Re: Philips PM 3070 100MHz Scope Repair No.1

[dazz1]

Hi
I paid far too much for a broken Philips PM3070 100MHz analog oscilloscope.  When these came out, I would have sold my children for lab experiments to get one.   
This particular scope came from the Royal New Zealand Navy.  It is in good physical condition.  Clean and no physical damage.

Note that I incorrectly titled this thread for a PM3072 scope.    Although the service manual is for both PM3070 and PM3072, I can't find evidence that a PM3072 was produced.  If it was, it was probably a higher frequency version of the PM3070 that I have.

Now I have to fix it. 

The power supply emitted a high pitched note, indicating distress.  Early checks of the power supply output voltages were all close to zero.  The service manual described a test load.  I made that.  This isolates the power supply from the rest of the scope.  This confirmed that the fault was on the power supply board.

[dazz1]

I don't mind admitting that the kilo Volts present on the CRT scare me.  There is no way I was going to stick in a screw driver to discharge the tube.  There is too much risk of the HV discharging across the handle through me.  If my grip slipped, my hand could end up too close to the screwdriver metal shaft.  If the crocodile clip lost contact, the screwdriver would be floating at kV above the ground I was standing on. 

I used a piece of plastic electrical conduit to add greater separation and insulation between me and the kV.  I cleaned the conduit to ensure no surface break over.   The probe at the end is sharpened to make it easy to get under the insulator and to maximise electric field intensity.  Even if I didn't touch the anode, it should arc over to discharge.    This is much safer than the usual screw driver probe.

[dazz1]

Having determined the fault is on the power supply board, I had to isolate the fault.   There are 8 separate output voltages plus other circuits on the board.

None of the supply outputs were greater than 0.5V.  Something is dragging down all supplies.
With mains power disconnected, I applied DC voltage to each of the outputs.  If the diodes inductors and capacitors are good, the current settles to near zero.  This also reformed the capacitors.
If a diode had failed closed circuit, the injected voltage would have excessive current.
Same if a capacitor had shorted. 

DC voltage injection measured milliamps of current.  Nothing that would drag down all of the supply outputs with mains power applied.

So now I am working to the hypothesis that the fault is in the EHT converter.  The voltage multiplier is a prime suspect. 
That is where I am now. 

I need to figure out a strategy for testing this circuit.  I do not like working near HV. 

[dazz1]

I have already removed the Rifa exploding capacitor.  It looked factory original and OK but not something I want to be near if it exploded.

I was originally going to re-cap the whole scope, but the cost of the components is much more than I paid for the scope.    The current plan is to identify and replace faulty components, rather than just replace them all.

[tautech]

Expect an identical donor from the same source coming your way soon.

I've had it for a few years but never looked at it or turned it ON and it was to go to recycling but you can have it Dazz.
I can't venture down your ways this time and we have to arrange the handover between yourself and wifey but we have 2 or 3 weeks to sort that out.
AFAIK she's booked somewhere in the Hutt so maybe you can collect from her.
Your address and contact #'s still the same ?

[dazz1]

Hi
Thanks for that.    A donor would be very much appreciated.
There is also a combi-scope Philips PM3375 for auction, but it has taken a bash.  A knob is missing and the pot shaft broken.   There may be other damage.   I was tempted but the sample frequency and memory depth are not enough to justify buying it.      It could easily be bricked with an unobtainable part.  No one has bid on it. 

I have given serious consideration to buying a modern Siglent DSO, but I simply don't need that level of sophistication.  A good old fashioned analog scope is sufficient.  I don't even need a 100MHz scope.  I am either dealing with RF or slow analogue. 

I could easily collect from SWMBO where ever she stays.   My details unchanged.  I foresee the biggest risk is baggage handlers. 

I am currently trying to figure out a diagnostics plan.  I don't have kV type test gear.  I don't want to be working close to HV.  It takes 3 supply voltages to make the EHT work.  They are all being dragged down somewhere.

[tautech]

Hi
Thanks for that.    A donor would be very much appreciated.

