Nixie DMM updateAs always while everybody was on Discord, I was at the bench working hard.
So, progress has been made ! There is some good some bad, it's been a roller coaster working on this thing this evening.
OK so first, I tidied my PSU repair. Removed the two 470uF 63V caps in // that I bodged under the chassis to replace that failed 1000uF 50V cap.
Replaced them with a single cap, 1000uF but only 35V. It's the closest match I had in my brand new Nichicon left overs. 35V is nowhere near 50V but.... in service the cap only sees 20V. So 35V is enough headroom I think to keep this cap happy, especially since of course I hardly run my meters 24/7 all year long to begin with...
![Confused :-//](https://www.eevblog.com/forum/Smileys/default/confused0024.gif)
The diameter of the cap was small enough to fit under the chassis while still clearing the cabinet. So I just used some hot glue to keep it in place. Use some heat shrink tubing to make sure it's never gonna short anything. This way the original can cap, screwed to the chassis, is still in place. That leaves me the option later to result it if I fancy so.
Having this cap sorted was a necessity so I can now put the meter in its normal orientation,; which I could not do with the two 4760uF cap hanging in the air with long exposed leads shorting themselves all the time...
So that one good thing done.
Then I powered the thing up to work a bit on troubleshooting my "free running counter from cold" issue.
It did not disappoint : problem showed up again, and again after a few minutes warming up, it vanished and all of a sudden the meter started working just fine.
So had only a few minutes to work on it. I concentrated first, as said yesterday, on the trigger section, and the ramp generator. That's where the bad stuff starts to arrive...
![Undecided :-\](https://www.eevblog.com/forum/Smileys/default/undecided.gif)
I wanted to set the meter to "Manual" Trigger, so that it does not generate a ramp at power up, so I can see if the output of the ramp was valid, as if it's not it might well play havoc with the comparators hence ripple that garbage to the Gate logic hence it would explain my free running counter.
That's when I eventually realized that the switch that selects from AUTO or MANUAL.. is dodgy ! It's a SPDT switch. In AUTO mode, the contact close and open as they should. However in NORMAL mode, it's neither clearly open nor closed. This no doubt explains why I noticed that in AUTO mode the meter displayed (once warmed up and working I mean, of course) normal/sensible results, but in MANUAL mode, the result of the measurement would be good only 10% of the time. 90% of the time it would be absolute garbage, even illegal values ( >= 5000 counts ! ).
The switch looks to be the same type/style/construction as the mains/power switch which, I remind you, is ALSO failed (open circuit). They looks like sealed plastic body stuff, so I can't open them to refurbish them.
So, for my troubleshooting purposes, I shorted the that switch so that it's hard wired to MANUAL mode. Then I realized that when I depress the push-button, I also get sometimes weird results... so this switch is probably also dodgy. Nowhere near as bad as the MAN/AUTO one, but still a little bit.
This meter has seen too much moisture for too long, I fear it's full of gremlins and I may never get this meter working 100%, or if I do it will develop new problem every other months... I am starting to be a bit worried and depressed...
Anyway, it still was good for my troubleshooting purpose. It showed that yes I have stuff to sort out in the trigger department, but that no, it does not cause my free running counter problem... because... I could see that the ramp was 100% valid : stuck to the negative 6V rail as it should. So I now know that both comparators are getting valid input signals when the counter is free running. So now I have narrowed it down to the comparators misbehaving, one of them at least. I mean one might give a false "Start" signal", which would get the counter started, and the other comparator, working fine maybe, would therefore of course not give any "Stop" signal, hence the counter free running.... Or, if the comparators are good, then the problem must reside in the Gate logic circuitry. So I need to probe that stuff, but its not easy because I first need to identify on the board WHERE are the relevant freaking nodes I need to measure, and it's absolutely not obvious ! I have looked at it a bit, I think I have narrowed it down to a small portion of the board, but now need to spend quite some time to pin point the precise resistors I need, and find which side of them to probe. No silk screen or board layout is A MAJOR PITA !!!!
