EEVblog #777 - Keithley 177 Microvolt DMM Repair

[EEVblog]

So many fail walks into the tall weeds, jumping to false conclusions, improper symptom recognition. However I couldn't stop watching the whole thing.     It was like reliving past battles.

To be fair, with my short playing with instrument before the video and also during the non-video breaks, there was no indication to me that the switches had a problem. It might look that way in the video perhaps, but it wasn't evident to me at the time, because yes, I am trying to shoot videos watching through the camcorder LCD screen.

I think good troubleshooting skills is as much an art as learned skill. I feel you proceeded much too quickly and didn't think carefully enough about each analysis, decision, and path you made as you walked down the problem.

Why?
I did a visual as you are supposed to.
Nothing mechanical was evident to me at the time (the switches only played ball later)
I had reason to suspect there was a component failure due to overload.
I had a simple and quick troubleshooting procedure so I started to follow it (that without the camera would have taken mere minutes)
A totally valid repair procedure.

Sure I could have sat there with the schematic for 30 minutes thinking why i would charge up and down etc and do the things it was, and I might have suspected the switched and then methodically started to play with then to reproduce the fault.
But meh, it went down how it went down. If I did this 10 times I would likely have gone down 10 different ways.

To be useful as a teaching moment, such a video troubleshooting task should be 'scripted' where you previously did the complete troubleshooting process and then create and follow a show and tell script to share your decision steps that lead to solving the problem.

I don't do scripts.
And I think it was a very useful video and did follow a reasonably correct procedure. I video were I just intuitively came to the conclusion that it was the switches from the get-go would have been the most boring video ever. I'm glad it didn't go down that way.
Sure in the end it turned out the problem could have been found earlier, but meh, the luck of the game.
I could have edited out all the dead ends, and made myself look like a repair genius, but that would not have made for an honest nor interesting video. I think there is great value in showing what happened warts and all.
If I had done what you suggested for this repair, it would be a 5 minute video with no electronics troubleshooting at all.
The warts and all real-time approach I take also gives views the time and ability to think about it and spot the fault themselves before I find it. There is huge value in that, and I think makes for a much more interesting video than a scripted route-step tutorial.

[Lightages]

I agree Dave. This video showed how to troubleshoot, and what can lead you down the garden path. No complaints from me! 

[SilverSolder]

A stab at a Golden List of Troubleshooting Intermittent Faults:

1. Thou shalt carefully inspect the boards and components looking for burned or damaged components, cracks in the PCB, bulging/leaking capacitors, etc., and repair as necessary.

2. Thou shalt measure / verify supply voltages and repair as necessary.

3. Thou shalt clean any switches, relays, pots, and connectors in the signal path with high quality cleaning fluids.  (This fixes 90% of older equipment problems in my experience)

4. Still problems? Thou shalt wriggle power resistors and any other PCB mounted component that gets hot, looking for open/bad, cracked, or intermittent solder joints (especially for lead free boards).

5. Still problems? Thou shalt use freeze spray, hot air gun etc. to thermally massage suspect components to see if that provokes a fault.

6. Still problems? - Now it's time to get creative and start a "divide and conquer" process of elimination, try to divide the circuit in half, inject signals if necessary, and verify each half at a time (binary search strategy).

[AF6LJ]

I agree Dave. This video showed how to troubleshoot, and what can lead you down the garden path. No complaints from me! 

While I agree mostly I think what was left out was the experience of a more adapt technician.
Note to Dave;
Don't get mad at me for saying this but it is true; Engineers make poor technicians, a tech is much better at troubleshooting. That is what we do from a repair standpoint we have component history and history that goes with various pieces of gear we have worked on. We know the points of failure and can set the analysis aside and get eh job done. The tests Dave-2 did should have been done in the first ten minutes, sorry but when you work as a tech for a living your job depends on finding the obvious. Dirty switches are obvious.
You do what you do wall as an engineer, there are damned few engineers who are also good techs, they are the one in twenty...

[AF6LJ]

A stab at a Golden List of Troubleshooting Intermittent Faults:

1. Thou shalt carefully inspect the boards and components looking for burned or damaged components, cracks in the PCB, bulging/leaking capacitors, etc., and repair as necessary.

2. Thou shalt measure / verify supply voltages and repair as necessary.

3. Thou shalt clean any switches, relays, pots, and connectors in the signal path with high quality cleaning fluids.  (This fixes 90% of older equipment problems in my experience)

4. Still problems? Thou shalt wriggle power resistors and any other PCB mounted component that gets hot, looking for open/bad, cracked, or intermittent solder joints (especially for lead free boards).

