Need help with a Weston 1242 Nixie Tube Multimeter that is acting weird

[Postal666]

Just picked up a Weston 1242 Multimeter, and it's exhibiting some weird behavior.  It displays garbage, and acts like it's locked in an over-range state.  The mA input fuse was blown when I got it, so I'm thinking input circuitry may be the culprit.  Perhaps the damage is deeper into the conversion/logic sections? Looks like the range switch is working, as can be seen by the decimal point changing, but from the schematic it looks like there is a dedicated deck switch for the decimal point, so maybe that is ALL that is working.  There is a manual/schematic on BAMA if you'd like to look at that and see if someone can point me in the right direction.  All I have done so far is measure the PS rails. All seem to be within spec.

+12V = 11.84V
-12V = -12.19V
+6.2V = 6.24V
-6.2V = -6.27V
5V = 5.00V
HV = 96.9V

Video:
Manual: http://bama.edebris.com/download/weston/1242/1242.pdf

[david77]

What a nice meter, please do try and repair it.

The first thing I'd check is if the 1MHz oscillator is running. Do you have a scope to do that?

Another thing is the Nixie tubes showing two numbers at once. That should not be happening at all, could possibly point to defective U201...U204. In the manual it says these are '9315' nixie tube decoders/drivers. I'm not 100% sure but IIRC they should be compatible with standard 7441 or 74141 TTL drivers. That said, even the standard TTL parts might be hard to find today  .

[free_electron]

Its not a driver problem as the digits react....
I suspect a logic problem.
On the logic board : replace all ttl ics (74xx) . Most likely that will fix it.
The two big memories are ok.

[Postal666]

Had a few minutes to probe the crystal today.  I could only let the scope & meter warm up for about 20 minutes, but this is what I got.  Looks pretty decent to me.  Not a "perfect" wave, but I think it's enough to say the oscillator is ok?  What say you all, I'm just a noob...

[edavid]

Oscillator looks fine.  Maybe the 99990 detector would be a good place to start?

[Andy Watson]

Hard to tell with the old TTL logic, but to me, the top of the waveform looks a bit odd. I would check that power to chips is correct on the actual pins of the IC (not on the PCB), including the 0v. I'd look particularly at the section where you were probing the HT in the video - looks like a bad joint.

[SeanB]

Remove any IC in a socket and reinsert them, and change the 7474 IC's in the unit, as these are the most likely to fail. Most likely failures are ripple on a supply rail or an open circuit on a socket.

[Postal666]

All 74xx chips are soldered, no sockets.  I'm mapping out the pcb traces now, and will check the voltages/ripples asap.  The lines are all exposed and solder coated making probing easy, and I was probing the HV line on the trace (red) not a joint, but there could be a dry joint somewhere no doubt.

I took the 1Mhz signal from the crysal output (yellow).  The ps cap (orange) has a little tear in the end seal, which I am assuming a bad cap (they are on order, just farting around in the meantime) but I don't know.  I'll check the ripple on them.  I'll keep you posted. Thanks for all the advice so far...keep it coming! 

[PaulAm]

That double digit display is probably a bad decoder chip.  The 9315 has the same pinout as a 7441 which was a BCD-> decimal decoder chip with open collector outputs.  One of those has most likely shorted.

The 7441 is pin compatible and might be a replacement for the 9315, which was an early Fairchild TTL chip.  Might be easier to find.

See
http://www.cs.ubc.ca/~hilpert/e/nixieref/

[edavid]

That double digit display is probably a bad decoder chip.  The 9315 has the same pinout as a 7441 which was a BCD-> decimal decoder chip with open collector outputs.  One of those has most likely shorted.

It could be that, but I don't know about the most likely.  You can't tell if both digits are on simultaneously, or if it's flickering back and forth.  I would check the BCD input lines first.

