• Please review our updated Terms and Rules here

Pet 2001 vertical collapse

mickyc71

Experienced Member
Joined
Apr 5, 2021
Messages
57
Location
North East England
Hi
I've recently acquired a 2001 pet and powering it up i get a horizontal line on the monitor so it has some form of vertical collapse. I have replaced some of the video logic chips and now get a pulse on the vertical video feed but unsure if its correct any assistance would be apricated.
Many thanks
 

Attachments

  • photo60910.jpg
    photo60910.jpg
    101.9 KB · Views: 2
  • photo60911.jpg
    photo60911.jpg
    186.3 KB · Views: 3
  • photo60912.jpg
    photo60912.jpg
    54.9 KB · Views: 3
Vertical sync should be @ 60Hz and a little less than 5v. I had a similar problem with vertical collapse but it was a bad gate that turned the vertical sync to 1MHz.
If you have a 60Hz vertical drive signal but still no vertical deflection, check the connector in the monitor. The connector pins can oxidize causing a poor connection.
 
One thing that will really help people help you a lot, when you post a scope photo, is to write on the photo the volts per cm on the vertical axis, and the time per cm on the horizontal axis and the exact circuit point is was measured at. Or the displayed waveform, say apart from being a rectangular wave, doesn't help much.

Without this information, the observer has to be some kind of super sleuth detective to try to work out what might be going on.

For example, on your photo, which just happened to show the position of the scope's timebase knob, it looks like the setting is on 20uS/cm, which means the recording you have made has no reference to the vertical timing pulses, but is instead a recording of horizontal rate pulses, which have no relevance to the problem at hand.
 
I did attempt to capture the time and voltage settings but i had to reduce the quality of the image to upload it didn't realise they are now pretty much unreadable sorry, yes its 20uS/div and 5v/div and it was measured on pin 3 of the video connector, from what i gather now it should be a short sharp drop in voltage so this is not correct, as i had no signal before i was hoping it was ok but seems more fault finding is required.
 

Attachments

  • photo60949.jpg
    photo60949.jpg
    120.9 KB · Views: 3
It's probably easier to post a high resolution photograph on a file sharing website and link to it.

It's also easier to enter the scales as text into the post.

If set to 5V/div - then this looks like a 4V swing - so that looks right. I would have used 1V/div though - it gives better resolution.

At a 20 us/div timebase setting - this looks like 60 us per cycle. Pin 3 is the VERTICAL DRIVE, so you appear to be measuring the correct point on the monitor connector.

There is another post around somewhere where we had a similar issue - and it was a broken 74LS74 in the timing circuit. I'll see if I can find it. I don't think it was a 2001 though?

Dave
 
good point on the 1v/div, i will use that going forward, its been a long time since i've used my scope so very rusty(like parts of this case), i've orders a few more74lsxx chips and sockets so will replace them later this week and report back with some better images, many thanks for those that have replied, im sure once i've sorted the vertical collapse there will be other issues this was a loft find not used in many years.
cheers
 
For analog scope images I generally paste the image into a drawing program like Picture it or Paint. Then add text on it and save it as a jpg, with a file size 800 x something pixels. Then it will usually be small enough to upload to this forum.

Generally the Horizontal pulses will be around 60uS per cycle and the vertical ones around 20mS for a 50Hz vertical rate or around 16.6mS for a 60Hz vertical rate.

A lot of digital scopes have screen data that helps without any added labels.

The scope waveform you initially posted looks right for the Horizontal drive signal, where the H output transistor is turned on for about 40uS, and off for the other 20 or so uS.
 
Last edited:
The scope waveform you initially posted looks right for the Horizontal drive signal, where the H output transistor is turned on for about 40uS, and off for the other 20 or so uS.

Hi Hugo,
Yes, that looks like the problem. The signal 'Display On_Not" which runs at the horizontal frequency must not be gated properly by C6-pin13 and is clocking the Vertical Sync circuit way too fast.

See middle left of schematic http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif. Something may be causing C6- pin 13 to be high all the time. Daver has mentioned that we have seen that problem in the past.
-dave_m
 
Hi Hugo,
Yes, that looks like the problem. The signal 'Display On_Not" which runs at the horizontal frequency must not be gated properly by C6-pin13 and is clocking the Vertical Sync circuit way too fast.

