Nixie DMM repair updateBeen working a bit on it last night, two aspects.
Free running counter when cold, issueSat down for a while, with the schematic and continuity tester on the board... was tedious but I located a few of the interesting test points for the gate logic.
Looks to me like the problem is not the gate logic nor the ramp, which means it must be the comparators.
Zero offset impossible to calibrate, issueI posted the same pics as the other day to avoid re-uploading the same again and waste storage space on the forum...
As I mentioned in that last post, did investigate that suspicious mundane looking 33k 5% resistor in the Zero-ing circuit of comparator 'A', which is supposed to be a precision 27K 2% instead.
I put a 150K resistor across it, so its value drops down to 27K. Result : next to nil. Offset drops a tiny bit, but really not significant at all, still way off. Hmmm....
So I got more aggressive.. much more. I placed a 33K in // with it,, effectively halving its resistance, 16.5K. Result ? Should be dramatic no ? No. Still next to no effect, offset still pretty much as high !
However, if I do the same test on the 27K resistor of the other comparator, 'B', the the offset goes through the roof and the meter overloads ! Which is more like what I would expect...
So there is something fishy with that Comparator 'A', I feel...
ALSO : there is another issue with this zero offset calibration. Not only does it never drop to a value low enough that it can be compensated/calibrated, but it also takes forever to settle.
I don't know how much it is when cold, because of the issue with the counter ftree running when cold... so I need to wait 10 or 15 minutes for the meter to stop free running and actually measure stuff... in order to measure the zero offset. But, when it does start to work, offset is like in the several hundred counts. Then it slowly, progressively decreases/get better. It's not linear : drops fast at first, then slows down more and more, like an exponential thing. After an hour or two it kinda settles for a bit under 100 counts. 75 something like that, Way more than 50 any way. Way more than the 5 or 10 counts grand max that it should be in order that it could be compensated/calibrated out.
BUT : when I swapped the board for the good one.... board is dead cold of course, was lying on the bench... yet not only do I get ZERO offset, but I also get it INSTANTLY, no need to wait a minute or 10 or 2 hours ! No. Cold board, plug it in, fire up the meter, bang you get your zero offset, no questions asked.
Sooo... to me that's hint as well, isn't it. NOt only the offset way too high, but it's also decreasing over time, and over an abnormal amount of time. There is inertia in there. So... thermal effect again ?
So, adding 2+2 : the free running problem AND the offset problem, BOTH appear to involve the comparators, and also have a thermal/from cold timing aspect to them.
So, maybe the two issues are one and the same, both caused by some thermal related problem in one of the comparators.
I think I am getting somewhere there...
So, this means, if I am right of course, that I really need to focus on only one of these two issues, and it should kill both birds with the same stone.
So I can chose what bird to shoot at...and I choose to go for the offset problem. Why ? Because :
- It exhibits itself even when fully warmed up, so I get as much troubleshooting time as I need to, instead of just 5 or 10 minutes from cold. This way I can get much more troubleshooting in one day than I could in one month if concentrating on the cold start issue... so should allow me to fix this meter much quicker.
- I's easier to probe than the gate logic approach : less probe points to search for on the board, no need to have 4 probes attaches to the scope... just generally easier to work with I think.
So, let's look again at the schematic regarding the offset compensating path.
I am starting to feel that that 33k resistor, is not a problem, a nice find but a red herring because... the aim with these two resistors and the trim pot they are connected to, is to allow the use to compensate within a certain range. With the good board I can go about 10/15 counts below and above zero. About 25/30 counts lock to lock.
Well, with the faulty board... I get the same range also !
So I think this circuitry is fine and the problem is just a massive offset super imposed on that, that should not be there... and that seems to be thermal / tme related. huge at first then slowly settle to .. less huge but still way too high a value.
So let's look at the schematic and try to understand what components might be involved here ? Problem is... I am no analog engineer, so I am shooting in the dark and don't know much what I am talking about...
So I will do the best my pea brain can come up with, which is :
- We have a big leak, and this board uses Germanium trannies and diodes which I understand are notorious for being "leaky".
- We have slow thermal effect, exponential looking... exponential fits well with a PN junction, because well.. physics. Thermal effect too because well.. physic again, PB junctions are very sensitive to heat.. just stick a diode between your fingers, measure the voltage drop with your DMM and watch it vary in real time with ease...
So... we are after a bad diode or transistors.
Let's try to narrow it down first.. the comparator is made of 7 trannies no less. However only the first 3 make up the comparator per se : two for the differential pair( S1 & S2), one for the current source in the tail (S4). The output of the differential pair then goes to another tranny for some amplification (S4), then it goes to a flip-flop made of two trannies (S5 & S6). Then the last one (S7) is connected to output of the flip-flop output, one of the anyway, and loops back to the emitter of the trannies of the diff pair. The manual says that when the comparator/flip-flop.... flips, this transistor conducts and it cuts the current in the diff par, so that the input current of meter is very low. How thoughtful.
So I guess if there is a leak, it must affect only the diff pair (S1/S2/S3), or possibly S4 too, also connected to the offset path.
So we could have either of these 4 trannies leaky, at first glance.
But what about diodes ? We have two in the vicinity. One type " OA95 " and one type " OA202 ". This one looks particularly interesting to me. Look, the annotation next to it reads " Triée ". Which is French for " Hand picked / Selected ). hmmm. why would one need to hand pick a freaking diode... what could be a sensitive part of the operation of this comparator / instrument that one would want to make sure this diode performs to some standard... what about.. zero calibration maybe ?!
Who knows.
However one can see that this diode is not connected to the zeroing/offset adjustment "part" of the comparator. Comparator input is twofold as you can see. One part is devoted to the zero-ing compensation, taken from the collector of the diff pair transistor S1, and the other, used to apply the signal to be measured, goes to the base of that same transistor, of course.
So that's another clue that the offset problem may NOT be coming from the compensation path, but rather coming from the MEASURING path, the base of S1 not the collector.
Hmmm.... I am starting to like that. it's starting to come together I feel, the pieces of the puzzle are starting to make sense to me....
Maybe that diode has too high a reverse bias current and disturbs the input/base of S1.
Yes.... I think I will look at that diode next. Could remove it, wire it up on a bread board and meausre it's reverse current. Then replace it with whatever, a 1N4148 as I have plenty, just to see if that makes any difference.
If that's not it, then I will look at the trannies S1 to S3 in the diff pair. I had a quick look on Ebay, found several sellers in Europe that have this transistor for sale. One of them sells 2 NOS ones for 20 Euros shipped. Might get that if I need to.
Anyway, making progress I think ?!
-group-therapy-thread/?action=dlattach;attach=1459438;image)
-group-therapy-thread/?action=dlattach;attach=1459444;image)
Home
Help
Search
Login
Register
Topic: Test Equipment Anonymous (TEA) group therapy thread (Read 18871425 times)