UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Computers (https://www.vintage-radio.net/forum/forumdisplay.php?f=16)
-   -   Non-working Commodore PET 3016 (https://www.vintage-radio.net/forum/showthread.php?t=174829)

ScottishColin 12th Mar 2021 11:44 pm

Re: Non-working Commodore PET 3016
 
So I'll start with UE7/8 tomorrow but try to reuse UF8/9 if at all possible?

Colin.

Quote:

Originally Posted by SiriusHardware (Post 1352596)
Yes, sockets for those too. You only want to have to desolder the video RAM ICs once. This is one case where I might suggest conventional sockets rather than turned-pin because the pins of the desoldered ICs (should you need to desolder them) will be quite short and will have a thin layer of solder on them.

They may be easier to plug into conventional sockets than into turned pin sockets, if you need to refit them for any reason.


Mark1960 12th Mar 2021 11:45 pm

Re: Non-working Commodore PET 3016
 
For the effort required to try and test the video ram compared to replacing the buffers and ram, I think Iím close With Sirius for the next steps.

Remove and socket UE7 and UE8, but do not fit the buffers.

Check the video pattern again without the buffers.

If the video pattern is random then fit the buffers to the sockets.

If the video pattern is not random, remove and socket UF7 and UF8 and replace with 2114 or equivalent.

Iím being a bit overcautious, but donít think we should fit new buffers if the ram is still suspect.

Iíd recommend cricklewood for the 2114s, but I could also send a set from my spares box, though I donít have a system to test them in first and its a slow trip from Canada.

SiriusHardware 12th Mar 2021 11:51 pm

Re: Non-working Commodore PET 3016
 
Quote:

try to reuse UF8/9 if at all possible?
Don't remove them until we see what the data from them looks like with UE7 / UE8 removed from the video databus. If it springs to life, you may not need to remove or replace them at all.

If you do eventually remove them you'll be removing them because we have reasonable grounds to believe they are faulty. Removing them in one piece would enable you to test them independently using an Arduino sketch or some other kind of tester for that purpose so you can definitely say that they are or are not faulty, but it's up to you decide whether that is important to you.

Cutting them up in the usual way is certainly less likely to damage the board than trying to get them out intact.

ScottishColin 12th Mar 2021 11:53 pm

Re: Non-working Commodore PET 3016
 
Ok. UE7/8 removal tomorrow. New sockets and test without ICs inserted as the first step.

Thanks.

Colin.

ajgriff 13th Mar 2021 9:16 am

Re: Non-working Commodore PET 3016
 
1 Attachment(s)
Doesn't the disappearing @ and ; screens suggest some kind of intermittency? Before Colin reaches for the soldering iron again I would have thought it's worth checking/rechecking continuity between relevant pins amongst this group of ICs.

Ultimately if it's considered necessary to replace the video RAMs these look like a good option, as highlighted by Colin earlier:

https://www.ebay.co.uk/itm/TMM314APL...IAAOSw7FBgMmcK

Colin's photo is attached for comparative purposes.

Alan

SiriusHardware 13th Mar 2021 9:30 am

Re: Non-working Commodore PET 3016
 
Colin is very good / meticulous at doing continuity checks should the need arise, I believe he does thorough checks every time he removes an IC and fits a socket, to make sure he has not inadvertently damaged any tracks or vias in the process.

The intermittent / unpredictable effects seen were all with the original UF9 (suspected faulty) fitted or with no UF9 fitted. Since a new UF9 device was fitted I think the output has been constant chequerboard, which fits with the earlier observation that a constant 11111111 state is present on the SD0... SD7 lines. No doubt Colin can confirm or correct me when he surfaces.

Those ICs certainly aren't badly priced, good spot. I was resigned to them being four times that price based on recent experience of buying slightly older SRAMs.

ajgriff 13th Mar 2021 9:51 am

Re: Non-working Commodore PET 3016
 
Quote:

Originally Posted by SiriusHardware (Post 1352654)
The intermittent / unpredictable effects seen were all with the original UF9 (suspected faulty) fitted or with no UF9 fitted. Since a new UF9 device was fitted I think the output has been constant chequerboard ....

Fair comment and I think you're right. However, it does seem strange that when the old UF9 was refittted the @s emerged. Also at one point earlier (before UF9s removal) Colin detected the ; code. Perhaps these anomalies were indeed caused by a faulty UF9. Difficult to be sure.

