PDA

View Full Version : PET 2001-8 Swap similar motherboards for testing?



snuci
August 18th, 2014, 02:32 PM
I recently acquired a PET 2001-8 with the fuse holder missing. The seller didn't know much about the machine (it was her brothers who left it in her basement) so I took a shot at it. I finally found a fuse holder (not similar to any other computer or PET I had) and fired it up. Below is a picture of the output. When the monitor fired up, it sounded like the degauss sound of a Sony monitor. I have another PET 2001-8 and I don't remember it doing that. The screen actually looks like it might be correct in terms of character place holders (with odd characters replacing them) and even has a flashing cursor but typing anything in does not move the cursor.

19953

Anyhow, the motherboard is a 320008 and I'd rather not mess with it until I have to. I have not reseated the chips as of yet. As I mentioned, I have another PET 2001-8 that has a 320137 motherboard. For troubleshooting purposes, can I swap motherboards and see if the older board works in the newer case and vice-versa? I don't think it should be a problem but I want to be sure. The power connector on J8 is slightly different and has an extra cable that doesn't appear to be connected on the working PET (it's odd and has a different longer connector). If so, I'll do it this weekend and see what happens.

I did some searching and couldn't find a similar issue. I also played with the vertical adjustment pot in the back of the monitor but that just moved what is displayed up and down. Any help is appreciated.

MikeS
August 18th, 2014, 03:33 PM
Interesting; AFAIK all four versions of the 2001-8 had identical power supplies and connectors, 5 pins arranged symmetrically so you could plug it in either way:

Transformer to the two outside pins (1 & 5), ground to the centre pin (3), filter cap positive to pins 2 & 4 and negative to ground (pin 3)

Can you describe the two connectors and specifically what is different?

snuci
August 18th, 2014, 04:12 PM
The earlier PET has the normal 5 pin molex connector. The newer 2001-8 ("Professional Computer" with chicklet keyboard) has a nine pin connector with six wires attached. Five are connected to the normal 5 pin connector on the motherboard. The last wire is not connected. A pic is below:
19954

It does seem kinda wasteful (for Commodore) using a nine pin connector that has a sixth wire that would never be connected to anything. It's black and the other end is connected to ground as is pin 3.

snuci
August 19th, 2014, 02:42 AM
A little more info on this one...

Last night I tried one last time to see if I was, indeed, getting a cursor (it appears like a flashing underline) and tried all of the keys. Turns out that two of the keys work and I can type! I tried several lines and the output got a little longer but my Enter key didn't work and there's no key repeat so after pounding two keys for a few minutes, I stopped. After I would say five lines or so, the output grew to an inch. Here's a pic (sorry about the lighting)

19969

I'll definitely try the swap the motherboards between the two PETs and see what happens. That should tell me if it's the motherboard or the monitor portion. I also have a good chicklet keyboard on the other machine so I can experiment to see how big the screen output actually is if the motherboard acts the same.

Sadly, the chicklet keyboard on this PET is missing one key on the numeric keypad. What are the odds that someone has either a spare key or even keyboard? I'm going to see what info I can find on cleaning a chicklet keyboard. Anyway, I thought I'd post the update. Thanks for looking.

MikeS
August 19th, 2014, 08:14 AM
The earlier PET has the normal 5 pin molex connector. The newer 2001-8 ("Professional Computer" with chicklet keyboard) has a nine pin connector with six wires attached. Five are connected to the normal 5 pin connector on the motherboard. The last wire is not connected. A pic is below:
19954

It does seem kinda wasteful (for Commodore) using a nine pin connector that has a sixth wire that would never be connected to anything. It's black and the other end is connected to ground as is pin 3.I don't think that's original; looks like someone just replaced the 5 pin connector with whatever he/she had on hand. The original connectors were prone to overheating and burning, so it's not unusual to see a replacement.

FWIW, it's probably the monitor (although the main board and keyboard may also have issues).

Good luck.

BTW, I didn't hear back on the disk controller deal...

snuci
August 19th, 2014, 09:45 AM
Hi Mike,

Thanks for the response. I really hope it's not the monitor. I haven't worked on video boards before and that 9:" monitor is pretty cramped in there :)

I'll definitely let you know how it works out.

BTW, you've got a long winded email on the way :)

dave_m
August 19th, 2014, 10:17 AM
I really hope it's not the monitor. I haven't worked on video boards before and that 9:" monitor is pretty cramped in there :)



If you have a scope or frequency counter, you might check the Vertical Drive signal for 60 Hz. If you have it, then the problem like Mike indicated is probably on the video board.

MikeS
August 19th, 2014, 02:43 PM
... I really hope it's not the monitor. I haven't worked on video boards before and that 9:" monitor is pretty cramped in there :) Yeah, you pretty well have to remove the tube and the board from the case to really work on it; I've got one in that state waiting for me on my bench... ;-)

snuci
August 19th, 2014, 03:23 PM
Good news and bad news. Good news: I swapped motherboards and the video is fine so the monitor is working great. Bad news: The motherboard exhibited the exact same video results on the new monitor. Other things of note: both monitors make that degaussing sound and the older PET monitor is white while the newer one is green. I thought those white monitors were only on the blue screen surround PETs?

I'm back to the drawing board but it's definitely the motherboard.

MikeS
August 19th, 2014, 03:58 PM
Sadly, the chicklet keyboard on this PET is missing one key on the numeric keypad. What are the odds that someone has either a spare key or even keyboard? I'm going to see what info I can find on cleaning a chicklet keyboard. Anyway, I thought I'd post the update. Thanks for looking.I sold a spare one that I had, but I think a couple of TPUG members still have some; if I remember I'll ask at the next meeting if anybody wants to part with one. Which key is missing?

MikeS
August 19th, 2014, 04:05 PM
Good news and bad news. Good news: I swapped motherboards and the video is fine so the monitor is working great. Bad news: The motherboard exhibited the exact same video results on the new monitor. Other things of note: both monitors make that degaussing sound and the older PET monitor is white while the newer one is green. I thought those white monitors were only on the blue screen surround PETs?Nothing is certain when it comes to PET models and configurations; it seems that sometimes the factory just used whatever they had on hand, and of course folks like you have also swapped a lot of parts and modules in the meantime.

