Confusing Diode Test

[Bodrey]

In this video () starting at the 14:34 mark this woman begins measuring 4 diodes. She determines that the first two diodes are OK, but the last 2 are bad. I certainly see why she deemed the last 2 she measured bad, but I'm confused about the readings she got on the first 2.

According to what I've read and seen in videos she should get a reading when the black probe is placed on the cathode and the red probe is placed on the anode. Instead, when she measures the first two diodes the meter gives her a reading when the red probe is on the cathode and the black probe is on the anode. The results she got were the opposite of what I was expecting to see. Confusing...

[RoGeorge]

Measuring while the diode is soldered in a circuit (any circuit) might be misleading, but it might give you a hint.  So she removes the suspicious diode, and measures it again outside of the circuit.  The bad diode is short-circuited (the instrument needle goes up - short - in both directions), while for the new (good) diode, the needle hoes up only in one direction.

I didn't watch the entire video, only 1-2 minutes after minute 14 you commented about.  Seems OK.

Modern DMM have a diode-test mode, in which they show the voltage drop (about 0.6...0.7V for a Si diode).  Older analog AVO-meters, like the one in the video, does not have a dedicated diode-test mode.  For that tipe of instrument, you would use the ohm-meter mode, and measure the diode resistance in both directions.  When the minus is to the bar line sign (the cathode), you will get low resistance.  With the +(red) on the cathode (line sign) you get high (infinite) resistance in the diode.

Of course, you have to measure the diode in an open circuit (desolder the diode, or lift at least one terminal from the PCB before measureing).

Another thing to take into account, for analog AVO-meters like the one in the video, zero ohms is in the right hand sign of the scale, so the needle doesn't move for high (infinite) resistance, and it goes up (to the right of the scale) for low resistance.  Else said, infinite ohms is on the left side of the scale, zero ohms is on the right side (maxim) of the scale. 

[Bodrey]

RoGeorge,

Thanks for the quick reply. I wasn't aware that for AVO-meters 0 ohms is on the right-hand side of the meter. That would explain why I saw it jump up. I guess it stands to reason that with analog meters the right side would need to be either 0 ohms or infinite resistance; it would have to be one or the other to distinguish between the two.

And, as it turned out, the third diode was not in fact bad. Only the fourth one was and she confirmed both after removing it from the board. The AVO-meter registered a false positive, perhaps because the third and fourth diodes were in series(?)

I hope most of this stuff ends up sinking in and I can retain it. 

[nonius_]

AVO-meters

That's a VOM;  AVO is a particular brand of VOM's

I consider AVO's special enough to be displayed in my living room, together with some other technical paraphernalia.

Besides a few DMM's and scopemeters there's also a VOM sitting here on the workbench. It's still quite useful and better than a DMM for certain tasks.

[Zenith]

Most analogue multimeters have the -ve probe positive and the +ve probe negative on the ohms range. Digital multimeters usually have -ve negative and +ve positive. For measuring resistors or continuity it doesn't matter. For go/no go tests on diodes and transistors, it can cause confusion.

 

[RoGeorge]

AVO-meters

That's a VOM;  AVO is a particular brand of VOM's

Never heard or used the VOM therm.  AVO-meter stands for Amperes-Volts-Ohms-meter, this is how I remember we used to call that type of instruments here, in Romania.  If there is also a brand AVO, to me that's just a coincidence. 

[Zenith]

VOM Volts-Ohms-Milliamps. "Vom" was used in the USA to refer to an analogue multimeter. CRO pronounced "Crow" like the bird was used in Australia for a Cathode Ray Oscilloscope.

AVO, Amps-Volts-Ohms was a UK manufacturer of analogue multimeters, and one of the first. They also made signal generators, valve testers and other test equipment.

[nonius_]

I stand corrected; turns out you (RoGeorge) were right after all; https://en.wikipedia.org/wiki/Multimeter

I've never come across the term 'avometer' so learned something new there myself.

Not surprized there aren't many AVO-brand meters in Rumania; there are quite a few in NL, but they were VERY expensive. The AVO-8 was made from the early '50s up until about 2010; in 2010 it cost about 1000 euro. Yes, 1000 euro for an analog meter! (and worth every penny).

(was it a post of you, or of another Rumanian, who explained how his father smuggled a Russian VOM inside a fish into the country? If that's the trouble you had to go to to obtain a Russian VOM, I can't imagine what it'd take to import/smuggle one from the UK)

And yes, the reverse-polarity of the measurement probes in resistance-mode can bite you if you aren't aware of it. But you normally figure out pretty quickly that something's a bit odd.

