Fluke/Philips PM3384a combiscope digital trace faults

[Guineafowl21]

Hi all,

In analogue mode, the CAL signal displays fine on all four channels.

In digital mode, channels 1 and 3 display the trace off the top of the screen, while 2 and 4 display it off the bottom. Occasionally, the trace wobbles into view, and is distorted, lots of roll-off etc.

The self tests all pass, but the autocal gets stuck after 3 minutes or so.

I have the service manual, but can’t make head nor tail of the DAC repair tool function described near the end.

Is there a firmware update?

Anyone else had this fault? I found a thread on this forum describing something similar, but can’t find it again and anyway it didn’t really conclude. I gather a user called HighVoltage may be able to help.
Cheers for now.

[Guineafowl21]

Quick update: For whatever reason, all the digital traces are now below the screen. Also, the autocal does complete, but takes about 10 minutes (should be 4) and ends with a message to repeat the process. Status display says ‘calibration error 101’.

[Guineafowl21]

I’ve traced the signal from the front end, and it makes it onto the digital board (using square CAL wave as test signal).

It goes through a buffer/splitter OK, then through the MASPU (main aquisition signal processing unit).

From there, it reaches an upright, brown-coated hybrid chip pins 1 and 2. The output is on pin 3. The wave is still there, but looks distorted. I’m not sure if this is my measurement technique, though.

This then goes to a heat-sinked flash ADC. No data emerges, but I’m not experienced in this digital stuff. The power supply appears to be GND and -5V, so I expect the data to be a fast 0 to -5V signal.

Perhaps someone might be able to suggest what signal to be expected at each stage? Should I be using a simpler, DC signal to trace?

[Guineafowl21]

Anyone? Bueller? Anyone?

The signal from the hybrid to the flash ADC is at -5.4V, when it should be -0.5V. If I bias it to the correct level with a suitable resistor, a trace appears on the screen.

However, something’s now happened and the scope won’t trigger in analogue mode. What might I have done?

[HighVoltage]

From there, it reaches an upright, brown-coated hybrid chip pins 1 and 2. The output is on pin 3. The wave is still there, but looks distorted. I’m not sure if this is my measurement technique, though.

Most likely your "brown coated hybrid chip" is the culprit.
Show us some pictures

[shakalnokturn]

I wouldn't play with the hard to source dedicated components too soon. If the problem is affecting all channels, is one brown hybrid likely to be the trouble? (Well, one could be dragging a PS rail low... Thermal cam?)
I haven't been trough the whole of the extensive service manual, neither have I ever worked on one of these oscilloscopes. It would be worth spending a couple of hours on the service manual to get a good enough understanding on how things work.
Also try so do some more basic troubleshooting like comparing displays (if you can get traces on screen at all) on different V/div settings, using four, two or one channel (Muxed, normal or interleaved ADCs).

As your problem hits all channels, obviously checking power supplies and reference voltages is a good starting point.
+5V, +18V, -7V to brown buffer hybrids OK? Check that differential inputs to the brown buffer are symmetrical too.

Another point I'd check is reference to the DAC (DACPOT) that provides all the control voltages, you can also look at the various "offset" voltages fed to the digital section by DACPOT, if out of range due to a flaky buffer (N8016)  it's possible that will throw the traces off screen and cause clipping somewhere.

[HighVoltage]

Usually only 2 channels fail in this mode.

Either channel 1+3 or 2+4
I have never seen all 4 channels fail in the same way.

Looking forward to what you find.

[Guineafowl21]

Initially the 1/3, 2/4 pairs behaved differently - former were up top, latter down bottom. Now all are off the bottom.

The hybrid voltages -7VA, 5VA and 18VJ are all there, slightly low but within 10%.

The hybrid output ANINA at pin 3 for channels 1&3 is at -5.67V, not -0.5v as it should be. Biasing this up to -0.5V with a resistor brings a trace onto the screen.

So why is the ANINA signal over 5V below what it should be? Giving the hybrid the benefit of the doubt for a bit, I checked over the MASPU-A. Some of the voltages are not what they should be.

-12VA1 at pin 19 was -5.8V instead of -12V, and was caused by a 1R resistor gone high. I replaced it, and it’s correct now.

Voltages at V8019 and V8021, a DC feedback loop, were not 1V and -5.4V but 6V and 2V. These seem under control of the MASPU, so I moved on. 10V ref and the components checked out OK, anyway.

The other problem was pin 68 BOOT, which was at 2.46V, not -0.5V, but apparently this is a feedback from the hybrid. It’s looking a bit like the hybrid, isn’t it?

[Guineafowl21]

Also try so do some more basic troubleshooting like comparing displays (if you can get traces on screen at all) on different V/div settings, using four, two or one channel (Muxed, normal or interleaved ADCs).

Another point I'd check is reference to the DAC (DACPOT) that provides all the control voltages, you can also look at the various "offset" voltages fed to the digital section by DACPOT, if out of range due to a flaky buffer (N8016)  it's possible that will throw the traces off screen and cause clipping somewhere.

As far as I can see, no combination of settings will bring a trace onto the screen.

[Guineafowl21]

Having fiddled a bit more, putting the scope in windows mode brings up sort-of traces on channels 2 and 4. This is the CAL signal, which as you can see is a bit distorted.

@shakalnokturn where would I check the DACPOT voltages?

[DC912]

Guineafowl21, were you able to solve the issue?  I've got a PM3394B with similar symptoms--analog is fine on all four channels, but switching to digital puts the trace above the top of the screen.  I also received an error code 101 when I ran the auto calibration, but the service manual is not a model of clarity regarding the use of the error code.  The service manual indicates that the error code directs you to "where in the oscilloscope" the fault exists, but there is no further explanation of how the error code relates to the operation of the scope, or the boards or components within it. Were you able to get yours working again?

[RetiredElectronicus]

Hi all,
I have more or less a similar problem on my combiscope Philips dual channel PM3370A.
Analog is working fine for both channels.
In digital mode, Channel 2 is working ok,however, CH1 is not on the display and when turning on the vertical position button of CH1, the CH1 beam floats from top to bottom or from bottom to top on the display but in the opposite direction of the rotation direction of the vertical position button.
I found a service manual for this scope on the internet but before dismanteling, I wonder of somebody could share information to find and/or to solve this problem.

[HighVoltage]

Unfortunately, you have a broken hybrid chip that is unobtanium for a few years already.

[shakalnokturn]

Is it the brown thick film hybrids that fail? Anyone tried de-potting and reverse engineering?
If you have any corpses around I'd be curious to scratch into one...