You might try turning out the lights and watching the back of the monitor while you turn it on, in case there's any visible arcing anywhere. The power transformer also often hums for a moment when you turn on the power.

Did you try the bad MB/good monitor with a good keyboard, i.e. does it look like it basically works except for the height problem?

snuci
August 19th, 2014, 04:20 PM
I sold a spare one that I had, but I think a couple of TPUG members still have some; if I remember I'll ask at the next meeting if anybody wants to part with one. Which key is missing?

Keypad "/" key. I would need the stock, key and I'm guessing a spring? I just cleaned the keyboard and it's pretty simple to clean up. There was a thin layer of dust on it. When I figure out how to fix the video, I'll give it a shot. Thanks Mike. I appreciate it.

snuci
August 19th, 2014, 04:35 PM
Did you try the bad MB/good monitor with a good keyboard, i.e. does it look like it basically works except for the height problem?

That is correct. If I type in a bunch of characters and hit enter on the good keyboard, I probably get a syntax error. The screen is never more than an inch after typing and having errors come up.

As a test, I just tried to load a little "Hello World!" looping program from cassette? No problem. The cassette stopped when the program finished loading. When I ran it I saw this:

19976

Seems to be working great except for the video.

snuci
August 19th, 2014, 04:48 PM
In addition to the above, it looks like I am only getting one line of the 7 or 8 pixel lines of the characters when I type. What I found odd was that if I typed the 2 key, I could be getting the highest line of pixels in the number 2 (I can't really tell) but if I went to the next line by hitting return, I would only get a dot or two when typing 2. Maybe that was the second highest line of pixels. Next line with the same key was a different pattern of dots (maybe the next line of pixels or a single dot). Oddly enough, the above example looks like well formed characters but I think it's showing one line of pixels in an alternate fashion from each letter on each of the 24 lines (that makes up an inch or so). If there are seven lines of pixels per character, it's showing one line of each "Hello World!" then the next line of pixel on the next line and so on until you see 24 lines of single row pixels. I hope I'm describing this correctly.

Maybe it's a character generator chip? Just a theory.

KC9UDX
August 19th, 2014, 05:18 PM
If the problem is vertical, it's not the character ROM. It's probably the glue logic around it.

Something appears to be wrong with the video hardware on the logic board. From what little I can see at this distance, unfortunately I can't narrow it down much better than that. You may well just have bad video RAM.

Do you have the schematics, and are you able to read them?

If I were you, I'd first socket and replace the video RAM, then build a breakout board for the character ROM, and work from there; but I'm not you. And, unless you have something to drive the circuitry, that doesn't help you much.

Here's my preferred way to diagnose PET video:
1998119980
Hardly ideal, but it sure works.

MikeS
August 19th, 2014, 07:43 PM
...Maybe it's a character generator chip? Just a theory.Not likely; you're not putting characters where they're supposed to be, so it doesn't matter what the characters are.

If you crank the brightness pot all the way up, do you see a full screen (blank, except for your garbled text line) with diagonal lines?

Do you have a 'scope?

snuci
August 20th, 2014, 03:35 AM
You may well just have bad video RAM. If I were you, I'd first socket and replace the video RAM, then build a breakout board for the character ROM, and work from there; but I'm not you. And, unless you have something to drive the circuitry, that doesn't help you much.

I haven't yet done one of the first things you should do when troubleshooting. I haven't reseated the socketted chips. Those irreplaceable ROM and RAM chips have me a little worried on this one but that's what I'm going to do next. Yesterday I eliminated the monitor but I didn't have time to reseat the chips. Today I'll reseat everything and see what develops.


If you crank the brightness pot all the way up, do you see a full screen (blank, except for your garbled text line) with diagonal lines?

I did that this morning before heading to work. It's very difficult to tell on this monitor but after inspecting closely, I'm pretty sure it's only the one inch that is located where the Hello World! example text is as noted in the pic above.


If you have a scope or frequency counter, you might check the Vertical Drive signal for 60 Hz.

I'm guessing vertical drive would affect the screen height? This sounds like a good possibility but I don't know how to check for vertical drive.

I do have a scope but I confess that I don't know how to use it yet. In diagnosing a KIM-1, where I thought I was supposed to get a square wave, I was seeing a sine wave so I put that on hold until I could learn how to use it properly. Haven't had the chance yet to figure it out.

I also have the schematics and can use them to locate ICs, check continuity and check noted voltages but that's the limit of my knowledge.

I do appreciate everyone's help. Tonight I'll reseat the chips and see if that makes a difference.

MikeS
August 20th, 2014, 08:14 AM
While you're at it you could try replacing the 6522 and the 6520 at G8 with known good ones from the other board, but it kinda looks like the problem might be somewhere in that B6/C6/D8 area at the top middle of sheet 3. If you can have both PETs running at the same time you could use the scope to compare the various signals in that area.

dave_m
August 20th, 2014, 08:57 AM
I'm guessing vertical drive would affect the screen height? This sounds like a good possibility but I don't know how to check for vertical drive.

I do have a scope but I confess that I don't know how to use it yet. In diagnosing a KIM-1, where I thought I was supposed to get a square wave, I was seeing a sine wave so I put that on hold until I could learn how to use it properly. Haven't had the chance yet to figure it out.

The Vertical Drive signal is a standard +5V signal located at chip D8-pin11 on sheet 3 of the 320008 assembly schematic (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif). At the top right of sheet 3, you can also see it is at connector J7- pin3. J7 is the signal connector that goes to the monitor. Note that J7-pin6 is missing and is used to properly orient the connector. The waveform is a skinny negative going pulse of about 2 mS and the pulse period is about 16.6 mS (60Hz). With power off, remove the J7 and hook the scope to pin 3 on the board. Power up and check the waveform. The monitor of course will be dark. You may see a dot for a second in the middle of the screen. When finished, power off and properly reconnect the J7 connector. Before removing any connector, it may be good to take a photo of the connector mated on on the board or mark pin 1 to make sure you reconnect properly.

