Test Equipment Anonymous (TEA) group therapy thread

[Ice-Tea]

Yeahhh...

[TERRA Operative]

Called it!   

[Ice-Tea]

Not at all (Schaeffner, not RIFA)(yeah, I know).

Which one is it, do you ask? Well, obviously the ultra-compatc unobtanium version, d'uh.

[med6753]

Not at all (Schaeffner, not RIFA)(yeah, I know).

Which one is it, do you ask? Well, obviously the ultra-compatc unobtanium version, d'uh.

What's the voltage rating of the capacitors inside? Same as here? 250V? Those line filters rarely blow their cookies over here and I'm sure it has everything to do with the lower line voltage. Putting 250V rated caps across a 240V line is just inviting magic smoke.   

[TaylorD93]

Came into work this afternoon to find a PL320 TTI PSU sat on my desk. Broken ofcourse.

Since they found out at work about my TEA problem and the fact i had at one point almost 18 Oscilloscopes, i suddenly found myself repairing company equipment.


In this case, a single bulk capacitor from 33rd Week of 1993 had failed (leaked over the board) so the 7.5V and +5V rails were dropping out at 100Hz as the inputs were dropping below the dropout voltage etc. Quick run down to the shopfloor stock room, one perfect match cap, 35V 1000uF same footprint etc. PSU now repaired and works.

Don't have the time to go through and fully recap it, but its now working and for their use (mostly 24V cooling fans on dummy loads) it will be fine. Some noise at 32V at full load, however given the age of the PSU below 30V it appears to be in spec as per the service manual.

[Ice-Tea]

Not at all (Schaeffner, not RIFA)(yeah, I know).

Which one is it, do you ask? Well, obviously the ultra-compatc unobtanium version, d'uh.

What's the voltage rating of the capacitors inside? Same as here? 250V? Those line filters rarely blow their cookies over here and I'm sure it has everything to do with the lower line voltage. Putting 250V rated caps across a 240V line is just inviting magic smoke. 

Don't know, they're potted. But I once opened one to fin RIFA inside, so.. But you're right off course, much more likely to go on 240V. This one has always been on 240V, though. Bit bummed that I can't just pull one out of my magic Schaffner box, but it seems FN9226 is a decent replacement.

[factory]

A digital power analyzer for the Panaplex fans, US only auction*
https://www.ebay.com/itm/353993185981

That looks broken, too many dots on and I guess they didn't make a model with a binary display.

Well I usually say "if it isn't broken, don't buy it". 

Can't help thinking part of the problem, is that none of the volt range switches are pressed. This probably won't do much for the extra decimal points on the Amperes/Watt displays though. 

Did anyone bid for it? it made approx $54.

David

[Neomys Sapiens]

Tadaaa!

Order placed today...

You're buying a Windows CE computer with a floppy disk drive?      

And wasn't he about to get one with USB? 

It has USB. 4 USB2.0 ports as well as a side-mounted CDRW drive. As well as a 80GB HDD. And it runs Windows2000 and not CE. I have no reason to complain about that, as the 7xxx series uses the same OS and I spent enough time with them to see that it is stable in this application.

[mnementh]

Not at all (Schaeffner, not RIFA)(yeah, I know).

Which one is it, do you ask? Well, obviously the ultra-compatc unobtanium version, d'uh.

Meh... Schaffner is just RIFAs wearing a Iron Man suit. 

mnem
"Jarvis... spew my guts all over that high-voltage SMPS!!!" 

[factory]

Not at all (Schaeffner, not RIFA)(yeah, I know).

Which one is it, do you ask? Well, obviously the ultra-compatc unobtanium version, d'uh.

Meh... Schaffner is just RIFAs wearing a Iron Man suit. 

mnem
"Jarvis... spew my guts all over that high-voltage SMPS!!!" 

Or WIMA paper cRaps too and one unidentifiable cap. 



David

[Vince]

Nixie DMM repair update


Zero offset impossible to calibrate, issue

...
...

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.

...
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.

OK so I looked at that special hand picked diode OA202.