[lowimpedance]

Looks more likely a logic failure. So from turn on I assume the meter does a lamp test (maybe/ maybe not?), anyway I would have expected the display to start updating as such and read approx. zero, so something is not clocking. Check your supply lines are clean with your CRO as suggested above.
 Is that clock you measured at pin 14 of U211??, if yes is there any activity on pins 8,9,11,12 (U211) ? if not replace it. Also the logic level is correct?
These vintage PCB's are not great for easy repair as the trace bond to the laminate is easly broken with heat and any force (ie when removing the component out of its holes)
Best bet since its a common part (use TTL LS to replace) get your side cutters out and cut each leg at the IC body, then after the IC is gone and there are just the legs its a simpler matter to heat the joint with your iron while holding the 'stump' of the pin with tweezers and when the solder melts remove the pin. When all pins are out use your solder wick or better a sucker to clear the hole.  Then solder a socket in and then put the new IC back in the socket.
 With such a nice old meter its worth the effort to repair, and its simplicity should mean a great fault finding exercise for you. (just be careful of the HV and mains areas!).

Just a thought is the update pulse from the A/D on pin 13 of U215 there??, see page 21 of the PDF.

 

[Postal666]

Ok, here's some more:  I checked the clock on pin 14 of U211, it's bang on to what the oscillator is putting out.  There is activity on pins 8,9,11,12 (see pics).  Funny thing is that the EXACT same data is on the outputs of U210.  You can see they feed U206, which feeds the drivers for the displays 2 LSDs.  This explains why the 2 LSDs are stuck on "7" and never change.

U214e is working, as you can see it is inverting the signal from U211 pin 11.  Pin 15 of U210 is outputting to the other chips (I forgot to snap a pic of the wave).  HOWEVER, pin 15 of U209 (and U208) is doing nothing, and the output pins are carrying different, steady, DC voltages. These voltages translate through U205 to the drivers, and explains(?) why the 2 MSDs don't know what to do.   

      U209                     U208
pin 11 = 1.50V       pin 11 = .12V
pin 12 = 4.24V       pin 12 = .13V
pin 13 = 4.30V       pin 13 = .14V
pin 14 = 4.34V       pin 14 = .14V
pin 15 = 0.12V       pin 15 = .11V

Regarding pin 13 of U215....I get this 2 pulse "blip" running at 1MHz, and it is riding on a 120Hz (ripple?) wave.  See pics.  So now it seems I have another problem that is off of the logic board.  Before I go yanking chips, I'm going to wait to get the ps caps replaced.  Just wanted to give an update on things. Anything else I should/could check in the meantime?  Thanks for all your help, everyone!

[lowimpedance]

 The 'blips' are from the main clock transistions, and it looks like there is no update pulse from the A/D converter. What is happening at U105 pin 5 ?. Pin 6 should be the inverse, is there a clocking signal on pin 1. If you get a sensible result on pin 5 check U106 pin 12 then pin 3 which should be the update pulse (maybe a bad interconnection between PCB's.) If no activity at U105 pin 1 something is wrong in the A/D.
 Note that U106 is ancient RTL not ttl and uses 3.6V supply (plenty of info on the net) , in this unit there is a zener from 5V. If that one turns out to be faulty it should be possible to bodge in a ttl inverter (different pin outs!), remove the zener etc. fingers crossed its not faulty.
 Anyway good first step is to make sure the supplies are good after the recap.

[lowimpedance]

Just noted that U209 pin 11 is 1.5V !, looks suspect . The data sheet archive has the data on th 93L10 (BCD counter) look at the National semi data sheet, gives all the required logic state info for the various pins etc. Truth table should help you check the state of your counter's.

http://www.datasheetarchive.com/

U209 looks suspect to me.

[Postal666]

Yeah, I figured (given the frequency) that the "blips" were related to clock edges, but my main concern is that it is just the 120Hz ripple noise that I'm seeing instead of the update pulse.  I think that it would be prudent to get the recap done before going much further. Besides, the wife has shut down both of my "labs" for the moment. (kitchen table and living room floor) 

Concerning U209:  I'm glad you confirmed my suspicion about that chip.  You say the tip off for you was the 1.5V on pin 11.  Forgive my ignorance, but shouldn't I be seeing "data" like on the first two chips, rather than a "stuck" DC votage?  The voltage measurements I reported were not peak voltages, but rather just steady DC output.

I appreciate your help and patience with this.  I am such a noob at this diagnostic/repair stuff, and well, electronics in general.       Your help has been more than a blessing.  I'll follow up as soon as I get the recap done!