As for the sine wave, you should be using a scope with a bandwidth of at least 5 MHz and usually about 40 MHz and use a standard 10X probe (attenuation of 10) to see a good square wave on a 1 MHz clock. Or are you using one of the USB digital sampling gadgets that hook up to a computer? In that case the sampling rate is probably not good enough to properly see high frequency signals.
-Dave

dave_m
August 20th, 2014, 09:10 AM
While you're at it you could try replacing the 6522 and the 6520 at G8 with known good ones from the other board, but it kinda looks like the problem might be somewhere in that B6/C6/D8 area at the top middle of sheet 3. If you can have both PETs running at the same time you could use the scope to compare the various signals in that area.

Hi Mike,
Right, hopefully the problem is just the generation of the Vertical timing signal and not in the screen refresh logic timing which can be messy.
-Dave

MikeS
August 20th, 2014, 12:17 PM
Hi Mike,
Right, hopefully the problem is just the generation of the Vertical timing signal and not in the screen refresh logic timing which can be messy.
-DaveAmen! I've got one that duplicates the first character (top left) about half way down the screen; really must look at that some day... ;-)

KC9UDX
August 20th, 2014, 03:07 PM
Amen! I've got one that duplicates the first character (top left) about half way down the screen; really must look at that some day... ;-)

Sounds like missing address lines at the video RAM.

snuci
August 21st, 2014, 04:05 AM
Yesterday I didn't have a lot of time but I did manage to re-seat all of the socketed chips and put it back in its original PET. That didn't help as I saw the same thing. After a good cleaning, my keyboard now works but I tried to load the "Hello World" program on this PET and the cassette didn't stop. Fast forward and rewind also don't work of the cassette is in it but do if the cassette is not there. The cassette tape is not tight at all. This is just a symptom I figured I'd mention.

I didn't check what dave_m suggested yet. I have a Tektronics 465b scope so I should be okay but there's a lot of buttons and dials and is a little intimidating for a newbie. I'll try to give it a shot tonight or tomorrow.

MikeS
August 21st, 2014, 08:10 AM
Fast forward and rewind also don't work of the cassette is in it but do if the cassette is not there. The cassette tape is not tight at all. This is just a symptom I figured I'd mention.Sounds like you need to replace the belt; is this the original white key drive or the later black key?

snuci
August 21st, 2014, 10:24 AM
Sounds like you need to replace the belt; is this the original white key drive or the later black key?

It's the black key version but I figured if play appears to work (or spin at least), it would use the belt?

KC9UDX
August 21st, 2014, 10:33 AM
What you're experiencing may be normal. A datasette connected to my C64 does exactly what you describe, and I always wondered why. PETs do not do that, from what I've seen, but maybe yours is like my C64 for some reason.

dave_m
August 21st, 2014, 01:55 PM
Those irreplaceable ROM and RAM chips have me a little worried on this one
If you have the old style ROMs and RAM (6540 and 6550), that will be a problem. However there is always the PETVet to replace all RAM and ROM in one shot.

dave_m
August 21st, 2014, 02:02 PM
Fast forward and rewind also don't work of the cassette is in it but do if the cassette is not there. The cassette tape is not tight at all. This is just a symptom I figured I'd mention.

Mike is right, replace the belt. The PET does not control the FF or REW buttons. It only switches +9V power to the cassette motor.



I didn't check what dave_m suggested yet. I have a Tektronics 465b scope so I should be okay but there's a lot of buttons and dials and is a little intimidating for a newbie. I'll try to give it a shot tonight or tomorrow.

Tektronix 465B, wow, I'll say you are in good shape. I wish I had one. I use a 453, an early predecessor to the 465. The 465 is small and bullet proof. Army tanks should be made so rugged. We can help with tips on the using the scope. Please tell me you have a set of probes. We'll let you know how to 'compensate' them for capacitance loading to get accurate square waves. There is usually a screw adjust on the side.

snuci
August 22nd, 2014, 02:32 PM
The Vertical Drive signal is a standard +5V signal located at chip D8-pin11 on sheet 3 of the 320008 assembly schematic (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif). At the top right of sheet 3, you can also see it is at connector J7- pin3. J7 is the signal connector that goes to the monitor. Note that J7-pin6 is missing and is used to properly orient the connector. The waveform is a skinny negative going pulse of about 2 mS and the pulse period is about 16.6 mS (60Hz). With power off, remove the J7 and hook the scope to pin 3 on the board. Power up and check the waveform. The monitor of course will be dark. You may see a dot for a second in the middle of the screen. When finished, power off and properly reconnect the J7 connector. Before removing any connector, it may be good to take a photo of the connector mated on on the board or mark pin 1 to make sure you reconnect properly.

On J7 Pin 3 I'm getting 4.28 volts which seems kinda low but just for clarification, I have tried a good board with this power supply and it was fine. I also see this on the scope.

20031

To me it looks like it's around 4 volts (volts/division is 2), drops down to 0 for .25 milliseconds then resumes for about 2.25 milliseconds? (time/division is .5ms). I could be reading this wrong so I provided the image.

EDIT: I did the same thing to my working PET 2001-8 and the voltage was at 4 volts measured and with the same settings, I did not see the same thing. I did see the 16ms delay and probably a drop for about 2ms like you mentioned when I readjusted the time/division to 5ms (as opposed to .5). I also saw a straight wave not like the tiny pulses on the top of the 4 volt square wave like the picture. Since my my connectors were different on this PET, I ws able to do it with video still plugged in (if it makes a difference).

As always, much appreciated.

dave_m
August 22nd, 2014, 03:45 PM
To me it looks like it's around 4 volts (volts/division is 2), drops down to 0 for .25 milliseconds then resumes for about 2.25 milliseconds? (time/division is .5ms). I could be reading this wrong so I provided the image.

OK, that would cause the 'vertically impaired' display you are getting. That is real progress. The 4 Volts is fine, anything above 3.2 V is OK.

Be patient, we will have to trace back and look at signals from the jungle of flip flops and other chips to see where the problem creeps in. It could be the JK flip flop B6 or perhaps C7, but I don't think we will be that lucky.