Looking at the schematic, me thinks it's simply there as protection to clamp the input voltage to safe levels, just above 5Volts, the capacity of this meter.
So it's always reversed biased. I guess it's reverse current, would mess up the input of the diff pair, the base of transistor S1, and cause an offset. I guess.
That might be why they put a special diode in there, proably special in that it has a low reverse current.

So I pulled the data sheet for both that OA202, as well as the OA95, the latter being the general purpose one that litters the instruments, tens of them all over the place.

I was not disappointed. the OA95 is given for 30uA reverse current, and the OA202 is.... no less than 3 orders of magnitude lower ! 
And it's Silicium at that, not Germanium unlike the OA95.

OA202 is given, typical, between 0.01 to 0.1uA at ambient temp. I pulled it from the board and tested it. Measured the current with my MX56 DMM as it's the only thing I have than measure low currents. 0.01uA resolution, just enough for what we need here. Well, it measured... just above 0.01. So not only is it within spec, but it's at the very best of what it can be. Hand picked it was, and they did a good job at that it appears !
So, that diode is not the problem 
Since it's just a protection diode, I tried the meter without that diode in place, thinking well at least I am SURE it won't leak anything at all !
But yeah, doesn't change anything, of course....

Oh well, it was worth a shot, I did need to rule it out anyway, so that's done...

So now the next suspects are transistors S1 ans S2, that constitute the diff pair itself.

Need to swap them. Problem is that it's not easy...as you can see from the pic of the board above, the 4 transistors, the two of each pair on both comparators, are held in place all together by some black plate, and the metal cans of the transistors are glued to each other and to the plate. I had a go with my soldering iron on that glue : it barely makes a dent ! I am fucked, can't remove these transistors 

How to remove that glue ? By mechanical means maybe.. maybe try to use hack saw blade to weaken it, then a tiny / precision file to grind it down... 

The fun continues....

[Cubdriver]

A digital power analyzer for the Panaplex fans, US only auction*
https://www.ebay.com/itm/353993185981

That looks broken, too many dots on and I guess they didn't make a model with a binary display.

Well I usually say "if it isn't broken, don't buy it". 

Can't help thinking part of the problem, is that none of the volt range switches are pressed. This probably won't do much for the extra decimal points on the Amperes/Watt displays though. 

Did anyone bid for it? it made approx $54.

David

I was busy with work and missed the auction end.  Probably just as well, I need it like a hole in the head and have too much stuff in the repair queue already as it is.

I'm still playing with the CMC 880B and wanted to monitor the internal oscillator with something more accurate than I suspect that the 5216A I was using the other night is, so I grabbed the 5340A I bought a few months ago.  Powering it up reminded me that the fan sounded like a worn out cement mixer filled with empty beer cans, so I went on a snipe hunt at lunch to lube the fan in an effort to quiet it down a bit.  It's a 3-1/4" muffin fan and if course wedged tightly in place, so for now I gained access to the back side and dribbled a few drops of oil in through the power lead entry passage onto the shaft; it's a temporary fix but has certainly quieted it down considerably.

May rig up the Lucent standard later to feed the frequency counters a better reference.


I'm pleasantly surprised that the two frequency counters monitoring the 880B's oscillator only disagree by about 4-5 Hz - did not expect them to be that close at all!  (Not sure off the top of my head what the 5216A uses as a frequency standard, but given that it's about 51 years old I'm impressed!  I suppose it can be considered well aged at this point in time.)

-Pat

[Cubdriver]

Nixie DMM repair update


Zero offset impossible to calibrate, issue

...
...

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.

...
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.

OK so I looked at that special hand picked diode OA202.

Looking at the schematic, me thinks it's simply there as protection to clamp the input voltage to safe levels, just above 5Volts, the capacity of this meter.
So it's always reversed biased. I guess it's reverse current, would mess up the input of the diff pair, the base of transistor S1, and cause an offset. I guess.
That might be why they put a special diode in there, proably special in that it has a low reverse current.

So I pulled the data sheet for both that OA202, as well as the OA95, the latter being the general purpose one that litters the instruments, tens of them all over the place.

