Capacitors leaking on LG 50PC3D

Just barely, and only over a fairly long time period (months). There are many different types of electrolyte. Some are mild acids, but most are water or alcohol based.

Just clean up the mess, replace the big caps (even if they haven't blown up yet), and be done with it. If you have an ESR tester, it might help identify which caps are the likely culprits.

More:

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

here

is an esr meter repository to choose from a schematic to build your own esr meter !

have a look. regards,

--
---
Kripton

** Anyone who needs to check the ESR of a few electros can lash up a test rig in seconds - all you need is a bench audio generator and a basic scope or CRO as poms and Aussies like to call them.

Set the audio gen to about 100kHz and connect the output across the electro under test - then connect the scope probes direct to the same cap ( not the generator). You should see a small voltage at 100kHz on the scope.

By comparison with known good electros of similar ratings, one can note the residual voltages and determine if a given electro is OK.

Tests can be done while the electro caps are still in circuit - but be careful to make sure they are fully discharged first !!!

..... Phil

If one takes that literally, the scope probes /will/ be connected directly to the generator.

Did you leave something out?

Nope, it IS connected directly to the generator. And the capacitor across both.

You see a reduced value of signal due to the ESR of the capacitor vs the source impedance of the signal generator.

Pretty simple.

Not direct reading, but accurate comparison.

Jeff

"Jeffrey Angus" William Sommerwanker

** Connection of the scope direct to the electro cap is important.

Otherwise, the inductive reactance of the cable from the generator ( at

100kHz ) to the cap becomes included in the scope display of ESR - and is often more than the ESR of a good cap.

..... Phil

On 1/24/2011 5:20 PM Jeffrey Angus spake thus:

Skimming through a bunch of ESR meter schematics, this seems to be the game plan for most of them:

Oscillator --> attenuator --> cap under test -->

op amp/comparator --> rectifier --> meter

(osc. is usually around 50-100 kHz)

Gots to build me one someday ...

(the attenuator delivers only millivolts to the cap being tested, avoiding electrolytic polarity issues and any semiconductor junction resistances)

--
Comment on quaint Usenet customs, from Usenet:

   To me, the *plonk...* reminds me of the old man at the public hearing
   who stands to make his point, then removes his hearing aid as a sign
   that he is not going to hear any rebuttals.

So the cap goes before the scope leads?

--
Live Fast, Die Young and Leave a Pretty Corpse

Retch.

That would work if you knew the exact source resistance of the generator and that the generator is not going to try and protect itself from what it considers to be an AC short circuit load.

If you have 1 volt of output swing, and an assumed function generator output impedance of 50 ohms, the short circuit current will be about

20ma. Shove that into a 0.5 ohm electrolytic cazapitor, and you'll measure maybe 10mv across the cazapitor. That's kinda hard to see. Of course you could supply more than 1 volt but then you will have some problems trying to use this technique without removing the cazapitor from the PCB.

For what it's worth, I did it this way until I bought an ESR meter. Later, I decided to actually measure the output impedance of my function generator. The data sheet said 50/75 ohms, but it was anywhere between 30 and 100 ohms depending on the position of the dial and output step attenuator. So much for accuracy.

If you look carefully at the schematics supplied below, most of them put a known series resistance at the output of the generator section to the cazapitor. You could do the same thing with the function generator. 100 ohms in series with an alleged 50 ohms will not magically make the method more accurate, but it will reduce the error to a tolerable level.

Nothing is simple, but these examples might help:

Measuring Capacitor Self-inductance and ESR

Oscilloscope ESR Tester

99 Cent ESR Test Adapter

More links at the bottom of this page:

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

"David Nebenzahl"

** The idea of including an "attenuator " is misleading.

The electro under test needs to be driven with a known *current* at

00kHz - 20 to 50 mA is good.

Simply shorting my bench audio generator with the output level set to max produces 29.7 mV into a 1 ohm resistor at 100kHz or 30mA, near enough.

If the electro under test has an ESR of 100 milliohms - the above set up produces 3 mV rms across the cap which is easily enough viewed on most scopes.

Long as you connect the scope direct to the leads of the electro and know the current level from the generator - actual ESR values can be found.

.... Phil

ing

gn

Not likely to do much better than the MUL3333 ESR meter from Mat Electronics for $50. The main 'quirk' on the unit is the battery case holds 6 AA cells but it's wired as 3 parallel groups of 2 cells meaning just use 2 AA cells and skip the other 4. Very good value for the money. With any ESR meter there are 'gotchas' to beware of. Any ceramic caps in parallel with a 'lytic will mask the true condition of the 'lytic. If it reads bad, it IS but if it reads good it might not be. May times I've run into boards with many caps of the same value. If one of those is bad I change all of that value figuring if one is bad the others aren't far behind. Also, the output caps in SMPS need to be nearly perfect. 'Good enough' rarely is. When in doubt toss em out.

G=B2

"Jeff Liebermann"

** What kind of POS scope do you have ??? 10mV rms = 28.3 mV p-p.

My scope goes down to 5mV per division - then has a 5X mag switch if needed.

