EEVblog #804 - HP1740A Oscilloscope Repair - Part 2

[MauriceS]

The voltage regulators look awfully close to LM723's in TO-100 (metal can) package...

http://www.ti.com/lit/ds/symlink/lm723.pdf

[Xenix]

Amazing video! Very informative and entertaining for a hobbyist like me

[ornea]

Great video Dave.  I think I would have been scratching my head for some time on that one.

After repairs always amuses me when you show the customer the faulty component and they say "Is that ALL it was?".  Repairs take time to setup, tear down, go down those herring holes ,fix and put back together.

[Stonent]

Good video. Troubleshooting is never as simple as one might expect, it's as much art as logic and experience is king.

Chasing rabbits down the hole?

My side of the pond (at least in my circle) called it chasing rabbits into the tall weeds. 

At one point he said chasing red herrings down a rabbit hole.

[sakujo7]

Seems odd that it looked like a +15V problem...but it was a +5V bridge problem...but there wasn't a huge ripple on the output of the bridge. That 0.25V drop also seemed too low.

So, current theory is that the +5V was fed mostly or entirely from the output of U2 (which can operate on its own without an external transistor), through Q5 which was acting like a diode rather than a switch. This added +250mA onto the 15V rail, causing it to hit current limit.

Had the +15V been a bit beefier, this failure could have gone undetected.

[oldway]

Seems odd that it looked like a +15V problem...but it was a +5V bridge problem...but there wasn't a huge ripple on the output of the bridge. That 0.25V drop also seemed too low.

So, current theory is that the +5V was fed mostly or entirely from the output of U2 (which can operate on its own without an external transistor), through Q5 which was acting like a diode rather than a switch. This added +250mA onto the 15V rail, causing it to hit current limit.

Had the +15V been a bit beefier, this failure could have gone undetected.

Yes, that is the right explanation of what happened.

[German_EE]

Up the garden path chasing a red herring down a rabbit hole so that you can get your winner winner chicken dinner.

And they say that German is confusing 

Ah well, a good long repair video with lots of useful tips, just try not to work on live scopes with both hands please.

[dr.diesel]

Of the many hundreds of repairs i've made, never seen a diode bridge fail like that either.

[wojt]

I started to suspect the bridge when the ripple trace on scope showed to be 50 Hz. After a good bridge it should be 100 Hz. Apparently one of the diodes was gone already.

[EEVblog]

After repairs always amuses me when you show the customer the faulty component and they say "Is that ALL it was?".  Repairs take time to setup, tear down, go down those herring holes ,fix and put back together.

Yep, this one could have easily taken a lot more time than it did.

[EEVblog]

Amazing video! Very informative and entertaining for a hobbyist like me

Thanks, glad you liked it. And welcome to the forum!

[Rasz]

Nice one Dave.
Personally I would probably start by substituting power rails with bench psu, or isolating psu and loading rails manually, but your fault was so intermittent and random it would probably not help at all in locating it.

After repairs always amuses me when you show the customer the faulty component and they say "Is that ALL it was?".

Its a bad idea to show them the culprit from consumer relations point of view. Consumers dont understand what goes into repair, and if you show them failed part will often think they are getting ripped off, even more if you mention its $0.1. It might be good for your ego, not much else.

some rants about billing customers:

[jolshefsky]

This is probably one of Dave's most informative videos. It demonstrates that you need to be systematic about troubleshooting.

Anyone who's fixed anything fast knows the feeling of adoration when we take a guess, replace a part, and everything works—but that's kind of toxic. It makes us lazy.

Dave proposed a hypothesis then tested it based on the criteria. The resistance to abandon the theory was palpable in the video—the primary-side problems seemed so promising. But when it failed to manifest when testing the unregulated voltages. Then the hypothesis that there was a downstream problem with the horiz./vert. boards was also promising, but it too was dispelled when it wasn't an over-current problem.