I was not disappointed. the OA95 is given for 30uA reverse current, and the OA202 is.... no less than 3 orders of magnitude lower ! 
And it's Silicium at that, not Germanium unlike the OA95.

OA202 is given, typical, between 0.01 to 0.1uA at ambient temp. I pulled it from the board and tested it. Measured the current with my MX56 DMM as it's the only thing I have than measure low currents. 0.01uA resolution, just enough for what we need here. Well, it measured... just above 0.01. So not only is it within spec, but it's at the very best of what it can be. Hand picked it was, and they did a good job at that it appears !
So, that diode is not the problem 
Since it's just a protection diode, I tried the meter without that diode in place, thinking well at least I am SURE it won't leak anything at all !
But yeah, doesn't change anything, of course....

Oh well, it was worth a shot, I did need to rule it out anyway, so that's done...

So now the next suspects are transistors S1 ans S2, that constitute the diff pair itself.

Need to swap them. Problem is that it's not easy...as you can see from the pic of the board above, the 4 transistors, the two of each pair on both comparators, are held in place all together by some black plate, and the metal cans of the transistors are glued to each other and to the plate. I had a go with my soldering iron on that glue : it barely makes a dent ! I am fucked, can't remove these transistors 

How to remove that glue ? By mechanical means maybe.. maybe try to use hack saw blade to weaken it, then a tiny / precision file to grind it down... 

The fun continues....

Vince, I'd conclude that metal bar is to keep the transistors in step with one another (coupled) thermally - I'd be careful mucking with it unless I was certain that one or more of them were the issue.  Any possibility of getting four replacements, and desoldering it as a unit to try the others as an uncompensated group to see if the behavior changes?  THEN if you positively conclude one in the bar is the issue you can sort out how to get it out to enable replacement.

FWIW.

Good luck with it!

-Pat

[factory]

A digital power analyzer for the Panaplex fans, US only auction*
https://www.ebay.com/itm/353993185981

That looks broken, too many dots on and I guess they didn't make a model with a binary display.

Well I usually say "if it isn't broken, don't buy it". 

Can't help thinking part of the problem, is that none of the volt range switches are pressed. This probably won't do much for the extra decimal points on the Amperes/Watt displays though. 

Did anyone bid for it? it made approx $54.

David

I was busy with work and missed the auction end.  Probably just as well, I need it like a hole in the head and have too much stuff in the repair queue already as it is.

I'm still playing with the CMC 880B and wanted to monitor the internal oscillator with something more accurate than I suspect that the 5216A I was using the other night is, so I grabbed the 5340A I bought a few months ago.  Powering it up reminded me that the fan sounded like a worn out cement mixer filled with empty beer cans, so I went on a snipe hunt at lunch to lube the fan in an effort to quiet it down a bit.  It's a 3-1/4" muffin fan and if course wedged tightly in place, so for now I gained access to the back side and dribbled a few drops of oil in through the power lead entry passage onto the shaft; it's a temporary fix but has certainly quieted it down considerably.

May rig up the Lucent standard later to feed the frequency counters a better reference.


I'm pleasantly surprised that the two frequency counters monitoring the 880B's oscillator only disagree by about 4-5 Hz - did not expect them to be that close at all!  (Not sure off the top of my head what the 5216A uses as a frequency standard, but given that it's about 51 years old I'm impressed!  I suppose it can be considered well aged at this point in time.)

-Pat

The 5216A uses a regular crystal, no fancy temp compensated or oven rubbish here  , note it's a PITA to get the PSU board out and yes this counter with burnt mains transformer is still on the bench, wired up to various external PSUs.


David

[syau]

just won a keithley 7001 unit on the evil marketplace. Hoping to get some inverter and VFD stuff out of it.
offer of 80$ was accepted ... VFD is shown operational in the photos ...
would be kind cool if it reVoved the silly money 2k1

I waited for ages for one of those to come up at  low price and none did. As far as I can tell they use the same inverter.