It's a 3070 c/w the front panel hard cover and probe soft case on top. Top physical condition but unknown if operational.

I could easily collect from SWMBO where ever she stays.   My details unchanged.  I foresee the biggest risk is baggage handlers. 

Hutt Park Rd so not far from you and sometime 1st week Oct IIRC but we can sort the rest by email or on the blower. She'll be well pleased to get some stuff outta the garage. 

I am currently trying to figure out a diagnostics plan.  I don't have kV type test gear.  I don't want to be working close to HV.  It takes 3 supply voltages to make the EHT work.  They are all being dragged down somewhere.

Just checked the PC manuals folder and I have the manuals for 3070 and 3072 and there's a PM3217 here you can have too....from same source as your 3072.

I guess I'd better study up on the 3072 to see if I can be of any assistance in the meanwhile.

[wasedadoc]

Removing L6201 will ensure that the EHT generator cannot come to life.  If the power rails come up check the transistor driving the EHT transformer.  If the rails don't come up you can fault find the power supply safe in the knowledge that the EHT generator isn't going to unexpectedly start producing those kVs.

[dazz1]

Removing L6201 will ensure that the EHT generator cannot come to life.  If the power rails come up check the transistor driving the EHT transformer.  If the rails don't come up you can fault find the power supply safe in the knowledge that the EHT generator isn't going to unexpectedly start producing those kVs.

Hi
My power board is slightly different to the service manual version and there is no L6201.  Also, I needed the EHT converter powered up because the +48V was down to about 32VDC.  I suspected a fault on the low volt driver side, or the fly back transformer.
I considered removing C6209 which is the capacitor that connects the 2kV output of the transformer to the voltage multiplier but I figured it was safer to keep the transformer loaded.

Just before I started writing this, I checked all the power rails driving the dummy test load I made.  All power rails are good and no high pitch tone.  The test load smells (and is) hot.  I think driving power back into the power supply outputs (mains power disconnected) has reformed the capacitors.  In hind sight, a break-out adapter that provided easy access to the power rails would have been a better option than the dummy load.    For the dummy load, it would be good to add jumpers to be able to isolate each load from the respective voltage rail.

I anticipate someone will say I should replace all the caps, and they would be right.    If the scope works OK and is in spec, then I will leave the caps alone.  I will replace as required.  I need to order enough parts to get free shipping. That means I will end up replacing the caps in batches.    One fails, I replace ten.

The next step is to plug everything back in and switch it on.  Hopefully all the smoke stays inside.

[dazz1]

Hi
Thanks for that.    A donor would be very much appreciated.

It's a 3070 c/w the front panel hard cover and probe soft case on top. Top physical condition but unknown if operational.

I could easily collect from SWMBO where ever she stays.   My details unchanged.  I foresee the biggest risk is baggage handlers. 

Hutt Park Rd so not far from you and sometime 1st week Oct IIRC but we can sort the rest by email or on the blower. She'll be well pleased to get some stuff outta the garage. 

I am currently trying to figure out a diagnostics plan.  I don't have kV type test gear.  I don't want to be working close to HV.  It takes 3 supply voltages to make the EHT work.  They are all being dragged down somewhere.

Just checked the PC manuals folder and I have the manuals for 3070 and 3072 and there's a PM3217 here you can have too....from same source as your 3072.

I guess I'd better study up on the 3072 to see if I can be of any assistance in the meanwhile.

I am starting to be infected with a serious disease.  It's called "I like fixing old test equipment that I can fix".    I'll be ready to catch anything you throw my way.

[dazz1]

OK so one step forward.
I reconnected everything and turned on the mains power.
All the smoke stayed inside.
Illumination worked and I got a single focused dot on the CRT.
Nothing seen on the LCD and none of the front panel controls work.

So fixing one fault has revealed another.  Just like whacka-mole.

Measuring voltage rails, even with the power supply partly inserted, is harder than it should be.  I think I will be making a breakout connector set because it looks like I might have to fix more than one Philips scope.

[tautech]

I am starting to be infected with a serious disease.  It's called "I like fixing old test equipment that I can fix".    I'll be ready to catch anything you throw my way.