See middle left of schematic http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001/320008-3.gif. Something may be causing C6- pin 13 to be high all the time. Daver has mentioned that we have seen that problem in the past.
-dave_m

Yes, that interesting "4 state machine" using the two 74LS109 JK flip flops. I really like that circuit. If that H rate waveform was measured at pin 11 of D8 (Vert drive), there is definitely a defect in that area. It would pay to double check that this was where the scope was connected and not elsewhere.
 
Thanks for the info C6 is next on my replace list awaiting for it to arrive i will post my progress when its replaced

No, wait! That may just be a symptom of the problem that is further down the logic chain.
Check to see if C6-pin13 is high because D8-pin6 is stuck high. If so, we will have to trace the problem even further back. D8-pin6 should only pulse high four times every 16.66 mS to be working right. Do you have an oscilloscope or only a voltmeter? This 4 State Machine circuit that Hugo mentioned can be devilishly hard to troubleshoot because when it gets stuck, everything looks bad until we find the actual problem and not the many symptoms.
-dave_m
 
i do have a scope, i had already ordered a few replacement 74lsxx chips, for the simple matter of swapping them out and the added benefit of putting in sockets im willing to do so, the new chips are so cheap, if after these replacements thee is no joy i will do more troubleshooting.
 
Yes, that interesting "4 state machine" using the two 74LS109 JK flip flops. I really like that circuit.
Hugo,
I agree it is an interesting circuit. It is doing so many things, not only creating the timing for the vertical drive circuit (60 Hz), but generating the addressing and timing for the video data and for the screen refresh.

However it can be a bear to fix. It probably has caused many PETs to stay broken.
I can see why Commodore replaced most of that logic with the 6545 CRT Controller chip.
 
Hugo,
I agree it is an interesting circuit. It is doing so many things, not only creating the timing for the vertical drive circuit (60 Hz), but generating the addressing and timing for the video data and for the screen refresh.

However it can be a bear to fix. It probably has caused many PETs to stay broken.
I can see why Commodore replaced most of that logic with the 6545 CRT Controller chip.

I once built another "4 state machine" but it used '74 flip D flops instead, so I could preset the initial states (from a pair of switches at either end of a travelling driveway gate) and clock the machine from a remote control through the 4 states too. So the four states were fwd, stop before rev, rev, and stop before forward. The thing is still working 20 years later. But I might have gone with that PET JK flip flop state machine, if I had known about that configuration at the time, I had not seen this before I joined this forum, so now it is in my bag of TTL tricks.
 
i do have a scope, i had already ordered a few replacement 74lsxx chips, for the simple matter of swapping them out and the added benefit of putting in sockets im willing to do so, the new chips are so cheap, if after these replacements thee is no joy i will do more troubleshooting.

Micky,
That is not a good plan. Every time one puts a soldering iron to a Commodore board, there is a chance of a lifted trace or pad. A little smart troubleshooting will keep unneeded parts removal to a minimum.
 
Micky,
That is not a good plan. Every time one puts a soldering iron to a Commodore board, there is a chance of a lifted trace or pad. A little smart troubleshooting will keep unneeded parts removal to a minimum.

I agree. The best plan, initially, at least, if nobody else has attempted repairs,, is to assume there is a single IC or component failure, and try to locate that by checking with the scope and also checking for thing like anomalous voltage levels ( not logic high or low ). And try to determine which gate or FF has failed to perform its task and is not obeying its logic truth table, considering all of its inputs. Then, only after there is very convincing evidence that a particular IC has failed, remove it for replacement.

If not, a very nasty spiral can occur with shotgun replacements where faults, including pcb track damage or even failed replacement IC's can be introduced creating multiple faults and leading one away from a proper solution, where the exact fault has been pinpointed and rectified.

Or worse, the shotgun technique works and it is cured...how could this be worse ?...because it rewards the person for this methodology and it will encourage them to fall into the same trap on the next repair job, which might not turn out to be successful.
 
Your advice is taken on board so i will attempt to diagnose any rouge signals/voltages while waiting for the parts will keep you informed.
 
is there a mistake on the circuit diagram, the symbol shows an OR gate but b8 contains NAND gates, or am i being dumb?
 

Attachments

  • photo60996.jpg
    photo60996.jpg
    45.1 KB · Views: 3
is there a mistake on the circuit diagram, the symbol shows an OR gate but b8 contains NAND gates.
I believe they are logically equivalent.
The truth table for an OR gate with inverted inputs is the same as an AND gate with an inverted output.
I think it's just a way of drawing it that makes the function more clear.
 
Last edited:
Back
Top