7001 is a good source of VFD for 2001/2, 24xx provided the display itself is still bright (which is hard to cmr by)

Note: The front panel pcb is not compatible along diff keithley meter as some has the inverter on the front panel while others have it on the main board.

[Cubdriver]

The nixies used are interesting in that they're dual anode tubes - odd digits on one anode and even digits on the other.

Hmmmm, like my Ferisol counter then !

Not knowing why such tubes even existed to begin with, I asked my old friend, if he had any idea. Not only did he know, but the answer was nothing I could have guessed on my own for sure.

He said back then it was a requirement from the military who, for reasons of ease of maintenance on the filed, required the tubes to be easy to remove and put on Noval sockets. Multiplexing the odd and even digits was the only way tube makers found to reduce the numbers of pins to satisfy this requirement...

Makes perfect sense - I realized that they fit in a noval socket last night when I looked up the data sheet for the tube; hadn't paid attention enough to notice when I had the board out.  It's definitely nice that it permits the tubes to fit in bog standard 9 pin miniature sockets, but in my opinion it's definitely a compromise readability-wise - 'regular' nixies have the digits in what seems like a very odd sequence - from the rear they go 1, 0, 2, 6, 9, 5, 7, 8, 4, 3 (based on looking at a broken B 5991 tube I have here); this seesm to be a compromise to enhance readability by reducing blockage by some numerals in front of the others.  The dual anode tubes require that they be grouped by odd and even, preventing this optimal arrangement.  I haven't looked closely at the order in the tubes used in the 880B, but definitely noticed that the numeral '1' is very prominent - at one point when I first started getting readings on the 880B, the uppermost digit was a 1 and at first glance I thought it was a neon lamp (like the elongated ones used in simpler meters to get a high digit that will only ever be a 1) as it appeared larger than the other numerals in the other tubes.  Later realized that it IS a tube, and that the 1 just looks off relative to the other digits.  Not necessarily a terrible thing, but I think I like the readability of the 'normal' ones a bit more.

Regardless, a very clever solution to a problem - simple multiplexing.

-Pat

[Cubdriver]

The 5216A uses a regular crystal, no fancy temp compensated or oven rubbish here  , note it's a PITA to get the PSU board out and yes this counter with burnt mains transformer is still on the bench, wired up to various external PSUs.


David

Damn!!  Even more impressive!!

-Pat

[Vince]

Vince, I'd conclude that metal bar is to keep the transistors in step with one another (coupled) thermally - I'd be careful mucking with it unless I was certain that one or more of them were the issue.  Any possibility of getting four replacements, and desoldering it as a unit to try the others as an uncompensated group to see if the behavior changes?  THEN if you positively conclude one in the bar is the issue you can sort out how to get it out to enable replacement.

-Pat

I agree with you but... the only way to be certain the trannies are the problem is... precisely to swap them... I am running out of options now 

The schematic does not say that these 4 trannies are matched, yet same schematic is not shy about annotations on specific components like this low leakage diode, or the special tempco resistors or hand picked Zener diode, on the Vref board... so chances are these 4 trannies are not matched. I can'"t believe that would NOT say it, if these were matched. Maybe because it's a comparator, a curde on/off switch, rather than a delicate op-amp working in it's linear region... that it does not need to be matched pairs ? The input leakage current however was definitely a concern of theirs because of the offset it can cause. but matching the hFE, maybe they didn't care for a comparator application ? 
Then again , I could pull all 4 trannies and test / compare their hFe on my 575 curve tracer !!! 
... once I am done restoring it that is 

They simply put that metallic plate around them to equalize the temperature among them, and deemed that sufficient.
And even then, that glue does not strike me at first glance like it could be some fancy thermal glue. To me it just looks like rock hard hot snot 

Also, there is no chance I could remove the 4 trannies in one piece still glued to that plate. For they have tiny legs and you can't even see them because of a plastic ring that covers them. A kind of standoff.

So pulling the lot is easy, but trying to put it back in, with leads that tiny that you can even see.. trying to line all those pins ? Not even a chance in a trillion. I would be screwed.