That reminds me of me prior to 2013 before I was in bed with Siglent.
Back then I had a hobby fixing scopes then there was no time for that hobby .....and now there is no time for any hobbies ! 

OK so one step forward.
I reconnected everything and turned on the mains power.
All the smoke stayed inside.
Illumination worked and I got a single focused dot on the CRT.
Nothing seen on the LCD and none of the front panel controls work.

So you can't move the focused dot around the display ?
Is it in X/Y mode ?

[dazz1]

Hi
The power supply board on my scope is slightly different to the one in the service manual.
The differences are evident when comparing the photo and the service manual diagram.
They include different (larger value) caps and a couple of additional connectors.

I suspect my scope is a late model, given the colour of the enclosure.

[dazz1]

Another step forward.

After lifting one of the PCBs, I found loose connectors.    This is bad sign because it indicates someone has already attempted a repair and failed.  I used to work where this equipment was repaired.  I know the quality of the techs there.  Failure to repair would not have been due of lack of knowledge or skill.   Worst case is that they discovered a faulty part with no available replacement.  Best case is that the cost of the repair (time and parts) was more than the value of the scope.  It would have been disposed of "Beyond Economical Repair".    So right now, I am not confident I can succeed in fixing this scope where others have not.

So with the connectors connected, I now have text displayed on the CRT and the LCD appears to be OK.  This is a major step forward because it indicates some critical unobtainium parts like the ROM are OK. 

I have no control with the buttons and no trace.  So a new phase starts in the fault finding process.

[dazz1]

Another step forward.
 I now have a trace A and the front panel controls all appear to work.  I think the problem was sticking buttons.  Some of the buttons are reluctant to return when pushed. 

The B trace is displayed but not displaying a signal.   This should be relatively straight forward to fix unless something weird is causing the fault.
The strategy is firstly to make sure the problem is not operator error. I need to read the user manual.
Next I will apply the cal signal to both A and B channels and compare the signals inside the scope.    I can use my trusty Philips PM 3217 scope to do this.
I will apply a binary division approach. 

That is, I will identify the mid point of the signal path and compare A and B. 
If both A and B are good, I will go down stream to find the next half way point along the signal path. 
If B is bad, I will go up stream.
Repeat.

This binary approach is MUCH faster than linear signal tracing.  Linear tracing is starting at one end and following the signal path. 

I am starting to feel more optimistic about getting this scope fully operational, but I am not there yet.  I will need to go through all of the controls and features to see if there are any latent faults.

[dazz1]

In my experience, you can always rely on operators to make stupid mistakes with complex equipment, and this is no exception. 
So when testing the B input, this operator simply plugged the scope probe into the connector next the A input.
There are 3x BNC connectors on the front of the PM3072 labeled A, Ext and B, in that order.

So when the operator moved the scope lead from Ext to B, suddenly the B signal appeared on the CRT.  Amazing!!!
In my defence, I can only say my trusty PM 3217 has the external input on the back panel. 

So at the moment, I am feeling reasonably confident that I most likely have a fully operational scope.  I need to read the user manual and work my way through every feature and function to make sure it all works.   I then need to do a calibration.  That will include looking for bad caps doing bad things to the performance. 

[dazz1]

Just taking a look around the insides while things are running OK.
I think I read something about really hot transistors in these scopes.
I have found 4x on the XYZ board.    These are too hot to keep my fingers on.  I see no indication of a fault. 
The heat sinks appear to be too small, but there is not enough room to fit bigger ones.
I am already thinking about how I could improve cooling.  Maybe 3D print cooling air ducts from the chassis fan.  I could even fit a tiny cooling fan.  Standard sizes go down to 25mm.

[tautech]

In my experience, you can always rely on operators to make stupid mistakes with complex equipment, and this is no exception. 
So when testing the B input, this operator simply plugged the scope probe into the connector next the A input.
There are 3x BNC connectors on the front of the PM3072 labeled A, Ext and B, in that order.

So when the operator moved the scope lead from Ext to B, suddenly the B signal appeared on the CRT.  Amazing!!!

