MSD 5971 shutting down within minutes after boot

[pernillux]

Hello,

I, a retired chemist, have been lurking on this forum for several years, but now it is time to surface.

I have for a long time tried to get an old HP MSD 5971 - HP 5890 Series II GC combination up and running. I was almost there (acceptable vacuum & tuning), but then the MSD developed the following problem: after about 3 minutes communiation from PC to MSD stops, and the PC displays the 'Spectrometer interface card is not active' message. If I power cycle the MSD, it immediately starts to communicate again, but after about 3 minutes the situation is the same. I am aware that if the MSD sees no GPIB traffic for 2.5 minutes, it will shutdown. After the automatic shutdown it is still possible to communicate with the GC. The MSD behaves in this way even without vacuum, with cold diff pump.

Has anybody seen a similar problem ? How was it solved ? Cable issue ?

Greetings from Finland
Pernilla Sund

[Consumer Products Guy]

Did you try a different GPIB cable?

Do you have cable GPIB cables linked to EACH component, like computer, GC, MSD, autosampler?

[pernillux]

I measured the cable while in place (opened MSD, disconnected the internal connector from back panel to card, and measured from the internal connector, trough the suspected cable to another cable end). No broken lines, no shorts. And all units are connected (autosampler is not in use, software is configured for no autosampler).

Also, I noted that the time from boot to crash varies a bit, from 1 to 6 minutes. So I'm going to measure the power supply lines tomorrow (+5, +15, -15, and +26 V test points on the main board). Also, I am not quite done with the HPIB cable yet...

Seems to be very hard to find information about 5971 electronics... There seems to be several test points, would be nice to know what they are...

[pernillux]

Ok, the fog is dissolving... I measured the +5V power line today directly after boot, and it was 6.2 V, slowly rising ! After what could have been 30 seconds some fail-safe mechanism (I could hear a relay clicking) kicked in, the power unit did a shutdown, and the +5V line vent to 0 V.

So, it seems that the earlier reported problems was because the power unit was slowly deteriorating...

Where the heck can I get a new power unit for *220 V* ?

[pernillux]

Epilogue: It turned out that it was not possible to repair the power supply: the electrolytic capacitors have leaked (no wonder after 35 years !) and soaked the board. That has made the board slightly conductive even after being washed with water and alcohol and dried with warm air (tens to hundreds of kohms). I also got two other MSD5971 power supplies, but they suffer from similar problems. I did a reverse engineering of the power supply, and found that it is not special in any way, a standard switching power supply. I also did a total reverse engineering of the Main Board when I was at it - would be interesting to replace some of the electronics with open source hardware.

So i purchased a standard MeanWell QP-150F power supply, and installed that instead. It gives all the needed voltages, 5, +-15 and 24 V. The power supply needs ballast resistors (about 6 ohms on +5V and 60 ohms on +24 V, both 25 W or more) to give good regulation. The power supply is smaller than the original power supply, so it can be fitted in place of the original supply. It has a metal enclosure and a fan.

So, it is now up and running, no disconnects, problem solved. Now back to looking for a leak...

[MichaelVW]

I also did a total reverse engineering of the Main Board when I was at it - would be interesting to replace some of the electronics with open source hardware.

Wow. Learn anything interesting?

Keep updating please, this is neat. I came across this while looking at a list of Pittcon presenters: http://roro-uk.org/en/index
I don't know if the organization is any good but I like the idea of donating 'obsolete' equipment to people who can make use of it. It's a shame to think of all the equipment in the dumb now that worked or almost worked, but for a little hacking.

[pernillux]

Thanks for the link to roro-uk ! Yes, it is really a pity that old, still usable instruments are trashed, that happened a lot at my university.

I'm a bit unsure about legal stuff, but according to EU legislation, I'm allowed to do reverse engineering for trouble shooting av interoperability reasons, but (as far as I understand), not allowed to publish it outright. Then again, the machine is 35 years old... but here is a brief desrciption:

So, the MSD is built as a separate processor board (M68000), mainly doing communication. It communicates trough a 16 bit bus to the main board. On the main board, there are two GAL IC:s (field programmable logics), that 'interpret' 8 bits as 'command' and the other 8 bits as a 'parameter'. Most of the electronics consists of DACs, one 12 bit (10 bits are used) and four dual 8 bit DACs, all buffered by OPs, generating control voltages for the Top Board. There is also a 16 bits (14 bits used) DAC (DAC701 or similar) generating mass axis voltage for the top board.

There is only one 12 bit ADC, with an ADG201 (?) analog multiplexer in front of it. The EM amplifier, coil dip, ion source temp sensor and foreline Pirani pressure sensor are connected to the ADG201. The EM amplifier consists of a couple of metal-can single OP amplifiers, connected as log amp with a dual transistor. If I understand this correctly a log amplifier is used to achieve a big enough range with only 12 ADC bits.

Then there is a small switch mode power supply to generate filament current, with the 70 V electron acceleration voltage overlayed. And finally, a DC-DC converter (15 V to +/- 250 V), generating voltages for the lenses & repeller.

The rest is glue logic with a lot of dual flip flops to hold control bits.

It would be very tempting to use a 4 channel sigma-delta converter for ADC, and 8 or 16 channel modern DACs... all the glue logic and flip flops could be skipped...

I should also check the Top Board... who knows, some day...