some of the other ICs, especially the large ICs, UC5 / UC6 / UC7, because we don't know what the exact physical nature of the failure of the 0V feed to UD9 pin 12 is. It may be that the track carrying 0V from there onwards to other ICs is still broken / open even after you restore the 0V connection to UD9 pin 12.

[ScottishColin]

I'll do this - what is confusing me is the physical track from pin 12 to the capacitor is obvious, but there is nothing obvious that I can see that connects earth to pin 12 on UD9 or the capacitor.

All tracing I have done before I have been able to see the connectivity. Not this time. Is this just a multi-layer PCB that has tracks I will never see?

Colin.

Originally Posted by SiriusHardware

As AJ says, one end of the capacitor should show continuity to 0V, the other end of the capacitor should show continuity to +5V (PROM pin 24) even if you can't see how it gets there.

Pin 12 of the PROM and one end of the capacitor should both be connected to 0V. You need to make sure this is definitely the case, by adding a repair link if necessary, before moving on to find other problems - but well done spotting that problem.

There is a 'rule' which says that the first thing anyone should ever do when starting work on something is to check all supply voltages - we all know it, but we all too often skip that step or at least minimise it. In our defence there is also the balance of probability - why would a physical IC power connection which was OK before suddenly not be OK? It obviously must have been hanging by a thread and could explain why the machine worked / didn't / worked / didn't. Thankfully, it stayed faulty (or 'mittent', as a colleague of mine used to say), for long enough for you to find it.

If we had asked you to check the supplies to all of the ICs and not just the RAMs earlier on, you would have been down on that fault a lot quicker, although it would have seemed like a lot of unnecessary, tedious slog at the time.

After you have fixed the 0V fault local to UD9 I think it might be as well to check continuity of UD9 pin 12 to the 0V pins of some of the other ICs, especially the large ICs, UC5 / UC6 / UC7, because we don't know what the exact physical nature of the failure of the 0V feed to UD9 pin 12 is. It may be that the track carrying 0V from there onwards to other ICs is still broken / open even after you restore the 0V connection to UD9 pin 12.

[ScottishColin]

So - the 'earth' end of the capacitor has continuity to UD9 pin 12 and UD8 pin 12 (which is a ROM with no wired temporary connection to UD9) so that looks OK.

The 5V end of the capacitor has connectivity to pin 24 of UD9 and other ROM chips' pin 24 too.

I have also checked pin 12 UD9 to earth and that's all good with the temporary wire in place.

Colin.

Originally Posted by ajgriff

Taking these points in turn:

- Good as this means that the socket is ok.

- The other leg should connect to pin 24 ie, the 5V supply line. Check continuity (PET off) between 'the other leg' and pin 24 and/or check the voltage (PET on) with respect to earth on 'the other leg' which should be 5V.

- Check continuity between the wire and any convenient earth.

- It's extremely unlikely that this capacitor has failed. Best to perform the above checks before considering the possibility.

Otherwise there are signs of positive progress here.

Alan

[ScottishColin]

With the temporary wire between UD12 and UD6/12 I have checked UD9 pin 12 to

UD8 pin 12
UD7 pin 12
UD6 pin 12

UD4 pin 1 and pin 21
UD5 pin 1
UD6 pin 1
UD7 pin 1

All good.

Colin.

Originally Posted by SiriusHardware

As AJ says, one end of the capacitor should show continuity to 0V, the other end of the capacitor should show continuity to +5V (PROM pin 24) even if you can't see how it gets there.

Pin 12 of the PROM and one end of the capacitor should both be connected to 0V. You need to make sure this is definitely the case, by adding a repair link if necessary, before moving on to find other problems - but well done spotting that problem.

There is a 'rule' which says that the first thing anyone should ever do when starting work on something is to check all supply voltages - we all know it, but we all too often skip that step or at least minimise it. In our defence there is also the balance of probability - why would a physical IC power connection which was OK before suddenly not be OK? It obviously must have been hanging by a thread and could explain why the machine worked / didn't / worked / didn't. Thankfully, it stayed faulty (or 'mittent', as a colleague of mine used to say), for long enough for you to find it.

If we had asked you to check the supplies to all of the ICs and not just the RAMs earlier on, you would have been down on that fault a lot quicker, although it would have seemed like a lot of unnecessary, tedious slog at the time.

After you have fixed the 0V fault local to UD9 I think it might be as well to check continuity of UD9 pin 12 to the 0V pins of some of the other ICs, especially the large ICs, UC5 / UC6 / UC7, because we don't know what the exact physical nature of the failure of the 0V feed to UD9 pin 12 is. It may be that the track carrying 0V from there onwards to other ICs is still broken / open even after you restore the 0V connection to UD9 pin 12.

[SiriusHardware]

It won't be a multi-layer PCB, only two sided, but the 'other' connection to the capacitor may well be hidden under the IC socket - it could approach it either from the top side or the bottom side.

Now that you have restored the UD9 pin 12 to 0V connection, could you check for good continuity from there to the 0V pins of other ICs, especially UC5, UC6, UC7 on all of which pin 1 is the 0V pin.

Edit: Crossed with Colin's last post. I don't see UC5, UC6, UC7 in the 'tested to' list but maybe that was just a typo.

[ScottishColin]

Apologies - yes it's a typo. Tested to the 6502, 6522 and bothe 6520s.

Colin.

Originally Posted by Mark1960

Then you might have found a fault. Add a wire link from the capacitor pin that connects to UD9 pin 12 and pin 12 of one of the other proms.

Originally Posted by SiriusHardware

It won't be a multi-layer PCB, only two sided, but the 'other' connection to the capacitor may well be hidden under the IC socket - it could approach it either from the top side or the bottom side.

Now that you have restored the UD9 pin 12 to 0V connection, could you check for good continuity from there to the 0V pins of other ICs, especially UC5, UC6, UC7 on all of which pin 1 is the 0V pin.

Edit: Crossed with Colin's last post. I don't see UC5, UC6, UC7 in the 'tested to' list but maybe that was just a typo.

[ajgriff]

I think we're now back to where we were at post #187 which is good news. With the ROM/RAM board fully active and the wire link in place are there any other signs of character corruption apart from the cursor itself? Broadly speaking is the PET behaving itself aside from the cursor issue? Just want to clarify the position before thinking about causes and verifying the integrity of the 5V supply.

Alan

[SiriusHardware]

Also try running the Slothie and Daver2 test PROMs now that the UD9 socket has power.

Daver2's test code (Place in UD8 position with original PROM in UD9 position) has a keyboard test near the end, and I think Slothie's may have too.

[ScottishColin]

OK. Slothie's PROM works. The keyboard testing in there works too.

Daver2 PROM gets struck at the caharacter testing screen and does not progress (see photo).

Boot with all PROMs inserted dumps me straight to the monitor.

So half a step forwards I think.

Colin.