Philips PM 3210 oscilloscope with slowly fading trace

[dmulligan]

I have a vintage Philips PM 3210 25MHz 2 channel analog oscilloscope which works well until it has been on for a while, the trace slowly disappears.  I haven't timed it but the trace starts to fade after about 15 minutes and has completely faded away after 20-30 minutes.   

I have only opened it up to look for burnt parts, nothing looks bad.  The service manual suggests a few things such as checking the supply voltages and switching or turning a few controls such as turning the intensity knob all the way clockwise, setting the channels to 0, setting the mode position to XY etc.  The neon bulb stays on when the trace has faded away, I have made the suggested control adjustments, and still no trace.  Setting the mode to XY shows a dot for a short time, but it too fades away.  I have not yet checked the supply voltages but I will once I know what the next steps might be.

I guess I am also trying to decide between getting a cheap scope or repair this one.  The high voltages and heat inside this scope makes me nervous to poke around inside it while it is powered on.  It doesn't help that one of the graticule bulbs are burnt out too.

Thanks,
David

[shakalnokturn]

This is usually symptomatic of a defective HV transformer on the PM32xx series.

I have a couple with the problem awaiting repair.
The HV winding is the outmost one so I'm considering cutting through it with a Dremel to be able to use the original transformer's other secondaries, then try using a single CCFL inverter transformer to supply the HV output for the CRT. I have done this successfully before on a scope (without PDA) that had an open HV winding.

[dmulligan]

@shakalnokturn how does a transformer cause a failure like this where the trace slowly fades away?  The only way I can imagine is heat causing an open or short which would be a sudden failure, wouldn't it?
Is there a safe way I can test for this without a HV capable meter?  Is there a side effect I can look for?

[shakalnokturn]

Insulation degradation, internal leakage, weak diodes (although rectification may be external here).
The easiest things to check without HV specific equipment are:
Evolution of primary current draw.
Preheating the transformer with a hair-dryer to check how shorter the uptime is.
Scoping the transformer on lower voltage windings, usually a defective one will show decaying voltages over time or extra bounces.
The decay is not always very obvious to spot as the HV winding has high internal resistance it can collapse enough to loose the trace with little effect on other voltages.

[dmulligan]

I wonder if slapping a computer case fan on the scope beside the power supply would breath more life into it.

[Bareter]

The problem with slowly fading trace is not related to the HV transformer, but to the mains transformer. I have the same oscilloscope and I had the same problem. This is due to the fact that the filament of the lamp's is at a potential of -1600V and as the mains transformer heats up, a leakage and short-circuiting of -1600V will lead to chassis. The solution is to add an independent, small filament transformer (6,3V), with good insulation between the windings and the casing, and place it e.g. in the place for the mains cable. I did it and the oscilloscope worked good for many years. Unfortunately, few days ago the HV transformer has damaged. High voltage winding break and short circuit with other windings. As a result, there is no high voltage -1600V and + 8500V and no other supply voltages. I am looking for a solution. Maybe some suggestions? First, I will try to remove the high voltage winding and I will check if the supply voltage +/- 22V and +/- 85V appears. Then I'll think about how to use a little HV transformer from an old black and white PC monitor. For this generator on NE555 or...?  I have a sentiment for this equipment, so I would like to fix it. Please any advice.

[George Edmonds]

Hi

I have been repairing such faults now for over sixty years, the chances that it is a faulty transformer is almost zero.

In my experience such behaviour is invariably caused by one or more the resistors in the high voltage chains that supply the CRT slowly going high in value as a result of the applied voltage.  The time scale given is exactly what I would expect to see.

The real problem is working with the high voltages present.  It is not possible to do valid low voltage measurements with a DMM of the resistors as they quickly change resistance to a lower resistance once the high voltage across them is removed.  DO NOT use standard resistors as replacements, there are high working voltage types available from the likes of Vishay, look at their VR series for example.

G Edmonds

[dmulligan]

Thank you George.

You said that it isn't possible to measure the resistors themselves with a DMM, I understand why.  In your experience are there any low voltage or temperature measurements possible to infer which high voltage parts I should focus on?

Thanks,
David

Hi

I have been repairing such faults now for over sixty years, the chances that it is a faulty transformer is almost zero.