Can I see myself do that....well yes of course. 
Might turn out to be an easy fix Daz. 


From just a few second on Gurgle:
BFW16A = TO-39 Final stage of the wideband vertical amplifier in high speed oscilloscopes.

IME all final stage CRT plate amps are under some stress and as such are a common failure point.
Those that appear to be their drivers OTOH need be thermally linked and as such will be harder to heatsink although with your skills Daz you could bend something up outta a wisp of copper sheet that could easily be better than OEM.

I'll have a rat through the dead PCB's stash to see if I have any TO-39 heatsinks and if so pop them into your October Red Cross package.

[dazz1]

So on further consideration, I think the unconnected connectors were not the result of someone else trying to fix a problem.  I think it most likely they simply vibrated out.  That connector type has white plastic clips to hold it in place.  These were missing on the downward connectors.  I will 3D print some replacement clips to prevent the same problem happening again.

Many years ago, this type of fault would have been repaired.  I introduced the concept of run-to-failure or repair-by-replacement.  For certain types of test gear it makes sense to use an item until it stops working, then repair by replacement.   It maximises the serviceable life of expensive equipment past the point of being supported.  Unsupported is does not make equipment unserviceable.    I didn't invent this concept.  I just introduced it. 

It is very likely that this scope was run to failure, then simply replaced with no attempt to repair.

[dazz1]

From just a few second on Gurgle:
BFW16A = TO-39 Final stage of the wideband vertical amplifier in high speed oscilloscopes.

IME all final stage CRT plate amps are under some stress and as such are a common failure point.
Those that appear to be their drivers OTOH need be thermally linked and as such will be harder to heatsink although with your skills Daz you could bend something up outta a wisp of copper sheet that could easily be better than OEM.

I'll have a rat through the dead PCB's stash to see if I have any TO-39 heatsinks and if so pop them into your October Red Cross package.

A Red Cross package for someone needing help.  Yes please.


The power supply board has an unused 5V connector, which I could use to power a small 25mm fan to supply cooling air directly to the hot transistors.   
This connector is not shown in the service manual so this fix may not be available to similar scopes.
There is a gap where I can 3D print a bracket that will just slide into place.  No holes, no cutting. 

[dazz1]

The sticky buttons on the front panel were very quickly becoming a serious issue.  I figured they had probably built up a sticky film of gunk over the years.
I took the panel apart to clean the buttons with alcohol.

All much better now.

[dazz1]

I am noticing that the LCD back light is feeble.  Looking at service manual, I could not find a part number for the lamp.   The manual says the LCD panel is not serviceable.   When I took apart the front panel, I did not see an obvious way to change the lamp.  Ideally I'd like to upgrade to LED back lighting. 

Any suggestions?

[tautech]

I am noticing that the LCD back light is feeble.  Looking at service manual, I could not find a part number for the lamp.   The manual says the LCD panel is not serviceable.   When I took apart the front panel, I did not see an obvious way to change the lamp.  Ideally I'd like to upgrade to LED back lighting. 

Any suggestions?

Yes IIRC it's been done and written up a few years back in the forum. We need to hunt it out and again IIRC it was in a 3070.

[tautech]

Gunna drop a few links here with worthy/valuable comments about these scopes:
https://www.eevblog.com/forum/buysellwanted/philips-scopes-opinions/msg119557/#msg119557
https://www.eevblog.com/forum/repair/philips-pm3070-scope-runs-very-hot!/
https://www.eevblog.com/forum/testgear/test-equipment-anonymous-(tea)-group-therapy-thread/msg1459884/#msg1459884

And jackpot !
LCD backlight repair:
https://www.eevblog.com/forum/repair/philips-pm3065-scope-repair/

[dazz1]

I found the LCD lamp replacement instructions here:
http://www.jetecnet.com/Scopes/backlight.html
That hot wire lamp looks like unobtainium.  I strip of LEDs would be an improvement.

Like the one here:
https://www.eevblog.com/forum/repair/philips-pm3055-scope-repair/