Alan

ScottishColin 13th Mar 2021 12:23 pm

Re: Non-working Commodore PET 3016
 
I've checked all the SDx outputs from UF9 and the respective inputs on UF10 and they're all high.

Is that the right check?

Colin.

Quote:

Originally Posted by SiriusHardware (Post 1352654)
Colin is very good / meticulous at doing continuity checks should the need arise, I believe he does thorough checks every time he removes an IC and fits a socket, to make sure he has not inadvertently damaged any tracks or vias in the process.

The intermittent / unpredictable effects seen were all with the original UF9 (suspected faulty) fitted or with no UF9 fitted. Since a new UF9 device was fitted I think the output has been constant chequerboard, which fits with the earlier observation that a constant 11111111 state is present on the SD0... SD7 lines. No doubt Colin can confirm or correct me when he surfaces.

Those ICs certainly aren't badly priced, good spot. I was resigned to them being four times that price based on recent experience of buying slightly older SRAMs.


ScottishColin 13th Mar 2021 12:26 pm

Re: Non-working Commodore PET 3016
 
For the record, no IC in UF9 = checkboard (that wanders a bit as per my previous post).

New UF9 = solid checkerboard

Old UF9 = inverse @ sign.

Colin.


Quote:

Originally Posted by ajgriff (Post 1352661)
Quote:

Originally Posted by SiriusHardware (Post 1352654)
The intermittent / unpredictable effects seen were all with the original UF9 (suspected faulty) fitted or with no UF9 fitted. Since a new UF9 device was fitted I think the output has been constant chequerboard ....

Fair comment and I think you're right. However, it does seem strange that when the old UF9 was refittted the @s emerged. Also at one point earlier (before UF9s removal) Colin detected the ; code. Perhaps these anomalies were indeed caused by a faulty UF9. Difficult to be sure.

Alan


ScottishColin 13th Mar 2021 12:32 pm

Re: Non-working Commodore PET 3016
 
OK - UE7 and UE8 are out - no sockets or replacements fitted yet.

Screen is a solid checkboard with both removed.

Colin.

julie_m 13th Mar 2021 1:12 pm

Re: Non-working Commodore PET 3016
 
1 Attachment(s)
With UF9 out, try connecting each of pins 2, 19, 5, 16, 6, 15, 9, 12 to 0V via a 470 ohm resistor. If the character generating circuit is behaving properly, each of these should produce a screen full of a different character; they will be hex codes FE, FD, FB, F7, EF, DF, BF and 7F respectively. The first 7 will all look quite different from the "default" checkerboard; with UF9 pin 12 to 0V, the pattern will be a checkerboard, but with the dark and light squares swapped over.

See this, borrowed from https://www.masswerk.at/nowgobang/2020/petscii:
Attachment 229059

ScottishColin 13th Mar 2021 2:12 pm

Re: Non-working Commodore PET 3016
 
With UE7 and UE8 back in place?

Colin.


Quote:

Originally Posted by julie_m (Post 1352739)
With UF9 out, try connecting each of pins 2, 19, 5, 16, 6, 15, 9, 12 to 0V via a 470 ohm resistor. If the character generating circuit is behaving properly, each of these should produce a screen full of a different character; they will be hex codes FE, FD, FB, F7, EF, DF, BF and 7F respectively. The first 7 will all look quite different from the "default" checkerboard; with UF9 pin 12 to 0V, the pattern will be a checkerboard, but with the dark and light squares swapped over.

See this, borrowed from https://www.masswerk.at/nowgobang/2020/petscii:
Attachment 229059


ScottishColin 13th Mar 2021 2:13 pm

Re: Non-working Commodore PET 3016
 
all -new UE7/UE8 in place - checkerboard.

Colin.

SiriusHardware 13th Mar 2021 2:56 pm

Re: Non-working Commodore PET 3016
 
Julie wants you try applying logic '0' to each of the UF9 output socket pins in turn (with UF9 removed) to make sure that when UF9 puts out different codes different characters are shown. I think we can already say that happens, up to a point, because with the (apparently) faulty original UF9 in place you get different characters generated, and when you have no UF9 in you get sometimes checkerboard, sometimes random, depending on what you poke with your fingers. That is because with UF9 removed the input pins of UF10 are not held in any specific state. However, do try the test Julie suggested.