The root of the problem may be a bad screen address decoder output at D8 pin 6. We may have to probe a lot of signals before we find the bad chip. In a later design, Commodore would scrape all that discrete logic and replace it with a 6545 CRT controller. However at least we can fix the problem by replacing a little 14 or 16 pin chip. We just have to find it. Hang in there. We need to think like Sherlock Holmes. Old MikeS will be of great help. He is called the bloodhound of Commodore Pets. With his help we will sniff it out.
-Dave

snuci
August 22nd, 2014, 04:03 PM
Thanks dave_m.

I was about to add one more pic. I do see a couple of ps that look a little crusty in the back corner nearest the transformer. I'm not sure if this makes a difference but I figured I'd post a pic anyway. Might these matter? The same two caps on my other PET look totally fine.

20032

MikeS
August 22nd, 2014, 05:09 PM
... We just have to find it. Hang in there. We need to think like Sherlock Holmes. Old MikeS will be of great help. He is called the bloodhound of Commodore Pets. With his help we will sniff it out.
-DaveOh no; I happily cede that title to you, Dave.

As I said, it might make it a lot easier if you could have both PETs running side by side with grounds connected, and compare scope readings in the suspicious areas.

MikeS
August 22nd, 2014, 05:11 PM
Thanks dave_m.

I was about to add one more pic. I do see a couple of ps that look a little crusty in the back corner nearest the transformer. I'm not sure if this makes a difference but I figured I'd post a pic anyway. Might these matter? The same two caps on my other PET look totally fine.

20032I wouldn't worry; just looks like a little mildew on caps on the cassette drivers; should just wash off with a damp toothbrush.

dave_m
August 24th, 2014, 02:49 PM
Oh no; I happily cede that title to you, Dave.

OK, Mike, enough said, I will quit making up titles for you! ;)


As I said, it might make it a lot easier if you could have both PETs running side by side with grounds connected, and compare scope readings in the suspicious areas.

That is quite an offer which might quickly fix the PET. As you know the feedback nature of that video controller makes it difficult to find the bad chip as so many signals will look bad. I wonder if the OP can attend the next Toronto PET Users Group?

MikeS
August 24th, 2014, 03:37 PM
OK, Mike, enough said, I will quit making up titles for you! ;) Just sayin' that IMO you're the PET guru here, especially the most active and most often helpful!


That is quite an offer which might quickly fix the PET. As you know the feedback nature of that video controller makes it difficult to find the bad chip as so many signals will look bad. I wonder if the OP can attend the next Toronto PET Users Group?I've invited him a couple of times; we'll see. There are also at least a couple of local PET folks whom he could visit or vice versa, but I think he actually has a second working PET himself; AFAIK there's little or no relevant difference in the video circuitry between the -008 and the -137 boards.

snuci
August 24th, 2014, 04:07 PM
Thanks for the replies guys. I actually like to try to figure stuff like this out. I'm just not yet sure how to use a scope or what to look for. Here's what I was trying today to to try to troubleshoot with a scope and two PETs (one working, one not).

First I calibrate with the working PET and then check the bad screen output PET. Starting from J7 pin 3 on the schematic I've been using the scope to check for a similar wave between the two units. I then moved backwards so the next item was D8 pins 11, 12, 13. I checked both PETs and the signals are not the same. So I went back to the next chip which is B6. 5 volts is fine on pin 10 and 13 but scope output on pin 2 and 11 is not the same on both PETs. So onto C6...

So I am basically working backwards until I get the same signal. If I ever find the same scope output, then I am guessing that the previous chip I checked is no good. Does this make sense or am I going about this wrong?

Thanks guys for the pointers.

dave_m
August 25th, 2014, 08:00 AM
So I am basically working backwards until I get the same signal. If I ever find the same scope output, then I am guessing that the previous chip I checked is no good. Does this make sense or am I going about this wrong?


No, this will find the problem eventually. But the problem is that the Q outputs of the B6 Flip Flops, that you are starting with, are the outputs of a sequencer. The State table is shown next to the flip flops. It sequences from state 00 (Vert Drive) to state 01 (Top Blank) then state 11 (Video On) then state 10 (Bottom Blank) and then back to the beginning state 00. All this should take 16.66 mS. Everything will look wrong there. The waveform timing is wrong because the gated clocks are arriving too soon at the B6 flip flops. The clock derives from the count decoder at the bottom of the schematic at D8-pin6. The question is what is the bad chip in the maze of counters, latches and gates in the middle of the schematic that is causing the incorrect timing. I would start at D8-pin6 and work back from there. I suspect that the problem is with the D8-pin3 chain of signals as that gate decodes the long count.


Be patient and you will find the problem. Take photos of the scope displays so we can all have fun helping with the detective work.

snuci
August 30th, 2014, 06:12 AM
I understand what you are saying but I just don't know exactly what to check and what I should expect to see. I would assume that at some point, the two PET scope outputs would be the same but I've checked a few points and none of them seem to match between the two. Could the different revs make a difference?

I really like this schematic with scope outputs at various points and wish there was one for the PET 2001 portion of the schematic I'm trying to diagnose:
http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/video-1.gif

I'm not sure if this will help but could you tell me the points in the schematic I can check and I'll produce something like this for both PETs? Maybe that would help me and others who find this thread later on.

One other question. It looks like we are circling around 4 or 5 ICs here. If I should happen to have one of these, can they be piggy-backed on to try it out? It may be a stupid question but I fixed an Osborne Vixen video problem this way and it would save me from removing chips that are fine.

I appreciate the help, as always.

dave_m
August 30th, 2014, 10:11 AM
I would assume that at some point, the two PET scope outputs would be the same but I've checked a few points and none of them seem to match between the two. Could the different revs make a difference?

Not in this circuit, as Mike mentioned, all 1st generation PETs use this circuit; I assume you are looking at the schematic sheet 3 for the 320008 assembly with old style ROMs and RAMs (6540 & 6550)? The problem is that when there is a problem (bad output), a lot of the signals will look bad due to the feedback nature of the circuit.




I'm not sure if this will help but could you tell me the points in the schematic I can check and I'll produce something like this for both PETs? Maybe that would help me and others who find this thread later on.
Yes, I would like to see the following points: D8-pin1, E8-pins 4,5,2 and 1 and finally D8-2. They are all at the bottom of the schematic (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif).