So I think trying to remove that glue, which again I don't even believe would be thermal glue, then removing the plate and extract just S1, would be the only feasible thing. Then of course put the plate bakc on and put fancy thermal glue as well BUT.... keep in mind that this thermal plate does NOT  create this offset problem, because the good board, gives perfect, zero offset FROM COLD. so the plate does not have even a chance to play a role in the offset issue.

So... really I don't need that plate for my trouble shooting purposes.



ALSO..... BREAKING NEWS, hot from the oven !!!!

Meter has been running for over an hour now, working well, as usual once it's hot.... I was scoping again some signals between the comparators and the gate logic, took some notes / waveforms down, and then all of a sudden I noticed that the meter STOPPED working ! Yes, the counter decided to free run again, like it normally does only from cold !!! 

GREAT !!!!

That means the condition of the meter degraded and now it's become a PERMANENT, SOLID failure !! That is .... MUCH EASIER TO TROUBLESHOOT !!!!!

Well, we shall see how it evolves, whether is really permanent or just comes and goes... but it doing it when hot after an hour, is definitely a first.

Please people, cross all your fingers and toes, let's pray that it has become a 100% solid fault.... please, please....   

[med6753]

I can insure it's a solid fault. Bounce it off the floor a few times. 

[Vince]

I might well do  Papy Smurf, I might well do !! Because this thing is driving me CRAZY !!!!! 

In the past 20 minutes or so it destroyed every little confidence I had here and there, I am now back to square one ! 

Well I mean... not really, but what happened in the past few minutes certainly forced me to reassess many things for sure....

So here is what happened. As I just said a few minutes ago, all of a sudden the meter started free running again, cool I thought, let's get probing around and finally fix it.

Then this happened :

1) All of a sudden, it started WORKING again !

2) 10 seconds later, back to free running again !!! 

3) That means it is NOT thermal related, but rather a bad / intermittent connection somewhere, so I was wrong ! 

4) Then as it was free running, I checked the output of the crystal oscillator : it was DEAD, nothing coming out of it, which means the counter/display is NOT "free running", it must be something else causing all digits to light up at the same time, AS WELL as the "overload/overflow" neon indicator coming on !

5) Some seconds later, still showing its "free running" pattern, Nixie tube on the second position from the right, decided to display a SOLID '0' digit ! Which means it is indeed definitely IMPOSSIBLE that the counter is "free running". 

6) Checked ramp generator : no ramp, dead as well.

NOOOOOO !!!!!  Let me jump off of the Bridge we have here in Saint-Nazaire, 60+ meters high, that's were the locals go to suicide (not even kidding), I might do it as well ! 


With so much going wrong all at once, immediate thoughts are...

1) Solder joints of the card edge connectors on the back plane PCB are broken, again... so had a quick look at them. Looked fine.
2) Power supply is having problems... so checked all the rails... WTF ! -12V reads -6V, +6V reads +12V, +12V reads +17.4V ?! 
3) A few seconds later meter starts working again... checking the rails again... they seem fine now ?!

Now a plain missing rail could be a dodgy solder joint, OK, I have already had that for the +200V rail.
However having -12V instead of -6 and +12V instead +6, and +17,4V instead of +12.... that's NOT a cracked joint !
I mean, not on the back plane anyway.

So... there must be bad joint on the PSU board itself, causing this insanity / weirdness... oh joy. 

I will be jumping off of that bridge soon, I am telling you !

[Vince]

EDIT !!!

OK it's back to "free running" again, but OK we now know it's technically, absolutely NOT "free running" .. so how shall I be calling this pattern it's displaying ?! 
Anyway, as I said that second Nixie board is not "free running" now, it's displaying a solid '0'. So... I just wiggled it a bit, and as I do this it gets the counter WORKING AGAIN !
Wiggle it some more, and it STOPS working, displaying this "free running" pattern again. I can wiggle that board and make the issue come and go as I wish !!!

NOW I am getting somewhere !!!  Aren't I ?! 

So... there must a cracked solder joint on the connector of that board, somehow causing all the other boards to go insane. Maybe the Reset signal ? The bad joint would send crap/illegal pulses to the counters on the other boards, and causing said counters to create garbage on their outputs, hence lighting up all their digits at the same time ?! Hmmmm.....

