• Please review our updated Terms and Rules here

[HELP] Commodore PET CBM 3008 3032 Screen Issue

Will take a break as i think i need to wait for a working ic7812... thanks, dave
 
Nope - no special load or anything.

These things are pretty 'bullet proof' normally... You can run them open circuit, short circuit and they generally survive (unless they are 'rip off' devices of course). Only purchase reputable replacements. We have had examples of counterfeit voltage regulators...

If you have it wired up like this https://www.componentsinfo.com/wp-co...uit-lm7812.gif (but with a resistor from pin 3 (OUTPUT) to pin 2 (GND)) and it is not working, then the regulator must be dead.

Note that the two capacitors identified within the circuit I linked to are for stability. If you are 'bench testing' I generally find these capacitors can be safely ignored.

Dave
 
I was looking at an earlier post so I thought I'd talk a little about rectified AC voltage. The AC voltage output of a transformer is measured in what is called RMS ( root mean squared ). What this means in human terms is that if you put a fixed value resistor across the AC, the resistor would get just as hot with an equivalent DC voltage of the same numeric value. If the transformer is rated for 12.6V AC and we placed a 1 ohm resistor across it output leads. That would create the same amount of power ( heat ) in a 1 ohm resistor across a 12.6V DC source.
Now, AC voltage isn't a constant voltage. It is constantly swinging up and down. Part of the time it is above 12.6V and part of the time it is below 12.6V. What most meters don't actually read true RMS. They measure the average and scale it down to the RMS value, assuming it was a perfect sine wave. If you were to measure a square wave it would not be a true RMS reading.
There are more expensive meters one can get that do measure true RMS but these are not needed for measuring things such as wall power.
So, what does the voltage mean when fed through a rectifier and filtered with a large capacitor?
If the capacitor is large compared to the load, the current would only flow through the rectifier near the peaks of the voltage. It would only charge the capacitor with enough current to replace what was used between the last peak and the current peak of voltage. So what does that mean for the output voltage. If we assume no voltage drop across the rectifier, a vary large value filter capacitor and a lite load, 12.6V RMS would rectify to about 1.414 * 12.6V = 17.8V DC. Of course, diodes are not perfect. Most silicon diodes will lose about 0.7V under load. Also, there will be some ripple between the AC input peaks so a more realistic voltage would be 16.5 V to 17 V DC someplace. If it used a full bridge rectifier, that would be another 0.7V.
So, the rectified and filtered voltage DC should be higher than the AC input voltage as measured on a typical volt meter.
I should note that if I have a full wave bridge rectifier, I will often check the frequency of the ripple on the filter capacitor. I have seen many cases where the ripple frequency it 60 Hz ( US ) instead of 120 HZ ( US ). This indicates that a leg of the bridge rectifier has gone open. The excess ripple often drops low enough to go below the regulators minimum input level.
Dwight
 
Thanks Dwight for some insights into rectification. Always fun to learn more. And thank you Dave for the 7218 test wiring. That is what i did minus the caps for bench testing

As I wait for the 7812 in strong hopes it fixes the dead board, maybe I can pick your collective brains on the nonlinearity board:

Shouldn't somebody who can actually understand the magical schematics be able to kindly find a small set of parts that creates the voltage slope that makes the crt beam go from left to right. And whatever determines the speed / slope of this voltage increase needs to be replaced. And as we can exclude the caps, what is it that changes horizontal coordinate on the beam, if it is not a cap?
 
Last edited:
Yep, Hugo.

@Hugo - Let's see if he has set the flag to get notifications when we mention him in a post...

If not, I will send him a PM.

I did suggest swapping the scan coils if you haven't tried that already.

Unlikely to be a transistor or a resistor. You have had the machine recapped - so it shouldn't be that. A faulty diode perhaps? I will check my TV repair "Bible".

It is clearly in the horizontal part of the monitor circuit. If you have an oscilloscope, you can use the monitor schematics and the waveform pictures within the Zimmers Commodore PET schematics part of the website to look for differences. Just avoid signals in excess of a few hundred Volts (or whatever your oscilloscope can handle in terms of input voltage.