5. Still problems? Thou shalt use freeze spray, hot air gun etc. to thermally massage suspect components to see if that provokes a fault.

6. Still problems? - Now it's time to get creative and start a "divide and conquer" process of elimination, try to divide the circuit in half, inject signals if necessary, and verify each half at a time (binary search strategy).

That is actually good advice and will get the problem solved 95% of the time.

[dentaku]

I suspect that good troubleshooting procedures and making a video simultaneously is an impossible task. As a 'professional troubleshooter' to component level for most of my working career (now retired), I was yelling at you almost non-stop watching this video.

 So many fail walks into the tall weeds, jumping to false conclusions, improper symptom recognition. However I couldn't stop watching the whole thing.     It was like reliving past battles.

 I think good troubleshooting skills is as much an art as learned skill. I feel you proceeded much too quickly and didn't think carefully enough about each analysis, decision, and path you made as you walked down the problem. That's probably because you also have to keep feeding audio talking points while trying to troubleshooting at the same time. Most brains are too small to pull that off successfully. 

 To be useful as a teaching moment, such a video troubleshooting task should be 'scripted' where you previously did the complete troubleshooting process and then create and follow a show and tell script to share your decision steps that lead to solving the problem.

 All that said, intermittent problems are the worst case symptoms to have to deal with. 

 

Shahriar has had a good stretch of interesting repair videos lately. He lives a place where it's easier to find good equipment to fix than Australia too.
https://www.youtube.com/channel/UCKxRARSpahF1Mt-2vbPug-g

[AF6LJ]

Note to Dave;
Don't get mad at me for saying this but it is true; Engineers make poor technicians, a tech is much better at troubleshooting.

Could it be said that engineers know how things work and technicians know how things fail?

I was thinking clean the switches. But rather than feeling righteous I admit it was because I don't know enough about ADC integrators to be misled.

I do like to see old test equipment rescued. I hope Dave makes more repair videos.

Something like that, some technicians know how things work, some engineers know how things fail.
Just like some engineers do make good technicians and some technicians know engineering.
When I use to work for Loral Data Systems we had a few engineers who were really great technicians. One of them designed the first PLL synthesized L band transmitter (200 channels 1.43GHZ to 1.54GHZ). Electronics was his hobby and I might add his first name was Dave.

So it depends; every once and a while you come across a truly great engineer, very seldom do you find a really great engineer who is also a really good technician. They are out there....

[poot36]

How about fixing the Tektronix scopes that you got in a previous mailbag?

[EEVblog]

While I agree mostly I think what was left out was the experience of a more adapt technician.

Sure, and those who repair vintage gear like this all time would instantly know with maybe 90% certainly the switches would come first.
As retrolefty said, try doing this as a video blog and shooting and talking at the same time trying to get useful content in the bag, it's a completely different game.
The only thing I did "wrong" here is not detect the bad switches earlier.
Should I have suspected the bad switches from the get-go? In a bit of kit this vintage, perhaps, and I did think of it as I was playing. But as the person who sent it in said, it worked fine just days before the event that killed it. And when I first played with it there seemed to be no correlation between the switches and what was happening. But I eventually came to that conclusion.

Don't get mad at me for saying this but it is true; Engineers make poor technicians, a tech is much better at troubleshooting.

I started out as a repair tech, thank you very much 

That is what we do from a repair standpoint we have component history and history that goes with various pieces of gear we have worked on. We know the points of failure and can set the analysis aside and get eh job done. The tests Dave-2 did should have been done in the first ten minutes, sorry but when you work as a tech for a living your job depends on finding the obvious. Dirty switches are obvious.

I did do those tests! as part of an initial touchy feely before the video, and there was nothing obvious. And several things lead me away from suspecting the switches as well. I can hear people saying "yeah, yeah, excuses excuses" but hey, that's the way it went down. Meh.
I would have been mightily embarrassed if I had never suspected the switches and David2 came along and found it in 15 minutes. But I had already suspected the switches twice by that point, and was literally the next thing I was going to attack in-depth, it's all it could be by that point.
Also, as I've said, the tests I did, whilst 30 minutes of video and commentary, probably amount to 5 minutes actual work before I came to the conclusion that the switches were at fault. So this thing, in the real world would have been fixed in maybe 10 minutes.

[EEVblog]

So it depends; every once and a while you come across a truly great engineer, very seldom do you find a really great engineer who is also a really good technician. They are out there....