One other question. It looks like we are circling around 4 or 5 ICs here. If I should happen to have one of these, can they be piggy-backed on to try it out? It may be a stupid question but I fixed an Osborne Vixen video problem this way and it would save me from removing chips that are fine.



That would be good to try, but remember if an output is stuck low then the piggyback chip might not be able to force it high. You would be looking for any signs of improvement (change) not necessarily a fixed circuit.

snuci
August 31st, 2014, 05:21 AM
Yes, I would like to see the following points: D8-pin1, E8-pins 4,5,2 and 1 and finally D8-2. They are all at the 008-3.gif"]bottom of the schematic[/URL]

Here are some pics.

D8 Pin 1 20194
D8 Pin 2 20195
E8 Pin 1 20196
E8 Pin 2 20197
E8 Pin 4 20198

snuci
August 31st, 2014, 05:23 AM
E8 Pin 5 20199

dave_m
August 31st, 2014, 09:37 AM
OK, now show the working waveforms. Please keep the horizontal time scale the same for all photos. Better yet show two periods of each waveform and tell he the time scale of each.

snuci
August 31st, 2014, 11:25 AM
This is where things look weird to me. If I was to guess by looking at the images, I would thing the "good" on isn't so go but it is. Remember, this is a different rev than the bad one (if it matters).

Here's what I got (trying to use the same calibrations)
D8 Pin 1 Good 20200
D8 Pin 2 Good 20201
E8 Pin 1 Good 20202
E8 Pin 2 Good 20203
E8 Pin 4 Good 20204

snuci
August 31st, 2014, 11:25 AM
E8 Pin 5 Good 20205

dave_m
August 31st, 2014, 12:18 PM
OK, good try, but I should have warned you that the 'good' PET waveforms were going to be at a lower frequency than the bad PET. So please pick a horizontal sweep setting that shows two periods of the slowest waveform which is at E8-pin1.

snuci
August 31st, 2014, 12:44 PM
I played with all settings and the best I could get was a waveform on D8 Pin 1. The rest looked pretty much the same.

20207

dave_m
August 31st, 2014, 01:55 PM
I'm not seeing the waveforms. Play with the trigger controls. Use Normal trigger mode and Channel 1 trigger and positive edge triggering until you get a some waveforms. You may have to increase the beam intensity and try different horizontal sweep rates. Keep the zero volt line in the middle of the screen for clarity. I noticed the vertical channel 'uncal' red light is on. Turn the red knob clockwise to calibrate the channel so the 0.2 V really means 0.2V (Note: when using a 10X probe it means 2V).

You will get the hang of the scope. You have a good one.

There is another approach we may take later. We can go down the counter outputs and check that each output is half the frequency of the previous one until we find an anomaly. The counter outputs have two uses. One is to decode the end of the vertical sweep and the other is to form screen addresses for the screen refresh logic.

MikeS
August 31st, 2014, 04:38 PM
Everything that Dave said.

This may help:
- Adjust V pos to put 0V on center line; set both the vertical gain and time/div pots so that the 'uncal' lights go out; leave vertical gain at 2v/Div (calibrated).
- Adjust the trigger controls so that you have a trace of some kind. Put the probe on a 5V signal to make sure you get a line across at 2.5 divisions.
- Turn the time/div switch from one end to the other until you see a signal (if you do; if not, is the trace at 0 or +5V )?
- If there is a signal, adjust time/div switch until you have only 2-5 blips on screen
- Adjust the trigger controls to try to make them stand still (not always possible).

snuci
September 1st, 2014, 07:36 AM
Okay, I was playing with E8 Pin 1 but never got anything but a straight line. I tried other points and got something. Here's some pics on the working PET 2001.

E8 Pin 1 20223
E8 Pin 2 20224
E8 Pin 4 20225
E8 Pin 5 20226

snuci
September 1st, 2014, 07:38 AM
Good PET 2001. I seem to get inconsistent output between tries. Maybe it's my scope?

D8 Pin 1 20227
D8 Pin 2 20228

MikeS
September 1st, 2014, 12:46 PM
Good PET 2001. I seem to get inconsistent output between tries. Maybe it's my scope?The trouble is that many of the signals you're looking at are not regularly repeating signals and with internal triggering you can easily get different images depending on which edge happens to trigger the sweep.

Suppose the signal looks like this: |_|-|_|-|_|---|_

Assuming your sweep is set to display 2 1/2 'periods' as in your first example, if it triggers (i.e. starts to draw) on the first falling edge you would see:
_|-|_|-|_
But if it triggers on the second falling edge you would see:
_|-|_|---

Since you should not see both a high and a low signal in the same time period, obviously the two images above are overlaid on successive sweeps as in your image.

But at this point we're really more concerned whether there is a signal at all and what the frequency is; I'll suggest again that you set the variable sweep control to turn off the 'uncal' light and use the rotary switch to put the display more or less back the way it was.

dave_m
September 2nd, 2014, 01:20 PM
The trouble is that many of the signals you're looking at are not regularly repeating signals and with internal triggering you can easily get different images


Mike,
Our friend is struggling with using the scope on complex signals. I was thinking that it might be easier to compare the State Outputs of the sequencer for timing differences (which is the opposite of what I first told him!). This may provide a clue as to what counter output is screwing up.

At least the signals there are periodic. The signals on the good PET will repeat every 16.667 mS (60 Hz), but the bad PET has shorter periods. By comparing timing on the top blanking, Video On, Bottom Blanking and retrace time, it may give us clues on the problem counter or decode logic.

If you agree, let's have him put channel 1 probe on B6-pin3 and channel 2 on B6-pin5, the outputs of the B6 flip flop (top of page) (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif). He should sync on channel 1, mode switch on 'alternate', trigger mode on 'channel 1 only' and A Sweep Mode to 'NORMAL Trigger' and A Triggering settings to SLOPE '+', COUPLING 'DC', SOURCE 'Internal'. For completeness I'll add Horizontal Display to 'A' and Magnification to 'Off'. You will not be using the B timebase (delayed sweep) so don't worry about those settings.