The characters generated might not be entirely predictable because while you are holding one pin low the other (unconnected) pins will still be free to do as they like, although they seem to prefer to stay in the high state most of the time.

Back to the main plot, am I correct in reading that with the new UF9, UE7, UE8 all fitted you still have a checkerboard pattern on the screen?

If so, then unfortunately the next item of interest is the video RAM itself.

ScottishColin 13th Mar 2021 3:01 pm

Re: Non-working Commodore PET 3016
 
Correct.

Colin.

Quote:

Originally Posted by SiriusHardware (Post 1352775)
Back to the main plot, am I correct in reading that with the new UF9, UE7, UE8 all fitted you still have a checkerboard pattern on the screen?

If so, then unfortunately the next item of interest is the video RAM itself.


SiriusHardware 13th Mar 2021 3:06 pm

Re: Non-working Commodore PET 3016
 
So it's up to you now whether you try to get the Video RAMs out intact so they can be independently tested, or just cut them up and hope that a replacement set will fix the display chain problem.

If you choose the former I am sure there is an already existing Arduino sketch for testing 2114 type RAMs, although it might be based on the UNO rather than the Mega. The problem you have is that if the results are bad, is that because the RAM is bad or because the tester, being used for the first time, is not working? Hard to know without a known good 2114 to test the tester with.

ScottishColin 13th Mar 2021 3:34 pm

Re: Non-working Commodore PET 3016
 
This all checks out - moving the resistor changes the characters on the screen.

But I think they're actually the inverse of the codes you have listed. Would it be possible for you to let me know which combination of pins I should use to get something 'sensible' - eg the letter A and I can try that to see if they are inverted or not?

Colin.

Quote:

Originally Posted by julie_m (Post 1352739)
With UF9 out, try connecting each of pins 2, 19, 5, 16, 6, 15, 9, 12 to 0V via a 470 ohm resistor. If the character generating circuit is behaving properly, each of these should produce a screen full of a different character; they will be hex codes FE, FD, FB, F7, EF, DF, BF and 7F respectively. The first 7 will all look quite different from the "default" checkerboard; with UF9 pin 12 to 0V, the pattern will be a checkerboard, but with the dark and light squares swapped over.

See this, borrowed from https://www.masswerk.at/nowgobang/2020/petscii:
Attachment 229059


ScottishColin 13th Mar 2021 3:37 pm

Re: Non-working Commodore PET 3016
 
Just to confirm, this is UF7 and UF8. I'll try to remove them intact, but I won't be reporting back until tomorrow.

I only have 'normal' 18-pin sockets, not turned pin - did you say thet would be preferable anyway?

Colin.


Quote:

Originally Posted by SiriusHardware (Post 1352777)
So it's up to you now whether you try to get the Video RAMs out intact so they can be independently tested, or just cut them up and hope that a replacement set will fix the display chain problem.

If you choose the former I am sure there is an already existing Arduino sketch for testing 2114 type RAMs, although it might be based on the UNO rather than the Mega. The problem you have is that if the results are bad, is that because the RAM is bad or because the tester, being used for the first time, is not working? Hard to know without a known good 2114 to test the tester with.


ajgriff 13th Mar 2021 4:17 pm

Re: Non-working Commodore PET 3016
 
In view of the relatively low cost of the replacement RAMs (if you order from your eBay source) I wouldn't risk unnecessary board damage by trying to remove the chips whole, in which case you might as well order turned pin sockets. Your choice of course Colin.

The video RAMs certainly look like prime suspects now, he says whilst keeping his fingers firmly crossed. A random character screen would be a great step forward but the READY prompt would be even better! Good luck.

Alan

SiriusHardware 13th Mar 2021 5:48 pm

Re: Non-working Commodore PET 3016
 
If you already have TP sockets just go with those. The original advice to go with conventional sockets was to let you have the option of plugging the original ICs back in, as 'second hand' ICs are usually easier to use in conventional sockets than in turned-pin ones.

Desoldered ICs can be used in TP sockets, but you have to get the pins really clean / clear of solder first.

Yes, let's cross everything and hope that a new pair of video RAMs moves us a step further forwards.


All times are GMT. The time now is 1:43 pm.

Powered by vBulletin®
Copyright ©2000 - 2021, vBulletin Solutions, Inc.
Copyright ©2002 - 2021, Paul Stenning.