It depends on what you do every day.
A good TV repair tech will fix a TV faster than anyone, engineer or tech.
A good system production tech will likely fix system production issues 10 times faster than that top TV repair tech.
A good RF design engineer who troubleshoots RF prototype circuits all day long will likely repair an RF amp quicker than any tech ever could.
It's horses for courses.

I used to be a repair tech on Z80 based CCTV gear, and I was the best one they ever had. At point I would have wagered I could repair that gear quicker and more efficiently than anyone else on the planet.
Would I be able to do that now 25 years later? Not a chance, I'd bumble around like anyone else would.

[EEVblog]

Shahriar has had a good stretch of interesting repair videos lately. He lives a place where it's easier to find good equipment to fix than Australia too.

Yes, gear is not easy or cheap to get in Oz. Majority of US sellers won't even ship outside the US or Canada.
BTW, I just scored a Fluke bench meter. Sold as not working, but I suspect it will be fully working...

[tech5940]

So happy to see a new repair video, thanks Dave!


Sent from my iPad using Tapatalk

[EEVblog]

How about fixing the Tektronix scopes that you got in a previous mailbag?

Those scopes are notoriously horrible to fix. Uses a ton of custom parts. My spidy sense tells me it would just end up like the DSA repair.

[beaker353]

Over the past decade as a industrial equipment bench tech I would say about 95% of my repairs come down to either the power supply or something electro-mechanical. I always throughly check both before pushing into component level tests.

-EM


Sent from my iPhone using Tapatalk

[dentaku]

Shahriar has had a good stretch of interesting repair videos lately. He lives a place where it's easier to find good equipment to fix than Australia too.

Yes, gear is not easy or cheap to get in Oz. Majority of US sellers won't even ship outside the US or Canada.
BTW, I just scored a Fluke bench meter. Sold as not working, but I suspect it will be fully working...

Shahriar also has the advantage of working at Bell Labs so his dumpster has way higher end stuff in it too, but good Ebay finds are definitely much easier to get in New Jersey.

[EEVblog]

Over the past decade as a industrial equipment bench tech I would say about 95% of my repairs come down to either the power supply or something electro-mechanical. I always throughly check both before pushing into component level tests.

In my repair days, almost all the repairs were component level failures.

[AF6LJ]

So it depends; every once and a while you come across a truly great engineer, very seldom do you find a really great engineer who is also a really good technician. They are out there....

It depends on what you do every day.
A good TV repair tech will fix a TV faster than anyone, engineer or tech.
A good system production tech will likely fix system production issues 10 times faster than that top TV repair tech.
A good RF design engineer who troubleshoots RF prototype circuits all day long will likely repair an RF amp quicker than any tech ever could.
It's horses for courses.

I used to be a repair tech on Z80 based CCTV gear, and I was the best one they ever had. At point I would have wagered I could repair that gear quicker and more efficiently than anyone else on the planet.
Would I be able to do that now 25 years later? Not a chance, I'd bumble around like anyone else would.

I'll bet you could once you been on the bicycle you never forget how to ride it.
I bought an S-Band transmitter off E-Bay a couple of years back, happened to be one I and an assembler built when I worked at Loral Corp. I wanted a piece of my past and a paperweight. As soon as I opened it up that experience all came flooding back. I have no doubt I could build / align those transmitters twenty nine years later.
I think you might be surprised as to how much you retain, it just takes a little stimulus to get it back out and into your hands.

[AF6LJ]

Shahriar has had a good stretch of interesting repair videos lately. He lives a place where it's easier to find good equipment to fix than Australia too.

Yes, gear is not easy or cheap to get in Oz. Majority of US sellers won't even ship outside the US or Canada.
BTW, I just scored a Fluke bench meter. Sold as not working, but I suspect it will be fully working...

Good deal, on the Fluke.
The US market for used (affordable) gear is starting to dry up

[EEVblog]

I'll bet you could once you been on the bicycle you never forget how to ride it.

Yeah, but I don't think I'd be as good as I was back then. I know a lot more now, and sometimes that can be a hindrance
i.e. a lot more complex stuff stuff that could cause various faults jumps into my head instantly more now than it did back then. There is definitely something in the "engineers can overthink things" line.

There is also a fair bit of psychology involved too. In the case the of the Keithley here, and most of my video repairs, I go into it secretly hoping it's going to be some elusive electronics fault that I can track down like an episode of Columbo. And that probably subconsciously influences how I do things.

[c4757p]

Over the past decade as a industrial equipment bench tech I would say about 95% of my repairs come down to either the power supply or something electro-mechanical. I always throughly check both before pushing into component level tests.

In my repair days, almost all the repairs were component level failures.