Those two signals will give us the timing of the four states. He needs to repeat this on the good PET for reference. He must set up the scope display to provide the 2 1/2 periods for each PET (3 positive pulses). The time base setting will be different on each PET). And have the scope with all "uncal" lights off (why do programmers like to tweak the little red knobs?). He should remember to provide us with time base setting (5 mS/div, etc) on each photo in case we can not read the knobs on the photo.

Mike, do you agree that this video timing logic in the early PETs is diabolical? I can't make up my mind if it is so tricky because they did not waste a single gate (overly sophisticated) or it is just plain kludgy. Knowing Commodore, I should go with "kludgy" :D
-Dave

snuci
September 2nd, 2014, 02:05 PM
If you agree, let's have him put channel 1 probe on B6-pin3 and channel 2 on B6-pin5, the outputs of the B6 flip flop (top of page) (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif). He should sync on channel 1, mode switch on 'alternate', trigger mode on 'channel 1 only' and A Sweep Mode to 'NORMAL Trigger' and A Triggering settings to SLOPE '+', COUPLING 'DC', SOURCE 'Internal'. For completeness I'll add Horizontal Display to 'A' and Magnification to 'Off'. You will not be using the B timebase (delayed sweep) so don't worry about those settings.

<sarcasm>Ya, this sounds much easier</sarcasm> :)

I really appreciate you guys helping so thanks for being patient. I don't appear to have button options with some of those names. Here ar a couple of pics. These are set for 2volts/division at 5ms

I couldn't really sync them (I don't think) but I can tell you what I saw. I have both channels going with B6 Pin 3 on Channel 1 and B6 Pin 5 on Channel 2 where applicable.

I do apologize for terrible scope skills but hopefully, this will help? The pictures are blurry because I was trying to take pics while holding onto one probe. Luckily I have one probe with a hook that I didn't need to hold.



Good Picture
Good Description
Bad Picture
Bad Description


20257
B6 Pin 3 only. It's a good solid square wave. You'll notice the line at 0. That's Channel 2
20258
B6 Pin 3 only. It is also a good square wave but a lot faster


20259
B6 Pin 3 and 5 - There was a slight flutter between the two (slight offset) but they were both the same wave.
20260
B6 Pin 3 and 5 - This also fluttere slightly with an offet but it was also both the same wave

dave_m
September 2nd, 2014, 02:35 PM
<sarcasm>Ya, this sounds much easier</sarcasm> :)
<smile> It's not as bad as it sounded....



I don't appear to have button options with some of those names. Here ar a couple of pics. These are set for 2volts/division at 5ms.

OK, a couple of things. It appears you have overlapped the two waveforms. We do not want that; use the vertical position knob on the two channels and separate them so we can see the timing relationship clearly. Play with the Trigger variable pot knob to get a steady display. The two wave forms are not in phase, there is an offset on the positive pulses. That is what we want to measure. You may need three hands or a friend to help with the probes and/or camera.

Also we want to see three pulses on the screen. Change the horizontal time base selector (for the bad PET) to obtain this (maybe around 1 mS).

I used the knob names from my old 453 scope. The names will be slightly different on the 465. The Trigger Source should be on 'CH 1' on the right side of your scope.

You are doing good.
-Dave

dave_m
September 2nd, 2014, 02:54 PM
Here is a gadget that makes life a littler easier when probing pins on ICs, It is called something like "IC Test Clip". It piggy backs on an IC and provides vertical test points.

20261

snuci
September 2nd, 2014, 04:55 PM
Okay, I think I got what you are asking.

This is B6 pin 3 on top Pin 5 on bottom GOOD one.

Note: Time scale is 5 milliseconds

20262

This is B6 pin 3 on top Pin 5 on bottom BAD one.

Note: Time scale is 0.5 milliseconds and slightly uncalibrated so that it would not move on the screen.

20263

KC9UDX
September 2nd, 2014, 05:11 PM
Instead of adjusting the timebase, try adjusting the Stability control, or changing the trigger level. It probably doesn't matter in this case; but, if you do that, you can still accurately measure the period on the graticule. You normally want to have everything in the CAL position as much as possible.

snuci
September 2nd, 2014, 05:24 PM
Okay, here's another pic of the BAD PET.

Time is .5 milliseconds and I played with a dial called "Holdoff" until it was not moving. I REALLY need to read the manual and find out what those knobs and buttons do :) I don't know if this is good enough?

20264

KC9UDX
September 2nd, 2014, 05:33 PM
I don't know that you'll learn much by reading the manual. What you're doing is the best way to learn.

It may help if you include the entire scope panel in your pictures. It's important for someone interpreting the display to see what the controls are set to. Something that isn't obvious to you might stick out like a sore thumb to someone else.

Oscilloscopes are fairly complicated. Unless you use one every day, you'll never be 100% proficient. Even if you use one every day, once in a while you may find you can't find a particular control on your own scope. At least, that's happened to me!

The good part is, you don't have to be 100% proficient.

dave_m
September 2nd, 2014, 06:07 PM
Okay, here's another pic of the BAD PET.

Time is .5 milliseconds and I played with a dial called "Holdoff" until it was not moving. I REALLY need to read the manual and find out what those knobs and buttons do :) I don't know if this is good enough?

20264

These look good enough to pick off time periods of the 4 States. Let the team do some calculations and we'll report back.

dave_m
September 4th, 2014, 02:41 PM
These look good enough to pick off time periods of the 4 States. Let the team do some calculations and we'll report back.

OK, some preliminary information. I had assumed that using the hypothesis of a single point failure, that only the long state (S3) called Video On would be different, and that the other states would match in timing period with the good PET. But it turns out that all state periods are different than the equivalent state on the good PET. So either there are two failed parts or continuing with the more likely single point failure, the two inputs to the counter decoder C8-6 NAND gate at the bottom of schematic (http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif) must share a common problem. Looking at the schematic, counter D7 (74177) is a suspect. If one or more of its outputs were stuck high, it would cause the short counts on both the long count decoder (S3) and the short count decoder (S0, S1, and S2). If there were a failure in the lower counter D6, it would also effect both decoders, but with a much smaller change of period than we are seeing.