** Utter bollocks !!!

As your own example proves.

** Just set it to max and that's it for ESR testing.

** More bollocks.

The vast majority of electros have less than 1 ohm ESR values - so any error in the calculated drive current is TINY in comparison to a 50 ohm source.

The variation in ESR values between new electros of the same type is far more.

.... Phil

With any ESR meter there are 'gotchas' to beware of. Any ceramic caps in parallel with a 'lytic will mask the true condition of the 'lytic.

** Totally insane BOLLOCKS !!!!!

A 0.1uF ceramic has an impedance of 16 ohms at 100KHz - while a typical

100uF electro has an impedance of 0.16 ohms at 100kHz. That is 100 times less !!!

Imbecile.

If it reads bad, it IS but if it reads good it might not be.

** Only time that is ever true is if the electro is shorted - internally or externally.

May times I've run into boards with many caps of the same value. If one of those is bad I change all of that value figuring if one is bad the others aren't far behind.

** Not a bad idea - if all the electros have been subjected to the same temps for the same times.

Also, the output caps in SMPS need to be nearly perfect.

** Well, that may be so in some cases.

But designers usually leave a margin for component variations and some deterioration in electros.

Once the ESR of an electro rises by more than 50%, it may be time for replacement - cos it is now gonna keep rising.

..... Phil

ypical

ly or

Who said anything about 0.1uF? The Sony EQ-45 board in the DVW-500 Digital Betacam has a lot of 100uF surface mount 'lytics, many of which read 'OK' but when removing the 'lytic, IT reads very bad and the pads on the board read 'OK' without the 'lytic.

Can't argue with facts.

G=B2

ypical

ly or

Also, your math is wrong hotshot.

G=B2

Tek 2247A. I have several others, but that's the one I like to use.

Read what I wrote again. I said "1 volt of output swing" which means

1 volt peak to peak. Sorry for not being absolutely clear. Also, all the ESR test circuits I've seen use a square wave, which is normally not measured in RMS units.

Mine goes down to 2mv/div. About 1/3 of that is noise. My guess(tm) is that anything I measure has a built in 1mv error due to this noise.

I think the X5 or whatever magnification on your unspecified model scope usually refers to the horizontal sweep, not the vertical gain.

How so? If the cazapitor happens to appear across a semiconductor junction, as will happen with the output of an integrated switching regulator with a reverse protection diode on its output, the protection diode is going to conduct if fed more than about 0.6V peak to peak. I think (not sure) that the original Bob Parker meter ran at about 80mv peak to peak to avoid this manner of problem.

Lets just say I don't like running my test equipment flat out into an AC short circuit (the cap under test). My cheapo Leader LFG-1300S function generator has some kind of overcurrent protection circuit on the output, that limits the square wave output swing, but also mangles the waveform.

True, except that the variation between what's considered a good and a bad electrolytic is sufficiently small that source impedance makes a difference.

(Notice the odd variations across the voltage ranges). I find myself working at the bitter edge of some of the recommended maximum ESR values all too often. I sometimes have to flip a coin to decide if the cap is good or bad. The last thing I need is a blurry oscilloscope trace or an undefined generator source resistance.

Huh? I've taken bags of brand new electrolytic caps and tested them to see if there was any way to predict which ones would fail and which ones would survive. For a given value, voltage, and temperature, they're all about the same. I see wide variations in value (i.e. +80% to -20%), but the initial ESR just follows the value. However, after running a crude accelerated lifetime test on the caps (high ripple current and high temperatures), I find the good and the bad caps diverging in ESR.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

"Phil Allison"

Who said anything about 0.1uF?

** Read you own post - f*****ad.

" Any ceramic caps in parallel with a 'lytic will mask the true condition of the 'lytic. "

The Sony EQ-45 board in the DVW-500 Digital Betacam has a lot of 100uF surface mount 'lytics, many of which read 'OK' but when removing the 'lytic, IT reads very bad and the pads on the board read 'OK' without the 'lytic.

** Really ??

Can't argue with facts.

** Let me know when you have one - f*****ad.

BTW

The Matelectronics ESR meter is an total pile of shit.

.... Phil

** Anyone who needs to check the ESR of a few electros can lash up a test rig in seconds - all you need is a bench audio generator and a basic scope or CRO as poms and Aussies call them.

Set the audio gen to about 100kHz (sine wave) use full level and connect the output across the electro under test - then connect the scope probes direct to the same cap, not the generator - this is important.

You should see a small voltage at 100kHz on the scope - say 2 to 100mV rms.

By comparison with known good electros of similar ratings, one can note the residual voltages and determine if a given electro is OK.

If you put a 1 ohm test load on the audio gen - you can easily find the output current and then use the voltage readings on the scope to get actual ESR values for electros.

Tests can be done while caps are still in circuit - but be careful to make sure they are fully discharged first !!!

.... Phil

My Leader audio generator doesn't do 100kHz. What 'audio generator' does?

--
Live Fast, Die Young and Leave a Pretty Corpse