![Rant :rant:](https://www.eevblog.com/forum/Smileys/default/rant.gif)
I got to that point, then as expected, meter was warm enough and it started to work properly. Since I now know I can count on the meter working fine after just a few minutes, I have decided to now also work on the other issues. So I wanted to adjust full scale. But that didn't go well because one of the two trim pots you need to tweak, is on the chassis, accessible from the top... so I need a long screw driver so I can get between the Nixie boards from the top, to reach the pot/chassis at the bottom. That means I also need an ISOLATED screw driver, because don't want to short things out, especially since it's not just the two nixe boards around, but also the PSU board with exposed terminals... don't wanna short the mains or secondary windings you see
![Roll Eyes ::)](https://www.eevblog.com/forum/Smileys/default/rolleyes.gif)
My flat screwdrivers are standard lenght, blade is 100mm long. I would need a 150mm blade at least, to do that job. So, Monday, I will go buy a log sparky screw driver.
Or at least a long non-isolated one, and isolate it with heat shrink tubing like I did the other day to one of my screw drivers to adjust the trimmers in my Tek 180A time marker generator.
Of course the goal of adjusting full scale was to determine the value of the resistor I need to put on the Vref board to achieve that. I would replace that resistor with a pot, then set the Zero adjust half-way, and adjust my pot. then substitute a fixed resistor as close as possible to the pot value. Then I would be good to fine tune full scale using the factory trim pots.
Anyway, i eventually occurred to me that, of course.. before you can adjust full scale deviation, you first need to adjust the ZERO. Isn't it...
So, I go to do that and... wow, the meter is completely out of whack : the zero is displaying 300+ counts !!!
![Scared :scared:](https://www.eevblog.com/forum/Smileys/default/scared.gif)
I tried the Zero pot anyway, but it looks like its range is about 25 counts (which sounds sensible).
So I have big offset problem in this meter, it's NOT normal !!
As the meter warmed up, it slowly decreased/improved, down to 100 or less. But still out of adjustment range, and still not normal/acceptable to me !
As a quick test, I swapped comparator boards, and the good board.. works just fine : meter show zero not 300, and the trim pot allows you to finely tune that 10 counts above and below 0. It's absolutely.. .perfect. It's funny though, that you can display both a negative zero and a positive one... two zeros... I went for the positive one after some discussion with myself. It's absolutely irrational and all to do with some OCD I think, for some reason I don't "like" a negative zero as much as a positive one ..
![Confused :-//](https://www.eevblog.com/forum/Smileys/default/confused0024.gif)
So, the problem clearly comes from that board. Schematic attached. One can see how the zeroing is done : one input of the comparators, via a precision 2% 27k resistor, is connected to one end of a 5k trim pot whose cursor is tied to the negative 12 volt rail.
I compared the two boards and quickly noticed something weird, see pic below. On the working board, I do see those two 2% 27k resistors, no problem. But on the troublesome board, I see only ONE of them. The other one, in its place, I see... a 33k regular 5% resistor !!!
Yes, 33k not 27, and 5% not 2% ! This alone I think could cause the big offset I am seeing. Worse part is .... I am 100% certain that this completely inappropriate resistor has been fitted AT THE FACTORY !!! For, it looks 100% like the other resistors on the board, even down to the little imperfections in the painting of the colour bands.
Leads looks like they have been bents with care, not by some cave man. But the definitive tell tale sign that this is factory is... on the solder side, the joints look factory and the conformal coating is INTACT !!!!
I can't believe it , there is no room for mistake here, this has been done at the factory for sure !
![Scared :scared:](https://www.eevblog.com/forum/Smileys/default/scared.gif)
I don't have precision resistors, only 5% ones, but I do that 27k ones. I will lift a leg of the lone 27k resistor to measure it reliably/accurately, then I will hand pick one of my 5% 27k to get as close as possible. However the trimmer being 5k, I don't think I don't need to be ultra close to 27k anyway, so not too worried. 27k 2% is 540 ohms tolerance either side. I should not have a problem finding one of my 5% resistors in that range. So wl do that... stay tuned !
![Cool 8)](https://www.eevblog.com/forum/Smileys/default/cool.gif)
![](https://www.davidpibworth.com/V1/sysGetThumbnail.asp?URL=c:\inetpub\wwwroot\files\152\images\the-arches-theatre-thats-all-folks.jpg&Width=950)