DSO Philips PM3323 (PM3320A)

[Tommy S.]

One week ago my Pm3323 starts freezing the crt after 30 min up time. That happend 2 times but was easily fixed by restarting the device. Unfortunately since 4 days ago it wount start anymore.    The memory tests where pass succsessfully and then it starts an Autosetup but during the Autosetup it resets completely and reboot again. So it hangs in an endles reboot loop.
4 years ago when i got the scope i replaced all SMD electrolytics and 70% of the axial electrolytics. The remaining ones was checked this time again and they still seems in good condition. The power rails and the 10V reference all are fine except that the +- 14V sits at +- 13.3V. But i guess that is not the problem. Until now i checked different signals etc. but found nothing suspicious. The service manual looks not so helpfull to me to pin down the problem.
Maybe someone else experienced this problem with a Pm3323 or Pm3320a (70% same hardware) and found a solution.
Any hint is welcome and
many thanks in advance...

I would love to use this old DSO again ....

[SaabFAN]

I had a similar failure, but didn't fix it.
The buyer on ebay later told me that a capacitor of a RC Time-Delay had the wrong value and was causing the issue that data didn't arrive at the right time at the ADC and therefore the autosetup-routine locking up.

Unfortunately, only a few components of the entire device are clocked. Most of the logic is asynchronous, relying on RC delay-circuits to provide the data at the correct time.
I actually used this to my advantage when I overclocked the CPU, ROM and RAM of the scope.

I suspect that, as with my old unit, some status line is coming up too slow or too fast. If I remember correctly, I have deleted the email when I cleaned up my PC a few years ago, the culprit was a cap on the aquisition-board (big one underneath the board with the CCD-Chips) or the Board that contains the Sample-Memory.
I'd suggest to get a logic analyzer with at least 32 channels to help in debugging.
Also take a look at the circuit descriptions of the boards. They are really comprehensive and explain in rather great detail how the device works.

[Tommy S.]

Thanks for your reply.
A faulty timing capacitor sounds convincing to me. That would also explain the slow development of the failure. I will look on that.
I also think that maybe the uP board A6 has a defective Rom. But unfortunately I'm not able to read out the checksum coz of missing programming hardware. In case I'm not successful with the timing capacitor I will build a simple eprom burner... would you have the firmware Version 5.1 so I can compare the check sums?
Thank you again for your effort...

[SaabFAN]

Yes, I do have the 5.1 ROMs. They're attached to this post.

I'd suggest getting a TL866 or equivalent programmer. They're really handy and can also be used to quickly program AtMega or AtTiny chips if no ICSP-Header is available but the chip is socketed. And they can test chips. I've got quite a few chips here from my grandfather and a few of the SRAMs and older ROMs were faulty. Avoids frustrating hours of troubleshooting a problem that's caused by a faulty SRAM-cell

[Tommy S.]

Thanks a lot!
 Going on a 1 week vacation... so work will continue next week ...

[Tommy S.]

 Today I checked some capacitors relevant for different timings and delays of control signals... but they seemed all good :-(   I found still 6 axial electrolytic capacitors on different power rails with horrible ESR and only 30% capacity left. But they where also not the problem...
Now I get a bit lost coz without a logic analyser and only an old uncalibrated 10Mhz Oscilloscope i see no big chance to analyse and measure all the timings in this complex scope...

[SaabFAN]

That is unfortunate.

Without a logic analyzer you probably wont be able to fix the device.

If your budget is limited, I'd suggest buying one of these cheap USB Logic Analyzers on ebay. They are about 10€ or 20€ if ordered from a seller with a warehouse in Germany.
If you are prepared to invest a little more, get a Philips Logic Analyzer. For example the PM3585 or something similar. They are pretty good, well built and have tons of channels and enough memory to see what's going on.

One thing I just remembered from the repair of my PM3323 (which is not entirely complete, but I've got no time at the moment to finish it): The leaking electrolyte from the capacitors destroyed a trace on the aquisition board that is connected to the substrate of one of the P2CCD-Chips. I'd suggest you check all traces near the replaced capacitors for damage. Corrosion could have continued after the caps had been replaced.

Also did you check the components on the board that sits between the board with the P2CCD-Chips and the large metal piece that separates the top and bottom of the case? Check caps and resistors there.
In my PM3320A I had a broken resistor on the ADC-Board that was used in an RC-Filter to reduce noise on the -15V Rail of the ADC-Board. The philips engineers used a lot of low value resistors (they could also be axial leaded inductors now that I think about it) to reduce noise on the supply rails. One of them might have failed open.

[Tommy S.]

I carefully checked nearly all relevant PCBs optically and measured on the semi analog boards the voltages on different components and compared it between channel A and B. But I found nothing suspicious.
It's only clear that the CPU stop after it enters the auto set routine and the watchdog initiates a Reset after 3sec.
So now I will build a simple ROM reader to be able to compare the Checksums.
To be honest I not want invest much in analysis hardware only for this DSO when I can get a similar working DSO for around  100 - 150€. I'm not much into digital electronics... I enjoy old analog things    so I wouldn't use that analysis Hw after anymore.
Anyway many thanks for your effort, Tips and Hints 
I really appreciate that.

[SaabFAN]

It's only clear that the CPU stop after it enters the auto set routine and the watchdog initiates a Reset after 3sec.

Ah, now that's interesting!
This means that the program crashes.
My scope didn't do that. It just remained in the auto-set function forever but didn't reset.

In my book this narrows down the probably cause to the CPU, ROM and RAM. As well as the bus, maybe. But most likely RAM and ROM (boards A5 and A6).
If you have compressed air or one of these dust-sprays for PCs, try cooling the ROMs and RAMs when the scope starts to lock up. If the lockup goes away, you have found the likely culprit.
If it's the ROMs, first try reprogramming them. If it is the RAMs, it gets a bit more tricky - these chips are hard to desolder. I managed to do it with hot air, but caused localized de-lamination on my RAM-Board when I overdid it with the heat. Didn't stop them from working with the CPU overclocked though

In case it is the ROMs, I've still got the 4 ROM-Chips from my PM3320A, which I replaced with more modern flash-chips to allow overclocking. I could program those with Firmware 5.1 and send them to you.

[Tommy S.]

The RAM test at startup passes so I believe more in a ROM problem. But it could also be, like you mentioned before, that a timing is wrong and the CPU not get the right data at the right time and stuck... maybe in your DSO was the watchdog disabled (jumper setting) ... if I do so in mine it also hangs forever in the auto set. The manual says:  the dso will work also with watchdog disabled.
Anyway ... I  will try to check the ROM now. But that will take time coz I need to setup some hard- and software to read them out.
Thanks very much for your offer to send newly programmed ones ... I will  take it if I'm not successful with my "Reader activities " .
Unfortunately I got today from a friend a professional Bosch car battery charger to repair.. I hope it will take me not to long ( mostly they have only a defective control button) .
Thanks again for your help.

[Tommy S.]

The Bosch C7 car battery charger works again... in the last 4 days I developed an EPROM-reader.
One day concept, another 2 days HW assembly and software development (VB6 it runs under win7) and on the last day I tested and improved it till it reads stable and fast ... 1Mb EPROM in 2 min.
But the sad news is: all my 4 Rom's are correct  :-(  So they are not the problem why my scope stuck in the auto set routine. So now I should get a logic analyzer to check all the timings but I think I rather prefer to buy a new scope....