Let me investigate that.... stay tuned !!! 


1AM here though, getting tired.. will have to wait tomorrow I think..... good night.... 

[TaylorD93]

EDIT !!!

OK it's back to "free running" again, but OK we now know it's technically, absolutely NOT "free running" .. so how shall I be calling this pattern it's displaying ?! 
Anyway, as I said that second Nixie board is not "free running" now, it's displaying a solid '0'. So... I just wiggled it a bit, and as I do this it gets the counter WORKING AGAIN !
Wiggle it some more, and it STOPS working, displaying this "free running" pattern again. I can wiggle that board and make the issue come and go as I wish !!!

NOW I am getting somewhere !!!  Aren't I ?! 

So... there must a cracked solder joint on the connector of that board, somehow causing all the other boards to go insane. Maybe the rest signal ? The bad joint would send crap/illegal pulses to the counters on the other boards, and causing said counters to create garbage on their outputs, hence lighting up all their digits at the same time ?! Hmmmm.....

Let me investigate that.... stay tuned !!! 


1AM here though, getting tired.. will have to wait tomorrow I think..... good night.... 

reminds me of a fault i had on a HP3528A Frequency counter, had 3 possible clock sources, Oven based, External or the basic crystal on the main board.

When the Oven Option was fitted the crystal on the main board was removed. But there was still a buffer circuit etc.

There was an issue with these logic gates which buffered the onboard crystal clock, as when the Oven is fitted, a logic level changes to essentially inhibit the internal Clock source, blocking any signal. Although the Crystal wasnt fitted it still needed to be inhibited.

But when the External Clock was being used, there was a fault where it would work for a short time then count errors would increase and suddenly read zeros. I traced it to this logic gate which was not inhibiting and was drifting randomly to either a high or low.  I couldnt find a replacement logic gate so just cut the track, hey why do you need a signal from a Crystal cct which isnt even fitted when you have the choice of an oven Crystal or the GPSDO source.

Was a real pain to find, but was a solid state issue.

Hopefully yours is more of a connection problem and will be eventually found!

[Vince]

Oh so you had some fun too I see !!! 

I just fixed it !!! 

It was the same bad connection problem I had on the comparator board with the 200Volt pin : a joint that looked fine, but once you wick all the solder to expose the pad... said pad looks lie bare copper, solder didn't stick to it AT ALL !

So I wicked all 12 joints on the back plane for that Nixie #2 board, and sure enough I found AGAIN such a pad that solder joint didn't stick to !!   

so what was this pin/pad carrying ? You just can't make ths shit up, it was GROUND !!! 
NOt only that, but coming to this pad, was a ground wire coming straight from the transformer, then feeding it to the entire backplane ! So basically this one single dogy joint was not just keeping ground from getting to that one Nixie board, no, it was cuasding the ENTIRE intrument to lose ground AND cause the PSU to produce illegal voltages due to the secondary windings of the transformer not being properly referenced anymore !
So.... I redid all the joint. With some flux, I managed to get that crappy pad to take solder.

Now the meter works perfect, and if I wiggle that Nixie board, meter stays solid, like champ !!!
Hell, I can even get the Manual Trigger mode to throw garbage anymore, like it used to ! So maybe the trigger circuit was misbehaving not so much because of a dodgy switch, as it was because of a dodgy ground ?!    I will keep an eye on that for sure... but so far so good. Have "new" switches coming soon anyway, so can replace on the spot if still required.

Soooo.... I am so happy ! 

I have now a "working" DMM... I mean, one whose remaining faults are solid, so I can concentrate on those now 

That is : the input offset problem, and that Nixie board #2 that displays only '0' and '1', and the Vref board to "rebuild"/fix.

NOW... I can go to bed.... 

[Cubdriver]

Nice sleuthing, Vince!  Intermittents are maddening to troubleshoot.

-Pat

[TERRA Operative]

Might be worth spending a little time to check every solder joint now you've found a few bad ones....