[Zenith]

For those particularly interested in AVOMeters, products of AVO Limited.

https://www.richardsradios.co.uk/avometersdata.html

I have two and I'm happy to leave it at that. I don't feel the temptation to acquire more.

[floobydust]

Testing diodes, it's important to know the multimeter's test current.

Modern DMM's are pretty weak- under 1mA, some 0.3mA test current. Basically a flea fart.
This is shit because a damaged semiconductor behaves a bit odd, if you get 0.55V instead of 0.7V it can look good when it isn't.

Old analog multimeters have high RX1 currents, some over 100mA- which cuts through oxides and corrosion on wiring connections, and makes the semiconductor-under-test work a little.
I found this is the reason a few techs swear by them. I like it for testing power semis like big IGBT modules.
But careful with the compliance voltage being crazy high like 22.5V for the old batteries on RX1000 range for example can push the gate junction or kill an IC.

[nonius_]

I'd say that for 99% (*) of the diode problems a DMM will work fine. But every once in a while you'll have to resort to other trickery to find the problem.

About two months ago I had a problem repairing a Philips PM5324 RF generator. There were the usual problems with bad electrolytic caps but after replacing them there was a pretty large 50Hz ripple on the unregulated DC. 100Hz ripple indicates bad caps but 50Hz ripple could mean only one thing: bad bridge rectifier, so I replaced the rectifier and the generator worked perfectly after that.

Out of curiosity I did a bit of an autopsy on the old bridge rectifier. Measuring in-circuit with the DMM showed nothing unusual. Measuring out of circuit also showed nothing unusual (see drawing with measured values). Measuring with the signature analyzer (which shows a curve, as opposed to the DMM which gives just one point on that curve) showed a slight deviation of one diode from the other 3 - but nothing dramatic. The next day, as I wanted to take a picture of the scope screen, it no longer showed a difference! So an intermittent fault when testing with a very small current (2mA and 20mA).

Only when putting quite a bit of load on the bridge, powering a 12V light bulb (measured current .56A), did the error manifest itself clearly. See scope pictures of the asymmetric rectified sinewave. Turned out the bad diode behaved as if it had a small resistance of about 5.5 ohm in series.

(In hindsight I should've used an analog VOM, that would probably have showed the problem immediately.)

To OP (a beginner): don't worry too much about my example, this is the first time in ~40 years that I've seen such an unusal failure mode. It's the exception, not the rule... Fortunately the 50Hz ripple was a dead giveaway. But always keep in the back of your mind that your measurement instruments provide clues and not definitive answers!

_{((*) 67.34% of all statistics are made up on the spot)}

[Zenith]

I'd say that for 99% (*) of the diode problems a DMM will work fine. But every once in a while you'll have to resort to other trickery to find the problem.

Yes. Usually diodes fail open circuit. Rarely they fail short circuit. Very rarely they show strange behaviour, temperature dependent or something else completely unexpected.

[Seekonk]

In 50 years of repairing electronics, I've never seen a power supply diode fail open. ORIGON had some stacked diodes made just for bias offset in amplifier output stages.  They had a run of them that would open, that is a special case.

I did have a diode problem that would have confounded many.  A guy built a Heath kit amplifier, and it didn't work.  He didn't want me to fix it, just tell him what was wrong for the $7.50 estimate fee.   The bias diode was in backward.  I think this was a factory assembled board.  The diode was properly installed as per the markings.  Except, the printed band was backward.   I reversed it and amplifier worked fine.  I told him it was a bad diode.   I wonder how many diodes he stuck in there backwards trying to get it to work.  I always wondered how that diode got marked backwards at the factory..

[floobydust]

I'd say that for 99% (*) of the diode problems a DMM will work fine. But every once in a while you'll have to resort to other trickery to find the problem.

Yes. Usually diodes fail open circuit. Rarely they fail short circuit. Very rarely they show strange behaviour, temperature dependent or something else completely unexpected.

I think you can't generalize diode failures like that. They can get overloaded, mechanical failure, die of old age/oxidation/ESD/ion migration/junction damage etc.
Actually the same for transistor junctions, I quickly learned that you need both resistance and diode-test measurements to catch all failures.
People are getting lazy and think a single diode-test voltage reading in the ballpark means it's good. Nope.

[SteveThackery]

Yes. Usually diodes fail open circuit. Rarely they fail short circuit.
[/quote]

What? That's the very opposite of what I've found. In my experience, diodes fail short circuit. I don't think I've ever had an open circuit diode.

[armandine2]

...as seen below, at the end the failed diode is still drawing current but is "burnt out" i.e. not emitting light.