AH - I remember now - this is the 'strange' monitor PCB that we may not have the schematic and waveforms for isn't it? Unless it is one of the 'other' types of monitors that could have been fitted by Commodore to this particular model.

Found it. The component layout is in the PET 2001 directory http://www.zimmers.net/anonftp/pub/c...deo-layout.gif

There are three (3) schematics in the same directory (referring to three slightly different schematic variants) - but (in this area) all three are the same.

You need to check:

1. The scan coils (horizontal coils) - I suspect the easiest way is by substitution.
2. C28. This should be a 10uF NON POLARISED capacitor (the red 'block' capacitor in your photograph).
3. C25. This is the yellow tube capacitor near C28.
4. CR17. This is a diode you can test with a multimeter.

The three separate schematics each have examples of the typical waveforms that you would expect to see on an oscilloscope when the PET has booted and is displaying the READY message.

Dave
 
Last edited:
Thanks Dwight for some insights into rectification. Always fun to learn more. And thank you Dave for the 7218 test wiring. That is what i did minus the caps for bench testing

As I wait for the 7812 in strong hopes it fixes the dead board, maybe I can pick your collective brains on the nonlinearity board:

Shouldn't somebody who can actually understand the magical schematics be able to kindly find a small set of parts that creates the voltage slope that makes the crt beam go from left to right. And whatever determines the speed / slope of this voltage increase needs to be replaced. And as we can exclude the caps, what is it that changes horizontal coordinate on the beam, if it is not a cap?

The beam is scanned from the center to the right by the current in the yoke rising at a near linear rate after the horizontal output transistor (HOT) switches on. Normally there is a yoke coupling capacitor and a series linearity coil which slightly alter this profile, but these two components are generally fairly reliable. The yoke coupling capacitor, called the "S" correction capacitor is generally a special high quality part with a very low ESR and non-polar. Essentially the rate of rise of current is V/L Amps per second where V is the power supply voltage and L the yoke's inductance. So obviously the power supply voltage must be stable, it normally is. It relies also on the HOT being in a saturated switching state, meaning its base-emitter current is high and its collector to emitter voltage stays at a low value during the time it is switched on.

Generally the width control coil is a plain variable inductor and the Linearity coil has a fixed magnet, there was a case a while back where the linearity coil's magnet had fallen off the assembly.

When the beam has been deflected to the right side of the CRT face, the HOT is switched off by its drive pulses that originate from the H drive signal. Then the magnetic field in the yoke begins to rapidly collapse, in a sinusoidal fashion as you now have an undamped tuned circuit with stored magnetic energy. This produces a high voltage positive flyback pulse on the collector terminal of the HOT. This is transformed up even higher by the overwind coil on the LOPT (Line output transformer) to acquire the EHT for the CRT. After 1/2 a cycle,it also reverses the yoke's magnetic field, so the beam has now flown back to the left side of the CRT face.

In any case, you only get to see 1/2 a cycle of that sine wave as the flyback pulse, the reason is that when the collector voltage of the HOT attempts to swing more negative than its emitter, the Damper diode (better called Energy recovery diode) is forced into conduction, in this case part D704 on your schematic. At this point the beam is on the left and the magnetic field energy in the yoke and LOPT returns in a leanear manner to the power supply, this scans the left hand side of the screen. When there are severe H scan linearity defects on the left side of the scanning raster, it always pays to test and or replace the damper diode.

The usual protocol to repair a set like this would be to use a x 100 scope probe (or x 10 if that is all you had, ok in this case but not a color set) and look at the voltage on the HOT's collector, it should be uniformly low throughout the scan indicating that both the HOT and the Damper are conducting properly. If they were not then investigate why. If the were then turn attention to the yoke coupling cap and linearity coil assembly.

Any additional resistance in series with the yoke tends to degrade the linearity, compressing the right side of the raster, be it, the HOT not switched on hard, bad connections/soldering or a higher than normal yoke coupling capacitor ESR. Normally though if it was the HOT not switched on properly, it would also be heating significantly.
 
