Nixie tube Fluke 8300A DMM saved! (and teardown)

[xi]

I was looking for a Nixie tube DMM since some time, but they were generally overpriced (there is currently an auction at $380 on ebay for a 8300A ) ; I ended by finding a "not working" 8300A for only $5! Not sure it had any option (the seller was unable to tell me for sure), but I decided to buy it anyway, and I got a nice surprise: it has all the most useful options: the Ohm and mV option board + the AC volts board, great!


I am going to explain what was faulty and then show some pictures of this old Fluke 8300A Nixie tubes DMM


This Fluke 8300A was manufactured in april 1972, it's a 6 digits (120000 count), auto range that uses beautiful Nixie tubes for the display. It has a resolution of 1µV DC / 10µV AC / 10mOhms, with a fairly good accuracy. Detailed characteristics and schematics are available on Fluke site: http://assets.fluke.com/manuals/8300A___imeng0400.pdf

Time to try to repair it:
I cared about selecting the right mains voltage, then powering it up: the display is working, but nothing changes on it, always displaying 00000 or 00001.
I quickly figured out that something wrong was happening with the front switches. Testing the switches' resistance shown some bad contacts (some of them where far above 1kOhms).

Disassembling these switches is a pain, they are all tied together, so I found another way: using some brake cleaner spray (some true contact cleaner would have been better, but I don't have any, and I already saved a radio potentiometer with this brake cleaner). Anyway, the brake cleaner worked, but it was not efficient enough for some of the contacts, so I decided to make pass a high current (several amps) through each contact, while manipulating the switches dozen of times. Before doing this, I carefully checked that I couldn't fry any low voltage component (like a Germanium transistor or diode), and I limited the voltage to 0.5V anyway.
The combination of the brake cleaner and the high current worked like a charm, and after this operation, all the switches contacts resistance were less than 1 ohm

Once reassembled, I powered the DMM (without the options boards), and ... it worked ... for a few minutes, until the first digit failed (all numerals were lit at the same time on this 1st digit).
I checked the digital signals and voltages coming to the Nixie tubes: everything seemed to be OK, and only one numeral was electrically commuted at a time (though all numerals were lit). I have no knowledge on tubes, but I read on the internet that the Nixie tubes may lit all the numerals when they are powered with a too high voltage.
However the voltage is within the Fluke datasheed specs (190V) ... time to disassemble the DMM again (this DMM is pain to disassemble, HP has much better mechanic designs).

I found that the failing tube had been replaced with another model, hence the problem I guessed: the replacement model doesn't cope well with the 190V voltage?
Looking at the Fluke schematics, I figured out that the HV power supply and serie resistor is common to the 5 rightest tubes, but the sixth tube uses a separate resistor and driver (the 6th tube only display "1"). So I decided to exchange the 1st and the 6th tube, and increase a bit the resistor of the sixth tube: it worked! So the replacement model of the Nixie tube didn't support such high voltage, that's why all the numerals were lit ;-)


Notice the original Burroughs B-5870ST on the left (brown socket), compared to the replacement model: a National_Electronics_1970-0042


Now sitting on top of my HP 3457A, and quite in spec (only had to tweak a really few settings and it was not used since year 2000 according to the seller!):


Posting some photos of the DMM:

This is the top view of the DMM, without any option board

[continuing on next post]

[xi]

Top view of the 8300A DMM, with the DTL logic on the left part and the analog on the right part.


Top view of the AC option board (takes AC input and transforms it into DC). There are some adjustments for all ranges (1V to 1000V, and both 1kHz and 50kHz frequencies each time)


Bottom view of the same AC option board, assembled to the main board


Top view of the Ohms and millivolts option board. This board allows the 8300A to measure DC voltages down to 1µV and resistors from 0.01 Ohms to 10 MOhms, using (or not) 4 wires terminals.


Bottom view of the same Ohms / millivolts board, assembled to the main board


Indicators are made of LEDs! The manual talks about some incandescent light bulbs, and I found some photos of these light bulbs, but my board seems genuine: the white reflector has been modified compared to the light bulbs version ; mine is deeper to cope with the LEDs' height.


Bottom view of the DMM's internals: everything is shielded and all the adjustments are identified, nice!

I am very happy with this newcomer in my "lab". I love this old electronic with DTL (Diode Transistor Logic), the analog to digital converter with recirculation-remainder entirely done with discrete components, the 5 voltages PSU with only one reference zener, .... It's a really instructive and interesting design, more particularly for guys like me who studied electronics way later than this DMM design

[factory]

Notice the original Burroughs B-5870ST on the left (brown socket), compared to the replacement model: a National_Electronics_1970-0042

That's a very nice meter and an absolute bargain at $5.

In case you didn't already know the 1970-0042 is a HP part number, unfortunately they don't list the NL equivalent number in their part lists so finding the manufacturers data for it would be very difficult.
You may be able to find a HP manual for something that uses it, to see what voltage & resistor values they used. Probably not much point though as you have fixed the problem anyway by swapping the tubes and done a bit of experimenting to find the correct resistor value to get it working.

David

edit;

Sphere list the HP 1970-0042 as being the same as National NL-5870ST (which may be wrong, see below).

David

edit 2;

Found a PDF for the HP 8600A that uses the 1970-0042, it lists it as being a B-5750-S, the circuit diagrams give an anode supply of 200V with 21.5K current limiting resistor for each tube.

David

[Jrelm7]

I have a favor to ask xi
I need photos of the top of that buffer board
I recently found 2  8300A that were in a scrap  I cleaned and put them together one unit it powers up and the nixie tubes light but I am missing the buffer board and want to make one from scratch.
I have been depopulating the non-working boards and have 90% of the parts I need and fluke has great documentation but to make my own board photos of it would sure help just removing the outer cover
and a shot of the traces would be great.

[Jrelm7]

I do see that there are photos of the top but none of them seem to have the whole board.
if I get it working I may try to build the option boards also
its a big ask so a no will be understandable.

[xi]

Hi,

No that's not a big ask and I will do that But I have to leave for the week-end now, and I will be on a business trip on next week... So I won't be able to send you the photo before next friday, sorry.
If you have no news from me on the next week, feel free to ask again.

Thank you for your patience!

[james_s]

The 8300A is a nice meter. I got one for free several years ago and gave it to a friend who I knew had been wanting a nixie multimeter and I already had a smaller Keithley. It had a minor issue but he fixed and calibrated it and now uses it pretty regularly.

[Martin.M]

actually restoration of a 8300A here,
last nixie dont work.

There are B5750 borrough on the board
Martin

[xi]

I have a favor to ask xi
I need photos of the top of that buffer board

Hi,
I finally found some time to open my 8300A and take a photo of the top of the buffer board! (sorry for the delay).
Here it is:

Keep us posted with your restoration.
Xavier

[Johnny10]

Great post!

[coromonadalix]

Nice to see it revived and working,       very good condition for 5$  loll  my friend who rebuild lamp radios would be pleased to have a meter like this one.