I would like to take a close look at the D7 counter outputs QA(pin5), QB (pin 9), QC (pin 2) and QD (pin 12). I think pin 5 and pin 12 stuck high or open would cause our "quick count" symptoms.
-Dave
PS When I am proven wrong, I will quickly place blame on MikeS and come up with a new hypothesis! This is called the Scientific Method. :)
-Dave

MikeS
September 4th, 2014, 05:26 PM
PS When I am proven wrong, I will quickly place blame on MikeS and come up with a new hypothesis! This is called the Scientific Method. :)
-DaveGlad to help with the process; besides, more often than not blaming me is justified.

Has anyone ever tried remote troubleshooting via something like Skype or Google Phone or Hangouts? Immediate feedback would be nice; keep the cam pointed at the scope screen, "put the probe on IC x, pin y," etc...

snuci
September 5th, 2014, 02:22 PM
More pics:

D7 Pin 2 Good
20296

D7 Pin 2 Bad
20297

D7 Pin 5 Good
20298

D7 Pin 5 bad
20299

snuci
September 5th, 2014, 02:26 PM
I tried to provide stable pics. Where I couldn't get a decent wave, I matched the settings with the good PET to be able to compare, at least.

D7 Pin 9 Good
20300

D7 Pin 9 Bad
20301

D7 Pin 12 Good
20302

D7 Pin 12 Bad
20303

snuci
September 5th, 2014, 02:28 PM
Has anyone ever tried remote troubleshooting via something like Skype or Google Phone or Hangouts? Immediate feedback would be nice; keep the cam pointed at the scope screen, "put the probe on IC x, pin y," etc...

That would be a great idea because i'm sure this is frustrating on your end but hopefully we are getting somewhere. I feel like this is a burden on you guys but I also think you like solving these so I persist. If, at any time, you want to give up, it's no problem.

Thank you guys for your help.

dave_m
September 6th, 2014, 04:25 PM
OK, the frequency of the pulse train out of the bad D7-pin5 is way to fast and is causing the states of the sequencer to happen too quickly.

Let's see where the clocks speed up. I'm guessing it's at the clock to counter D7, but let's make sure the clocks are normal until then.

On the two PETs, compare the clock into counter D6-pin8. They should both be at 250 KHZ and square waves. If that is correct then compare the clock inputs at D6-pin6. Are they the same?
-Dave

MikeS
September 6th, 2014, 05:24 PM
That would be a great idea because i'm sure this is frustrating on your end but hopefully we are getting somewhere. I feel like this is a burden on you guys but I also think you like solving these so I persist. If, at any time, you want to give up, it's no problem.Nah, no problem (especially since Dave's doing all the work ;-) ).

snuci
September 7th, 2014, 02:35 PM
Okay, we have some similarity between the two!

D6 Pin 6 Settings
20339

D6 Pin 6 Good
20340

D6 Pin 6 Bad
20341

D6 Pin 8 Good
20342

D6 Pin 8 Bad
20343

dave_m
September 7th, 2014, 03:12 PM
Okay, we have some similarity between the two!



Yes, good. Let's go to to the clock at D6-pin12. I'm guessing they may differ at that point. You should have gotten steadier waveforms than I am seeing. Make sure the Trigger mode is NORM and not AUTO and play with the trigger level knob if necessary.

Also don't forget to give us the time base setting so we can check frequencies.

snuci
September 7th, 2014, 03:42 PM
okay, here are the pics. These are not consistent and it was difficult to get a solid picture on the bad one.

D6 Pin 12 Good 2volts/div 0.5ms
20351


D6 Pin 12 Bad 2volts/div 0.5ms
20352

dave_m
September 7th, 2014, 04:08 PM
okay, here are the pics. These are not consistent and it was difficult to get a solid picture on the bad one.

D6 Pin 12 Good 2volts/div 0.5ms
20351


D6 Pin 12 Bad 2volts/div 0.5ms
20352

Now we are getting somewhere. Notice how long the high time is on the bad signal. That should not happen. I'm starting to think counter D6 or latch E6 may be bad. Let's give it some thought and try a few more signals. Try D6-pin11 and the counter clear signal D6-pin13.

snuci
September 7th, 2014, 04:51 PM
D6 Pin 11 Good 2v 50ms
20364

D6 Pin 11 Bad 2v 0.2ms
20365

D6 Pin 13 Good 2v 2ms Counldn't really see much more than a flicker below the line for both PETs
20366

D6 Pin 13 Good 2v 2ms
20367

dave_m
September 7th, 2014, 05:14 PM
D6 Pin 11 Good 2v 50ms


D6 Pin 11 Bad 2v 0.2ms


D6 Pin 13 Good 2v 2ms Counldn't really see much more than a flicker below the line for both PETs


D6 Pin 13 Good 2v 2ms


OK, make sure 10X Mag (red knob) is clockwise to off (I see the light on some photos). Keep everything in CAL (no red lights). Show only about 2 periods ( 2 repeating pulses). Remember to give the timebase setting like you have. Play with trigger level to get best sync. I'll need to see those last signals again. 50 mS/div setting is incorrect. The longest period of these waveforms is about 16.6 mS.

snuci
September 10th, 2014, 01:23 PM
Okay, I think I have it.

D6 Pin 11 Good 2v 0.5ms
20410

D6 Pin 11 Bad 2v 50microseconds
20411

D6 Pin 13 Good 2v 0.2 microseconds (looks the same as bad but I could not get a wave at all)
20412

D6 Pin 13 Bad 2v 0.2 microseconds
20413

dave_m
September 10th, 2014, 03:19 PM
Okay, I think I have it.



You are getting better at this, but if you notice the two D6-13 waveforms appear to be the same. The problem is that you did not show two periods of the waveform so we can not tell the frequency of the signals which I think will be different. Can you just verify that the bad D6-13 is a much higher frequency? I don't need a photo.

Please get a photo of the both D6-pin1, the clear to the counters? I think they should be the same with a frequency of 15.625 KHz. We are getting a little closer. I think the latch is properly following the counter outputs which is making me suspect the counter D6.

Mike, are you following any of this hocus pokus?

snuci
September 10th, 2014, 03:55 PM
You are getting better at this, but if you notice the two D6-13 waveforms appear to be the same. The problem is that you did not show two periods of the waveform so we can not tell the frequency of the signals which I think will be different. Can you just verify that the bad D6-13 is a much higher frequency? I don't need a photo.