Last edited:
I feel its a capacitor but while I have quite a bit of hands on experience with CRT circuits, its still a bit, err, by intuition and tests rather than look at circuit and really get to grips with the design.

So the cap I have identified isn't based on 'real circuit investigation' rather a starter for 10.

The zimmers does have the scope traces for each part of the horizontal deflection circuit and checking them for test points 13 to 17 would be worth while, but as I said, the horizontal deflection circuit is also used as an oscillator to drive the HV coil for the CRT's anode and thats High Voltage with a capital H (even for me and I work with 400kV circuits) SO BE CAREFUL

This https://www.eeeguide.com/television-...ction-circuit/ gives a bit of a guide to how this part works (though its not the same circuit at all)

That is somewhat unfortunate. I just read the description there of how the H output stage works and it is riddled with errors, some of the remarks apply to tube output stages. I guess things don't get checked very well before they end up on the net.

Out of interest, the inventor of Energy Recovery scanning for the Horizontal scanning, prior to WW2, in TV's and CRT VDU's was Alan Dower Blumlein ( the inventor of stereo audio and a Wartime Radar researcher) he died in a plane crash testing Radar gear and Winston Churchill described it as a national tragedy. Despite Blumlein figuring this out in the UK before WW2, the UK were slower to move to it. RCA in the USA had a brilliant engineer, Otto Schade, he worked out all the math & physics to make it practical, but he credited Blumlein with the original patent.The idea was you could use the stored magnetic energy at the end of scan to scan the left side of the raster, and return that energy to the power supply, greatly improving the efficiency of the circuit and also producing the flyback pulse useful for EHT generation for the CRT.

In tube sets, the H output device (tube)does not act entirely in a switchmode manner and the shape of its drive voltage, to a degree, affects the scan linearity. This is not the case in a transistor set and the drive voltage is rectangular, switching the transistor hard on or off and it has no effect on the linearity of the scan, unless say it was inadequate to keep the transistor saturated. This would show up right away scoping the HOT's collector, preferably with the x100 probe.

One thing that is not often alluded to, is that the base drive current for the HOT, derived from the driver transformer is not due to the driver transformer's primary being driven at that time, it is due to the stored energy in the field of the transformer, when that field collapses immediately after the driver transistor turns off. The core energy in that driver transformer was acquired when the driver transistor was previously turned on. This is not obvious and many manufacturers of TV's VDU's did not mark the polarity of the driver transformers windings and sometimes they are labelled in error. The point being that the active drive to the driver transformer, turns the HOT off, not on.

I once saw a very unusual fault where the E-I ferrite parts of a driver transformer had separated a little and it had lost inductance. As a result the driver transformer stored inadequate energy and failed to turn on the hot properly during scan time, corrupting the H scan that way.
 
Wow. Very informative! I gained some understanding in the deflection process but sure cant grasp all of it yet. Will do some additional readthroughs of your posts. IC still not in but I borrowed a scope so will see if I can get some patterns to you. However I understand these patterns are high voltage so ai got to see if I have the right probe. What kV am I looking at? Understood that Ls and the HOT are also suspicious now. Did not see any ferrit damage yet and will see where to find replacements for the inductors. Thanks will update.
 
DDE1D312-01C5-4254-A690-7E75485E4F7B.jpeg

Epic win part 1 thanks to all of you. Case dead crt board is closed! Replaced FS7812 and the glow is back at the yoke and the screen comes back on. Quite the journey to find a broken 0,3 $ part.

Well some keys dont work like the a in thanks but that is another thing.

Danke an alle!
 
I might have declared victory slightly early. now that i actually looked at the screen closer than 2 yards i realized there is quite an unpleasant wobble / ripple / tremble in horizontal direction especially in the first 5 text character lines. so from one frame to the next it seems the scanline starts more to the left or right. problem seems not between lines in a single frame but between frames. any ideas? low quality cap or hot transistor?
 
>>> Wow. Very informative!

I said we needed Hugo!