The thing I take away from this is, "don't guess" or at least "don't take your guesses on faith." Divide the problem somewhere—the power supply is a good start: "is the power supply output okay?" If no, then you can start troubleshooting the power supply. Divide at another point: "is the transformer output good?", then another point, "is the unregulated power good?" ... and at any point you've narrowed it down to "point A is good" and "point B is bad" so you can figure there's a problem between point A and point B.

Another thing—this also from just getting older and less cocky—if there is troubleshooting information, follow it. I say "from just getting older" because when I was younger, I'd think "why do are we doing things the hard way? we can skip this and that..." but now I know enough to recognize when someone is more experienced than I, and to just do what I'm told until I understand why it's done that way. Troubleshooting information just might be written by someone with more experience than you—particularly with what you're currently working on. So use it until you hit a dead-end.

[station240]

You were kinda right about a current limit being involved, just it was the 50% limit as a result of half the rectifier being out of the circuit. The meter test setup shows the current drops in half at the same time the voltage does.

My guess is the diode bridge is physically constructed with the 4 diode dies bonded between the pins, so it's possible for one diode to loose connection. Maybe the pins got bent when it was originally installed.

Btw I do have a faulty diode that behaves like this, its one of the big bolt in ones and the input terminal got twisted so it's now loose.

[vindoline]

Great video. I love the troubleshooting videos! 

[DanielS]

I have seen diodes fail open or short. Intermittent though is something new.

[boffin]

Like many others, I'd love to see the output of that bridge, pre and post failure on a scope.

Any chance you can hook it up to 9VAC, give it a load, and slap the Rigol on it and take a couple of screen shots?

I assume one of the four diodes in the bridge goes open?

[retrolefty]

Like many others, I'd love to see the output of that bridge, pre and post failure on a scope.

Any chance you can hook it up to 9VAC, give it a load, and slap the Rigol on it and take a couple of screen shots?

I assume one of the four diodes in the bridge goes open?

 I'm not so sure. Recall the ripple voltage on shown on the scope was not bad when the voltage output dropped from the bridge. Again it seemed to just go high resistance which I too had never recall coming across. But the test you suggest would still be worth seeing.

[jancumps]

This is probably one of Dave's most informative videos. It demonstrates that you need to be systematic about troubleshooting.

...

Or even "that you need to be analytical...".  A fully systematic person would have followed the service manual first straight down into that rabbit hole

[tautech]

Like many others, I'd love to see the output of that bridge, pre and post failure on a scope.

Any chance you can hook it up to 9VAC, give it a load, and slap the Rigol on it and take a couple of screen shots?

I assume one of the four diodes in the bridge goes open?

 I'm not so sure. Recall the ripple voltage on shown on the scope was not bad when the voltage output dropped from the bridge. Again it seemed to just go high resistance which I too had never recall coming across. But the test you suggest would still be worth seeing.

I too wanted to all to see the frequency of that ripple. I missed it as Dave had V/div set very low.

wojt spotted it:

I started to suspect the bridge when the ripple trace on scope showed to be 50 Hz. After a good bridge it should be 100 Hz. Apparently one of the diodes was gone already.

[HighVoltage]

I started to suspect the bridge when the ripple trace on scope showed to be 50 Hz. After a good bridge it should be 100 Hz. Apparently one of the diodes was gone already.

Very good point, wojt

Great fixing video!
Thanks Dave

[Ampere]

What a strange problem. I was under the impression that diodes could only fail open. A catastrophic failure or none at all.

[pickle9000]

@Dave

Seeing as how the bridge made for such an entertaining video why not frame it and hang it on the wall with the episode number?

These troubleshooting skills are what makes a good engineer, hobbyist, repair person. You cover a great deal in the video, an excellent job. Congrats.

[BFX]

Finally most interesting video.
My suggestion is that you could make one repair video blog every 1-2 months
 
 

[nfmax]

I see from one of my old lab notebooks that on 9th July 1996, I replaced a faulty bridge rectifier in the low voltage PSU of my HP 1740A. For one minute I thought it was the exact same part, but no, it was A16CR4, on the +15V supply. My unit has serial number 1751G03795; what is the serial number of the unit you fixed? The bridge rectifier solder joints all look OK on mine (however I still need to fix the delayed timebase!)