In my experience such behaviour is invariably caused by one or more the resistors in the high voltage chains that supply the CRT slowly going high in value as a result of the applied voltage.  The time scale given is exactly what I would expect to see.

The real problem is working with the high voltages present.  It is not possible to do valid low voltage measurements with a DMM of the resistors as they quickly change resistance to a lower resistance once the high voltage across them is removed.  DO NOT use standard resistors as replacements, there are high working voltage types available from the likes of Vishay, look at their VR series for example.

G Edmonds

[dmulligan]

Based on what George said, I am thinking about replacing R1013, R1014, R1015, R1016, and R1017 from the schematic on page 114 of the PM 3210 manual.  I plan to use Vishay 1% HVR25 or HVR37 resistors.  The manual specified carbon E24 series resistors but the Vishay resistors are metal film.  Is that ok?

https://archive.org/details/manual_PM3210_SM_PHILIPS/page/n57/mode/2up

[Zenith]

Metal film resistors are considered a superior replacement to carbon, so that should be perfectly fine.  The main thing is to select resistors with a high working voltage, which you are doing.

[dmulligan]

Now I am not sure that I am looking at the right "resistors in the high voltage chains."  Originally I thought George meant R1013-R1017 but now I think he meant R1006, R1007, and R1012.  Perhaps he meant both.

[shakalnokturn]

Please let us know how it goes, if you're successful I'll be placing an order for resistors.

[dmulligan]

What have you tried?  What are your exact symptoms?  Have you gone over the troubleshooting guide on page 72/38 in the PDF I liked above?  I re-read this recently and I now believe that section XI B1 covers this issue.  I intend to run through these before I place the order.

[dmulligan]

Interesting thread (use google translate) https://www.circuitsonline.net/forum/view/45068/1

[dmulligan]

When the service manual says "Check the emitter voltage of transistor X" does it mean ground to emitter or base to emitter?  I am pasting the relevant portion of the manual here and use the schematic diagram a few posts above for reference.

[Chris56000]

Yes, all voltages quoted in manuals are normally with respect to chassis/0 V/earth unless specifically stated otherwise, both the voltages on the fault–finding notes  and in figures on the Circuit Diagram are with respect to chassis!

Chris Williams

[dmulligan]

Now that my new DHO has arrived, I have finally opened up my PM3210 CRO to check the levels mentioned in the troubleshooting section of the service manual.  Several of them are waveforms thus needing another scope.  I was very cautious since there is high voltage present on the combination unit board. Of course the PCBs of this scope predate solder masks so there no helpful markings on the board.  To locate the probe points I was extra careful deciphering the black and white PCB photos in the service manual pdf as they are quite faded and blurry.

As part of the effort to be find the proper probe points I removed the combination unit board to confirm some of the bottom traces.  Luckily the high voltage section of the board is isolated to corner less than a quarter of the board.  The bottom of the HV section looked singed at first glance, it ended up being dust clung to the area and turned toasty brown over the decades.  I cleaned that up and found what from certain angles could have possible been two cold solder joints.  I really wasn't sure but I resoldered them to be sure. 

I reinstalled the board and observed that all of the transistors are socketed.  So I pulled each of the plastic (Maybe ceramic?  They were black.) transistors, looked at them and put them back. I also decided to twiddle each of the 5 potentiometers present as well.

Finally I powered it up, took the power supply measurements once the scope warmed up enough to show a bright trace.  Then I waited for it to fail, which normally takes 10 to 15 minutes.  After about 10 minutes my wife needed me.  I came back to it after at least 30 minutes, probably more than 45 minutes.  There was still a bright trace showing.  I have my fingers crossed as for now it looks like it is working again.  I checked a bunch of stuff for a while and all issues were attributable to dirty pots and switches.

Thank you for everyone's input.

[artag]

Glad it worked out .
The potentiometer twiddling sounds risky. Were you aiming to adjust them or just move them to clean the contact and hopefully put them back in the same setting ?
 

[dmulligan]

Glad it worked out .
The potentiometer twiddling sounds risky. Were you aiming to adjust them or just move them to clean the contact and hopefully put them back in the same setting ?

Exactly.  I moved each just far enough to say that it had moved and then moved.  I was trying to make sure that I didn't change their values while cleaning the contacts.

Looking back I now wish I had performed each step one at a time, testing between each one.