At least we can 'see' the problems on the monitor now! Well done with the regulator exchange. Of course, the question remains of did it commit suicide or was it murdered?! We may still have a fault that is hiding.

Are you able to post a short video of the problem you are observing with this monitor now? A picture paints a thousand words - and a video many more...

The EHT is in the order of 10kV to 11kV - so bank on 15kV for safety. This is only on the very final output stage of the LOPT. If you don't go there - the highest voltage is something like 100V. The expected voltage levels and waveforms are marked on the associated schematic.

Dave
 
It will not let me upload mov files... it essentially wobbles in the top left quarter of the screen. i do attach an image. aside from the tilted yoke the image is also squished on the top. the scope says 400v at the probe plug and the probe is 1x or 10x so i can only go to 4kv, right? also i really need probes with a clip i do not feel like manually holding the probes here. thanks Click image for larger version  Name:	555BB2F1-5CA6-4440-94E1-A826C5E88B3D.jpeg Views:	0 Size:	122.0 KB ID:	1236859
 
Last edited:
I would centre the yoke first.

I would also put the video on a public file sharing site (or YouTube) and link to it. Waste their file space rather than VCFEDs!

It is possible that the ‘wobble’ is as a result of a few things:
  • The magnetic field given off by the transformer.
  • A 50 or 60 Hz PET operated at 60 or 50 Hz (i.e. an import).
  • A power supply ripple issue. Possibly inadequate smoothing or a defective bridge rectifier diode for example.
The probe specification will detail the maximum voltage it can be operated at. You can’t just multiply two numbers together and call that the limit. I certainly wouldn’t use a standard probe at 4 kV...

Avoid the EHT - you shouldn’t need to measure it...

Dave
 
Thanks, will do. What does the 10x on the probe mean? Is it not
it attenuates the signal to 1/10?

On your hypothesis I can exclude the import. Power supply ripple is quite an option. I scoped one of the rectifier diode legs (cant remember which sorry) and it sure looked like a pretty crooked sine, half sine. will update.
 
Quick Q: Do I need to get L1 and L2 as they are both scan coils or just one? Thanks, Alex

1. The scan coils (horizontal coils) - I suspect the easiest way is by substitution.
2. C28. This should be a 10uF NON POLARISED capacitor (the red 'block' capacitor in your photograph).
3. C25. This is the yellow tube capacitor near C28.
4. CR17. This is a diode you can test with a multimeter.
 
Yes, the signal is attenuated by a factor of 10 - so that a more accurate reading can be taken (because the impedance is increased by a factor of 10).

However, the maximum withstand voltage of the probe is a function of the user's safety and the material that the probe is made from - not from multiplying the maximum voltage that can be accommodated by the oscilloscope input and the factor of 10. It may work out that way - depending upon the probe of course.

I just checked an RS-Components part ( RS-TP 6351R - RS Part-No.: 1799558 ) and the maximum specified Voltage for the probe is 400 V (RMS).

By feeding it 4 kV - you may damage the probe and then damage the oscilloscope and give yourself a shock in the process. Read the manual...

L1 and L2 are the horizontal and vertical scan coils - and are built into the yoke itself. You can't get one without the other. I was proposing (before we got the completely dead monitor working) to just swap them over from one monitor to the other monitor. You may not want to do that now we have the dead monitor working again...

Dave
 
Most common x 10 probes can handle around 400V rms or roughly 1.4 x 400V peak.

In most transistorised VDU's of the monochrome variety, the voltage on the collector of the HOT would unlikely exceed this value, so its ok to scope there with a x 10 probe. But keep away from the neck board of the CRT with the probe as the voltages there can be a lot higher. Generally if you want to be kind to a scope, working with CRT VDU's it pays to buy a x100 probe as generally these have a max voltage rating of around 2 to 2.5kV. So with these you can scope the collector of the HOT in a color set, that often peaks to 1.2kV.

