ESR Values for Electrolytic Caps

[tigr]

Why you NEED it ?

6 capacitors. In all capacitors, the capacitance is in the range 1-1.5 uF. ESR 2-15 Ohm. How would you define a bad capacitor in the circuit without measuring ESR?

[PA4TIM]

I already explained why I do not measure ESR in situ (= in circuit) and why you should not do that too. More here: http://www.pa4tim.nl/?p=3775
Just use a scope or desolder them and use a decent LCR meter to measure them. Ever wondered why serious brands do not make ESR meters ? 

Some math. A 1 uF cap has a reactance of 1326 ohm at 120 Hz. If we take 15 ohm we have a D of 0.01. The few datasheet, I looked stated a D of 0.1 up to 0.3 as new.  That is an ESR of 132 ohm at 120 Hz.
So for a non switching powersupply 15 ohm at 120 Hz is nothing.

For audio signal it is a different thing. at 15 kHz the ESR is about the same as the reactance.  Replacing a cap with an ESR of 15 ohm for one with 0,001 ohm in a feedback loop can do Wierd things.(phase wise)

[tigr]

Just desolder them and use a decent LCR meter to measure them.

Who told you that it is necessary to remove capacitors? In 99%of cases, no.

[tigr]

Ever wondered why serious brands do not make ESR meters ?

Look here.

[tigr]

And also the tables are necessary to correctly understand the results of measurements.

[nuno]

Recently measured an unused 25yo 2.2uF 50V electrolytic, which measured 2.2uF and 3.2 Ohm ESR (using scope, fungen and calculator).

[PA4TIM]

Just desolder them and use a decent LCR meter to measure them.

Who told you that it is necessary to remove capacitors? In 99%of cases, no.

Nobody, if you know something about electronics you know why not to measure in situ. The point is that it often is possible but it can be way off so you simply can not trust the readings. It a sort of Russian roulette.

- Without analysing the schematic you do not know if there are parts parallel that can make a cap with sky high ESR look good. And that alone is reason enough !!
- A 100 kHz squarewave can have frequency components upto 1 MHz.  So inductive components like traces, inductors, resistors can polute the readings and skin effects start to play a role.
- the biggest problem for ESR in or out circuit is that you do not know what the measured value is. ESR at 100 kHz is almost never stated in datasheets. 
- The tables are miles apart from each other and often specially made for a meter (most ESR meter do not measure ESR but |Z| measured  at 100 kHz so you can not use them for another meter.
- There is no low ESR. Nowhere is stated when the ESR is low in absolute numbers. A low ESR cap from brand A can be twice as high as a low ESR cap from brand B.

But you are the specialist
 I just measured a few 470 uF 25V capacitor. I measured 0.182 ohm  and  0.223 ohm in situ and the same were 0.115 ohm  and  0.205 out of the circuit. I used a IET LCR meter, one from Hameg and my own designed ESR meter. Calibrated with ESI standard resistors.

ps: I do not consider a ledbar ESR indicator from BK as professional gear.

[wraper]

PA4TIM, there is on issue with measuring out of the circuit. When desoldering and therefore heating the capacitors, ESR often becomes significantly lower. In some cases, especially for small caps, it may fool you.

[PA4TIM]

Wraper, I know but it is good to mention because beginners tend to heat them up a lot due to crappy soldering tools.. I have a very good desoldering handpiece (Pace SX-100)   They are out within a few seconds so most times they are still cold. If not I let them cool. ESR for a hot cap can be half the value of a cold one.

This afternoon I checked a bunch of caps in a secutest that had some problems. I checked them again in situ and out.   My own design meter can do 10kHz to 100 kHz and as suspected (by me) ESR in situ at 12 kHz was much closer to the out situ measurements. Here I state the 100 kHz ESR as measured with my IET LCR meter at 100 kHz

So a little quiz: 
- 10 x 100 uF 35V measured in situ and a few minutes after desoldering and again 1 hour after removing just to be sure. Differences were neglectable.

out          in
0.209   0.178  - 14.8%
0.195   0.231  +18.4%
0.194   0.238  +22.7%
0.198   0.158  - 20.2%
0.214   0.183  - 14.5%
0.271   0.347  +28.0%
0.305   0.346   -13.4%
0.298   0.181   -39.3%
0.304   0.352  +15.8%
0.313   0.332  + 6%

So that gives an uncertainty % of -39 to +28 That is a lot. Besides that a cap can be shunted by others (that was here not the case) and so a cap with an very high ESR can still measure low. That increases the uncertainty even more and makes it useless other then a quick check to expose the really bad ones.
because If the ESR measures very high then there is a very big change that cap is bad. But why bother, if you can or will not test them the right way, just replace them.

And then there are other problem. A dead-shorted cap has a very low ESR. But to much leakage is not detected.  The same for dielectric absorption. And another parameter, the capacitance itself, the most important parameter. Multimeters are only usable for good caps so you need an LCR meter and desolder them after all.