Interesting, what sort of things were you repairing? Electromechanical problems are by far the ones I see the most as well.

[krivx]

Over the past decade as a industrial equipment bench tech I would say about 95% of my repairs come down to either the power supply or something electro-mechanical. I always throughly check both before pushing into component level tests.

In my repair days, almost all the repairs were component level failures.

Interesting, what sort of things were you repairing? Electromechanical problems are by far the ones I see the most as well.

I feel like the amount of human handling is going to influence things - think of broken switches in things that get touched often versus dried capacitors in things that don't. I would guess that CCTV systems are mostly set-and-forget, so mechanical failure doesn't happen (assuming stationary cameras...)

[SeanB]

Bad connections are common, along with failed power supplies. Funny enough the most reliable connectors are the F connectors, they only give grief after a few years. A shot od spray lube on the end and reconnect and it works again for a decade. Worst is those phono/RCA connectors, they do horrid things.

Then you get the painters who paint everything, including the camera lenses.......

[retrolefty]

I'll bet you could once you been on the bicycle you never forget how to ride it.

Yeah, but I don't think I'd be as good as I was back then. I know a lot more now, and sometimes that can be a hindrance
i.e. a lot more complex stuff stuff that could cause various faults jumps into my head instantly more now than it did back then. There is definitely something in the "engineers can overthink things" line.

There is also a fair bit of psychology involved too. In the case the of the Keithley here, and most of my video repairs, I go into it secretly hoping it's going to be some elusive electronics fault that I can track down like an episode of Columbo. And that probably subconsciously influences how I do things.

 But grasshopper, Colombo episodes were all scripted, and you don't do scripted. 

[AF6LJ]

I'll bet you could once you been on the bicycle you never forget how to ride it.

Yeah, but I don't think I'd be as good as I was back then. I know a lot more now, and sometimes that can be a hindrance
i.e. a lot more complex stuff stuff that could cause various faults jumps into my head instantly more now than it did back then. There is definitely something in the "engineers can overthink things" line.

Technicians fall into that trap also, it has happened to me a few times and I have seen it often happen to others, especially being an RF tech for nearly all my career.

There is also a fair bit of psychology involved too. In the case the of the Keithley here, and most of my video repairs, I go into it secretly hoping it's going to be some elusive electronics fault that I can track down like an episode of Columbo. And that probably subconsciously influences how I do things.

Here is a good one for you that had three senior techs baffled (yes I was one of the three).
When I worked at Loral Corp I was building 1.5 GHZ FM exciters, they would drive a ten watt three transistor PA(aligning testing and some assembly but the majority was left to the assemblers...)
We had the second transmitter of 175 kicked back for failing vibration testing. The transmitter signal would start breaking up on the spectrum analyzer and it would do it at a nice linear rate. as you brought up the random vibe level the carrier would become more broken up in proportion to the amount of Gs the transmitter was subject to.

So the transmitter gets kicked back to QA, the top assembly tech ruled out the PA and the voltage regulator and kicked the exciter back to me. I ran the exciter did some "thumping on the Loop can, Miser board, VCO and Crystal oscillator modules. The loop was broken and I came to the conclusion that the VCO (for whatever reason) was sensitive to mechanical shock. The VCO is potted by the way. So back to the tech who aligned it, it goes..
Four days later he sends it back with a half dozen pages of work done and it is suppose to be good... I preform the same test and say it is still bad... Well I can't just send it back and I am busy turning out those other 174 exciters so reworking the VCO myself is out of the question...The third trip around my Boss's boss gets involved.
He says "Well Sue maybe it is the loop" I say "No sir I married the VCO with another exciter and the VCO is bad...
So now this exciter becomes a political football...
We set it aside and as we approach the end of the contract the exciter takes a trip to engineering. One morning an Engineer (who just happened to be named Dave (not the first Dave but another one) comes back with my exciter minus the VCO. He says this is good we are holding the VCO...
By this time that VCO had all the active and passive parts in the VCO cavity replaced still didn't work. There was an isolator a 100pf feedthrough cap, a 100pf feedthrough and a CK05 type ceramic cap soldered to the outside. of the VCO cavity. This almost stumped the Engineering department until the last remaining parts were tested with a newly purchased vibration probe. Turns out that CK05 ceramic cap had become piezoelectric, and was modulating the VCO supply voltage.

That is the kind of problem I know you would love to solve.

[poot36]

Those scopes are notoriously horrible to fix. Uses a ton of custom parts. My spidy sense tells me it would just end up like the DSA repair.

But even if it does it would be a great teaching tool and if it doesn't then you get a working high end (for the time) scope.