There is no need to go under the CRT's anode cap for general CRT servicing. If you don't go under there, there is no need to discharge the CRT's stored charge as it remains safely inside the CRT. The usual reason, if one did go under there would be to measure the EHT with a special probe, or to remove the LOPT or CRT, which mostly is not usually required. In the case where you go under there to measure the EHT, the set is running in this condition and you would not want to attempt to discharge it.
 
Well guys. :'( Time for my xmas story. :)

So I was getting settled with the wobble and convinced myself this is good enough and went ahead to reassemble. Adjusted yoke. Fixed the final inoperational keys. So I am done and think to myself, why not play some more with the oszilloscope and maybe learn more about the power supply stability. So I measure pins 6 & 7 from the big transformer of the PET.

Get some readings and then suddenly something happend: Christmasly sparkle with matching sounds. I don't know if I unintentionally connected pin 6 & 7 by touching the ground ring of the probe with some open cable or if it just happend anyway. I pull out the probe and notice a strong smell like soldering. No visible smoke though. Of course I pull the plug. After resting things for a while I try again and the machine is dead but makes electrical sparking noices and smells like solder. I can't locate it well but I strongly suspect the power transformer. My power meter at the socket indicates the PET box is now pulling 400W (!). For whatever reason the fuse (0.8A) seems to have no problem with that... So there is a major short there now. Darn.

Currently, I think I am not going to touch that thing again. Too frustrating. Thank you all it was great learning journey and I guess there isn't always a happy end to a vintage computer. Maybe spirits pick up after the holidays.

Merry xmas to you all and have a great start into 2022!
 
That’s a shame.

However, if you go back to my post #71 it was always a possibility...

Some machines are easy to fix, others not so, and others are just pigs! A few members I know have had pigs to repair and, to their credit, they have persevered and chased all the badness out of their machine. After that, it has been fine!

So don’t give up - have a few slices of turkey, some wine of your choice and regroup back in 2022.

I learnt the hard way when I was twenty something about what the small bit of plastic was that came with the probe - it protects the circuit being probed from the ground ring surrounding the probe (as I shorted the collector (metal can) of a PNP transistor to ground via the probe and burnt it out). This was on a home-built theatre lighting controller. Needless to say we had no more light that evening until a replacement arrived!

It is unlikely that you have damaged the transformer. I would disconnect the power cable from the PET main logic board and disconnect the power cable from the monitor and measure the transformer secondary voltages with your multimeter on AC volts.

Examine the top and bottom of both the PET main logic board and monitor board for signs of burning. There must be some evidence of it somewhere.

These are relatively simple tests to perform...

Have a good Christmas.

Dave
 
That’s a shame.

However, if you go back to my post #71 it was always a possibility...

Some machines are easy to fix, others not so, and others are just pigs! A few members I know have had pigs to repair and, to their credit, they have persevered and chased all the badness out of their machine. After that, it has been fine!

So don’t give up - have a few slices of turkey, some wine of your choice and regroup back in 2022.

I learnt the hard way when I was twenty something about what the small bit of plastic was that came with the probe - it protects the circuit being probed from the ground ring surrounding the probe (as I shorted the collector (metal can) of a PNP transistor to ground via the probe and burnt it out). This was on a home-built theatre lighting controller. Needless to say we had no more light that evening until a replacement arrived!

It is unlikely that you have damaged the transformer. I would disconnect the power cable from the PET main logic board and disconnect the power cable from the monitor and measure the transformer secondary voltages with your multimeter on AC volts.

Examine the top and bottom of both the PET main logic board and monitor board for signs of burning. There must be some evidence of it somewhere.

These are relatively simple tests to perform...

Have a good Christmas.

Dave

That is good advice.

Most likely if there was a brief short , it would not harm the power transformer. After all, that is what transformers are totally brilliant at; surviving brief periods of overload. If this were not the case, then our national grid power distribution system would have fallen apart long ago.

Most likely if you shorted a terminal out and there was a component failure it points to a semiconductor based part, like a diode, series pass transistor or an IC or transistor in the power supply. So, since you are already halfway there fixing it; Look at the area you probed when the event happened .Try to consider, looking at the schematic and where you were probing, what part got stressed the most?
 
Back
Top