[lowimpedance]

Of course there is a good chance data should have been present on U209 if the corresponding digits were changing, however another thing to look for is the logic levels for TTL to satisfy a logic 0 or 1 and not some indeterminate level like 1.5V, (search for the different logic families and their characteristics to give you an idea what to look for when looking at steady state conditions!). Should note it may not be U209 but U205 causing the problem with the logic level!!. 
 OR it simply could be the power supply causing wide spread effects with all the logic.
Repair the PS and post the result.

[Postal666]

Got the caps changed, but it's still doing the same thing.  No update pulse, but if I understand correctly, the update pulse is dependent on the 93L10 counter circuit completing it's cycle, right?  I think I'll start with replacing U209, since the TC output dies there, and see what happens.  Thanks for the help!

edit: from what I can tell, a 74LS160A will replace the 93L10 no problems, right?

[lowimpedance]

Quick look at the 74ls160 it appears to be a drop in pin/function compatible, so 'chop' U209 out clear the pins and holes and solder a socket in then try the 'LS160' , or maybe even an 74HC160. (remember to be careful with the heat on the pads, use enough power so the amount of time the tip is on the pad  to do the job is less. IE more heat less time is better than not enough and too long a time!. A temp controlled iron is your friend here.)
  Yeah it looks like the counters need to be working correctly to  enable the update pulse to display the conversion counted result. BTW was there any activity on pin 1 of U105?.
 Fix the counter section and I suspect it will come back to life. Update your results when U209 is replaced.
It maybe a few days before I will probably be able to offer any further 'opinion' !.

[LaurenceW]

A note of CAUTION, if i may! Your nixie tubes will be running on the higher side of around 150V DC (but it may be a rectified AC withOUT a smoothing capacitor, which is why you are only seeing ~100V). Point is, this will cause you grief if you come between the HT supply and ground, so please do be careful! Also, if anything from the HT supply finds its way into the 5V TTL circuits, you will do untold damage...

The Crystal output looks fine to me.  Before you start desoldering the TTL logic chips (U209 et al), can you give us some screen shots of the digital signals "Update", "99990", "20000", "Under Range", etc? I am expecting to see some sensible TTL data, there. Until you DO see sensible signals there, i would not start ripping out TTL chips...

This instrument is entirely repairable, but i agree that the PCB tracks will be very weak, so when you do get to swapping chips out, do cut the leads and desolder one at a time as suggested, rather than trying to remove the whole IC at once..

Nixie tubes ROOL! WWW.MrNixie.com

[lowimpedance]

 There was a photo of the update line which was not showing the required pulse at 100mS timing to latch the counter data to the display.  It was noted that counting activity had stopped at U209. Counter chip U209 has an output pin at 1.5v (stuck?). IF the counter chain cannot clock through a result an update command pulse will not  be present to latch the result.
 And yes some serious respect for the Nixie HV should be applied.
 Nixie tubes have that character that LCD's could never match, so any serious test gear collector should have one or two bits of kit with them .

[Postal666]

Thanks for the warning Laurence. I used to do residential wiring, and I do my own guitar amp repairs/maintenance (~400V) so I am VERY respectful of higher voltages. I regularly employ the "one hand in pocket" rule. 

Anyway, all of the signals you inquired about are not present (correctly at least). As lowimpedance said, the photo of the "update" line in an early post shows the signal.  This same signal is on the pins you asked about.  I could be wrong, but all of these signals seem to depend on the counter.  low- I did check U105, but it was late and I didn't snap a pic.  I'll check it again. While I'm waiting for the chips to come in I was planning on charting all of the logic levels on all the chips.  Probably won't find anything interesting, but like you said, there may be an intermediate value somewhere else.

I will keep you all posted with any progress over the next few days.  I forget what sparked my interest in Nixies (it was one of the bajillion posts here on EEVblog), but I totally became hooked.  Also, I don't mind spending some time and money to get her working.....I only paid $7 for the beauty.

[LaurenceW]

Yes, a nice project, and all the more rewarding a fix, given the bargain price you've paid! Having the manual and schematics is a HUGE boost, and saves a lot of reverse engineering.

I am considering a "retro" bench multimeter project, using nixies as the display. 6,000 (or more) count, V/A/R/Hz/C and true RMS. An off-the-shelf front end chip, and a PIC processor to command the front end chip and send  control signals to the display. Trouble is, who would buy it, when you can get the same functionality with a crappy LCD screen for 50 bucks?