Systron Donner 7004 digital multimeter from 1970's nixie display repair

[trobbins]

I recently managed to get my 1970's Systron Donner 7004 full four digit multimeter functionally working/calibrated.

I had been waiting patiently for a decade to see if two National Electronics NL-5750-S nixie tubes would land in my lap for nix (pun intended) as 2 of the display tubes were not displaying.  With some renewed incentive I opened up the dormant unit again to see if I could characterise the failed nixie tubes, and soon realised that both 'failed' tubes still worked when placed in the far left display position for '0,1,'dot' display.  I then realised that those two displays weren't working because the anode voltage was being suppressed sufficiently for each 'bad' tube not to strike. 

With a spare 16-pin header socket and some soldered on IC pins I was able to raise a SN74141 bdc to decimal driver IC above its normal pcb socket and just connect selected nixie tube cathode pins through to the driver IC.  A bit of trial and error showed that the 'best' of the two bad tubes still allowed all numerals to be displayed except for 7 and 9, so that tube was used for the most significant digit next to the 0-1-dot display, giving me effectively a working multimeter again - and for nix outlay.

The operating manual (https://dalmura.com.au/static/Systron-Donner%207004%20Multimeter.pdf) technical description and calibration procedure were next to useless, so it took a few loops to work through a calibration scheme.  The main concern was the minimum zeroed display for ACV, which required the 3 main power supply electrolytics to be replaced.  Without the guard plate covering the main analog pcb I was able to zero the last digit (0.01mV) down to 3-4 being displayed for the x0.1 ACV range, but putting the guard plate back on then raised the min display to about 47 (which was still just within spec).  I don't have the calibration facilities to accurately set the absolute FS levels, but they were able to be set against my best meter, so I'm pleased with the outcome.

[Jeff L]

Do you know about "cathode poisoning" of Nixie tubes. I don't remember if it is reverseable, but it is worth a look.

[trobbins]

The operating characteristic of the 'bad' nixies did not indicate a 'cathode poisoning' effect (requiring a higher strike voltage and having a lower conduction current - ie. higher resistance of a digit).  The fault characteristic I observed was that some cathodes appeared to continuously load down the anode such that the anode voltage (a DCV but with intentional high rectified mains ripple voltage) was noticeably suppressed and if particular 'bad' cathode pins were connected to the driver then that stopped the other ok digits from striking. 

The fault I was observing would still be related to aging, and may be via sputtered material causing a resistive path from the bad digit to anode.  One of my bad tubes had many such bad digits (but luckily not 0, 1 or the LHS dot), and the other bad tube's bad digits (7, 9) are luckily little used in the MSB position.