There is one simple test that you can do that tells you quickly if the caps fail:  Measure the ripple voltage with a scope. After that desolder the suspected bad one and test it with an LCR meter. Look up the max value of D or tan-d in the datasheet if you want to be sure. And if you really want to exclude everything do a leakage test.

Sorry for the members of the holy ESR church but this it the way it should be done.  And yes, in cheap TV's and other consumer crap 99% of the times so much caps are bad that the chance of shunting a good one with a bad one is small. That lonely not completely shot cap that on its own tries co keep thing running will not influence much.

I do not repair consumer stuff and bad caps that are so bad that things do not work anymore are rare in the things I must repair. An ESR meter is for me not very useful. I can measure ESR, I'm not stupid, if it really was a good thing I would do it. I once was a believer, do you think I designed and build several ESR meters if I was a non-believer ?

So dear members of the holy ESR church , what to do with these caps ?  Replace them or keep them ?

[tigr]

A dead-shorted cap has a very low ESR.

Show the photo and the result of measurement by your device of this capacitor.

[tigr]

- A 100 kHz squarewave can have frequency components upto 1 MHz.  So inductive components like traces, inductors, resistors can polute the readings and skin effects start to play a role.

Give a specific example of some scheme.

[PA4TIM]

Give a specific example of some scheme.

Just basic electronics:
https://www.allaboutcircuits.com/textbook/alternating-current/chpt-7/square-wave-signals/
http://mathworld.wolfram.com/FourierSeriesSquareWave.html  (the math part)
A squarewave has in theory an infinitive amount of odd harmonics. How much depends on the rise and fall time. A faster rise/fall time is more harmonics.

Show the photo and the result of measurement by your device of this capacitor.

As soon as I find one that is not yet exploded (like a short tantalum) If a cap is short it leaks DC current. If it is dead-short the DC resistance is close to zero ohm. What do you think, is the resistance for the signal from your ESR meter in such a case ? 

But first answer my questions above, are my caps bad ? I know the answer without tables, but I like to hear from you which table is correct and should I use. You must know the answer, otherwise there is no use for your ESR meters 

[tigr]

Give an example instead of 100uF, some low-impedance capacitor. You can 3300uF/6.3v. Or you do not have such?

[PA4TIM]

Why not the 100 uF's ?  I have given you all the measured values so now you can tell me if they are good. What do you do if you had to repair this instrument ? call the owner, sorry, it has 100 uF caps in side . I do not know how to test them ? 
http://schneiderelectronicsrepair.nl/?p=518 There you see the instrument that houses the 100 uFs

 I have new caps  on stock up to 6800 uF. In several voltages and type. Besides that a lot of SMD solid polymers and tantalums.  I repair test and calibration gear for a living so I only use high quality caps (Panasonic)
 

[tigr]

Look here.

[PA4TIM]

So in that case some need to be replaced. But why that table ?

I rather use the table below. It is called a datasheet.  D is max 3 x 0,14 at 100 Hz. The measured D value of the worst cap of my list is 0.039
So all my caps are still as good as they can get. 
https://www.vishay.com/docs/28316/116rll.pdf  the complete datasheet

[tigr]

 

[PA4TIM]

Yes, an often made mistake, Impedance is not the same as ESR.   Besides that, my caps are 35V and not 10V .  Z=impedance= (R+jX),  R is ESR,  jX is reactance. R is frequency depended. The lowest value is not at 100 kHz. Often is the SRF of an through hole cap lower as 100 kHz, the cap is inductive above the SRF.
The lowest ESR is often somewhere between 50 and 100 kHz. That is de reason the datasheet for through holes state the ESR (in the form of DF or tan-d) at 100 or 120 Hz.

So I guess you did not read my page about ESR I linked several posts ago 

[ez24]

Reading

[tigr]

 

[PA4TIM]

That is an impedance table, not an ESR table. 

Lets take a cap from my measurements  with an ESR of 0,2 ,
Z of 100 uF @ 100 kHz , Z = (0.2-0.0159j) |Z|= 0.2 ohm (the squarroot of (R squared + jX squared))

[tigr]

Tomorrow I will make measurements of capacitors 100uF/35v. I will show the measurement in the circuit and out of the circuit. There is no difference.
1000uF/50v.ESR-0,01 Ohm

[The Soulman]

 

How about this one?

[tigr]

 

[tigr]

So a little quiz: 
- 10 x 100 uF 35V measured in situ and a few minutes after desoldering and again 1 hour after removing just to be sure. Differences were neglectable.

out          in
0.209   0.178  - 14.8%
0.195   0.231  +18.4%
0.194   0.238  +22.7%
0.198   0.158  - 20.2%
0.214   0.183  - 14.5%
0.271   0.347  +28.0%
0.305   0.346   -13.4%
0.298   0.181   -39.3%
0.304   0.352  +15.8%
0.313   0.332  + 6%

I measured my capacitors. The ESR is within 0.2-0.4 Ohm. All the capacitors are not new. You have good capacitors.