I couldn't get more than a straight line on any time/division I used. I was just "high"?


Please get a photo of the both D6-pin1, the clear to the counters? I think they should be the same with a frequency of 15.625 KHz. We are getting a little closer. I think the latch is properly following the counter outputs which is making me suspect the counter D6.

These both look the same. 2V 20 microseconds

D6 Pin 1 Good
20414

D6 Pin 1 Bad
20415

One thing to note and I don't know if this matters. The D6 Pin 1 good has a flat top and bottom of the wave. The bad D6 Pin 1 has a wavy top and bottom. Just something I noticed.

MikeS
September 10th, 2014, 04:11 PM
These both look the same. 2V 20 microsecondsSorry, is that what the scope is set to, or is it the signal? I.e. if the scope is set to 2V/div then that's a 4V signal.

@ Dave: No, sorry, only time to follow superficially and leave it in your most capable hands.

snuci
September 10th, 2014, 04:27 PM
Hi Mike,

Those are the settings.
-2 volts per division
-20 microseconds per division.

dave_m
September 13th, 2014, 08:38 AM
I couldn't get more than a straight line on any time/division I used. I was just "high"?



Apparently you are having trouble triggering the scope. Let's go on to a NAND gate C2-pin 10 and 8. You should see a pulse with a period of 512 uS which is 8 horizontal scan times or one complete row of characters. This pulse clocks the latch to hold the first address of each row of characters.

snuci
September 14th, 2014, 03:17 PM
Here's some more pics. Al pictures are at 2V per division and 0.2ms per division.

C2 Pin 8 Good
20452

C2 Pin 10 Good
20454

C2 pin 8 Bad. Can't get anything but a line at 2 volts it seems.
20453

C2 Pin 10 Bad
20455

Thanks again.

dave_m
September 14th, 2014, 03:35 PM
Here's some more pics. Al pictures are at 2V per division and 0.2ms per division.


C2 pin 8 Bad. Can't get anything but a line at 2 volts it seems.
20453



I think if this signal is stuck high, it would cause our problem!! Let's think about it and also check that the source signal at B1-8 is also stuck at 2V. Maybe a bad 74LS20 NAND gate?
-Dave

snuci
September 14th, 2014, 04:02 PM
B1 pin 8 looks like C2 Pin 10 in both cases (same pics almost exactly). From the schematics I can see why what you said should be the case

dave_m
September 14th, 2014, 04:52 PM
B1 pin 8 looks like C2 Pin 10 in both cases (same pics almost exactly). From the schematics I can see why what you said should be the case

OK, that's probably the problem. You can take a quick look at the NAND input pins to make sure they are all pulsing (B1-pins 9,10,12,and 13) [no photos needed] and if so, the output B1-pin 8 is bad. Go ahead and replace the 74LS20 (dual 4 input NAND).

You should be on your way to a full 25 row screen! Or at least on the the next problem...:D

snuci
September 14th, 2014, 05:09 PM
There's definitely activity on them so I'll try to replace B1 with a new 74LS20. Hopefully, I can find a board with it that I can borrow.

Thank you for all of your time and wisdom. Let's hope this is it.

dave_m
September 15th, 2014, 12:37 PM
C2 pin 8 Bad. Can't get anything but a line at 2 volts it seems.
20453



Wait! I have it wrong. C2-pin 8 is the output of C2. I got it mixed up. I thought we were talking about the input at pin 10. Do not replace B1. It may be C2 that is bad, the 7400 NAND. Does that make sense to you? If C2-pin8 is stuck high (2V), it will keep the latch in constantly following the data input and never latch. That will make the counters appear to count fast as it will never reload the counters properly to repeat the addresses for 8 scan times as it should for each row of characters.

dave_m
September 19th, 2014, 05:17 PM
What ever happened? Did you get a chance to replace the C2 part (7400)? Maybe Mike can mail you a part if you don't have one.
-Dave

snuci
September 19th, 2014, 06:26 PM
Maybe this picture will help :)

20528

You did it! C2 was it. I can finally get these two PETs and the scope off of my living room floor :)

I am very grateful for all of your help. I really want to know how you were able to figure it out and hope one day I can work through this to be self sufficient in figuring something like this out.

Thank you to all who chimed in.

PS: Just so you are aware, I found and soldered that C2 7400 in minutes ago. I've been busy lately so thank you for your patience.

MikeS
September 20th, 2014, 07:47 AM
Excellent! Congratulations to both of you, Santo and Dave!


I really want to know how you were able to figure it out and hope one day I can work through this to be self sufficient in figuring something like this out.TTL signals as used in the PET should always be either below 0.8V or above 2.0V and any values between around 0.5V and 4.0V probably indicate a bad IC, either the one driving the signal or one of the ones receiving it; also an IC output signal that is always the same value is very suspicious unless it's just a spare. Often just following signals around with a scope (or a good logic probe) will find a bad IC even if you don't understand what it's supposed to do.

Again, well done!

Now, let's see if we can get wbochar's working...

dave_m
September 20th, 2014, 09:53 AM
Excellent! Congratulations to both of you, Santo and Dave!

Mike,
Santo deserves a lot of credit for his persistence in taking so many scope photos. He taught himself quickly on operating a scope. We got lucky and found a stuck signal that was clearly a smoking gun. Sometimes the signal is not so clear. :-)

Later, he may find small issues with the PET such as some keys not responding and cassette player problems, but they can be fixed too.
-Dave

MikeS
September 20th, 2014, 10:03 AM
Mike,
Santo deserves a lot of credit for his persistence in taking so many scope photos. You both deserve kudos for persistence and patience; you lost me about 30 posts ago ;-)

dave_m
September 20th, 2014, 12:48 PM
You both deserve kudos for persistence and patience; you lost me about 30 posts ago ;-)

Mike,
I had no idea this thread has gone on for 90 messages. It reminds me of the marathon thread from some years ago with the English fellow in Japan. Come to think of it, it was a problem in the same video control area. He had worked that problem on and off on other forums for quite a while before we came to the rescue.

I tried to find that old thread but maybe it rolled off the archive cliff. I can never find anything with the search engine on this forum.
-Dave