HP 3466A digital multimeter with flickering range LEDs (SOLVED)

[dlucy67]

I've just acquired a used HP 3466A digital multimeter (with the battery option but no cells in place) and it seems to work OK, but the range / selection LEDs flicker no matter what range or mode I select. The same behavior is apparent in auto-range.

https://youtu.be/9foXmHjH6Ok

Anyone seen this before? Have any suggestions on where I should start looking for the cause?

[wn1fju]

From the schematic, it looks like the 7 range LEDs are all multiplexed in with the normal digits via the controller U907.  If that is the culprit, you are probably out of luck since it is an HP hybrid.  However, I would check the driver transistors, Q906 and Q908, for they may be leaking (it looks like several of your LEDs are dimly lit).  Also, the ohms LED isn't lit in your video, so put an ohmmeter across it and make sure it is not shorted or anything.

[dlucy67]

check the driver transistors, Q906 and Q908, for they may be leaking (it looks like several of your LEDs are dimly lit).

Ok, I'll desolder them and run them through the tester, looking for leakage.

Being a novice, I'm not sure what made you suggest that, but I'm looking to learn. What made you suggest they might be leaking?

Also, the ohms LED isn't lit in your video, so put an ohmmeter across it and make sure it is not shorted or anything.

I did diode-check all those individual LEDs, in-circuit, and found the ohms LED was not shorted. It lits, very dimly just like all the others, when I did the diode-check on it.

[wn1fju]

Looks to me like current for the 7 LEDs is sourced out of Q906, which is turned on at the specific time by Q908, which is ultimately fired by U907.  The LED current is then sinked back to the controller chip on the same lines that are used to control the numerical digits.  So if, for instance, Q906 is not shutting off, but actually passing a small amount of current all the time, then the LEDs will fire every time one of the shared digit control lines is fired.  That would account for the flickering of the non-used LEDs.  If the transistor then is turned on to full current for the proper LED, that LED will light normally.  Just a thought.

[dlucy67]

Looks to me like current for the 7 LEDs is sourced out of Q906, which is turned on at the specific time by Q908, which is ultimately fired by U907.  The LED current is then sinked back to the controller chip on the same lines that are used to control the numerical digits.  So if, for instance, Q906 is not shutting off, but actually passing a small amount of current all the time, then the LEDs will fire every time one of the shared digit control lines is fired.  That would account for the flickering of the non-used LEDs.  If the transistor then is turned on to full current for the proper LED, that LED will light normally.  Just a thought.

This sure does sound like a plausible reason. I investigated, tested Q906 and Q908, found they were operating OK, replaced them anyways and the problem remained unchanged. Then I replaced all the LED segment driver transistors (Q900, Q902, and Q904), but no change.

The only other components I see in that part of the circuit are the resistors, one capacitor C907 inductor L901. Would failure in any of those bleed some current through to all of the units LED segments? Or is the IC a more-likely culprit?

There is a rattle coming from this board every time I flip it over. I've looked carefully at it from every angle and I don't see anything loose that would be causing a rattling noise. The only component that might have insides, as far as I can tell, is the crystal Y900. Is it possible for a crystal to come loose inside its can and rattle around? I was thinking the flickering could be caused by a too-high sampling rate or display refresh rate. The crystal manages that, doesn't it? Still wouldn't explain all the other units LED segment glowing dimly, but....

[wn1fju]

I don't see C907/L901 as the problem, and a "loose" crystal (even if possible) can easily be verified with an oscilloscope to see if it is working correctly.  I guess at this point you want to rule out the non-HP-hybrid ICs, since these are at least replaceable.  It is fairly easy to look at things with a scope.  Pay close attention to the drivers (U901, U905, U906, etc) since these are often the culprits.   Output pins that either don't change or are floating at some intermediate level between logic 0 and logic 1 are suspicious.  It's tedious, but you'll eventually find the culprit.

Try to find the "loose" part.  Who knows, maybe something is shorting a trace somewhere!

[dlucy67]

When I clip leads on to both of the legs of the crystal to look at its output, the display flickering suddenly calms down and looks like a normal 3466A. So, I'm thinking there is something up with that timing circuit and first target will be the crystal itself

[dlucy67]

I pulled the crystal, it rattles, and everyone tells me that's a sure sign of trouble. So, I've ordered a replacement crystal and will report back in with the results.

[Kleinstein]

The update rate seems to be really high. For this mode it may be just normal that the other LEDs that normally should not light up show a little glow.  There should be a way to get a slower update rate and less noise, which might be a more normal mode of operation.

A rattling crystal is not a good sign - you should try a replacement. Though it is a little unusual to have the circuit so much faster from a failing crystal - often the circuits won't ran much faster at all, even if one would try it's best.

[dlucy67]

The crystal has been replaced and the LED display has calmed down and it only changes at a reasonable, viewable rate.

Thanks to all who helped me track down the issue!