ESR Values for Electrolytic Caps

[The Electrician]

Yes, an often made mistake, Impedance is not the same as ESR.

It's true that given Z = R + jX, ESR (which is the same same as R here) is not the same concept as |Z|.  ESR is only a part of Z; the other part is X.  But the low cost ESR meters measurement of |Z| is a perfectly adequate substitute for measurement of ESR if the measurement is done at 100 kHz and component being measured is a typical aluminum electrolytic.

I explain why this is true in another thread: https://www.eevblog.com/forum/projects/impedance-lcr-esr-meters/msg459262/#msg459262

Your calculation in reply #47 shows this equivalence; the numerical value of ESR is the same as |Z|:

"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]

100uF/35v.

[PA4TIM]

Yes, an often made mistake, Impedance is not the same as ESR.

It's true that given Z = R + jX, ESR (which is the same same as R here) is not the same concept as |Z|.  ESR is only a part of Z; the other part is X.  But the low cost ESR meters measurement of |Z| is a perfectly adequate substitute for measurement of ESR if the measurement is done at 100 kHz and component being measured is a typical aluminum electrolytic.

Yes I know, you are right, as the capacitance increases Z comes closer to R. A bit a bad example. Now beginners think it never makes a difference because they do not understand complex numbers and math.

But something else, as far as I know the 100 kHz in datasheets is measured with an impedance analyser using a true AC sinewave. The signal from my IET is a sinewave, as goes for all my other LCR meters and bridges. But many ESR meters use a squarewave with an DC offset.

My own ESR meter reads the same value with a 12kHz  squarwave as my IET at 100 kHz(sinewave), but not around 100 kHz, then they differ. I have done a lot of ESR measurements with VNA's. But that is also with a sinewave.

But Tigre does not understand my point.
- in situ measurement CAN be usable, but the problem is you need to know if there are components parallel that influence the readings. In many circuits there are at least several 1 to 100 nF caps, for instance on the Vcc pins of IC's. So a reading can be spot on or way off. As the reading is to high you replace a good cap, so no problem and a rare occasion, but the change the reading is to low is bigger and that makes in situ measurements handy as an aid and not a replacement for real troubleshooting with scope and multimeter.

- The table thing. After I told my caps are good he comes with yet another table that shows the impedance and according that table they are bad but he says they are good . The ESR is in my case on its own already higher as the reactance. But according the datasheet they are still within the new-specs. However, if I was believing this table I would have replaced them and then still the secutest would not have worked. I knew they must be good because ripple voltage was good but I did it just to see how usable an in situ measurement was.

So you get measurment data that could be way off, and the dangerous part is that the biggest chance is, the cap is worst as measured.
you can not really check the value because the tables are to different from each other and based on "nothing" without data about the caps and meter.

But on the upside, the ESR meter is cheap and makes it possible for beginners to repair some consumer gear without much knowledge.  (because there caps are number 1 and often so bad the ESR is very high). It can be a handy help to save time for pro's to0. They often know upfront the problems with TV's and know the ESR values they must see. The secutest is something like that. If I check diodes I know there are two that always measure a short in the diode function of the DMM.

And what about this: see picture

[rsjsouza]

Interesting discussion; I learned since an early stage in electronics to be very careful with in situ readings of component parameters, as I was constantly being fooled by skewed readings. Later in life I found out that any electronic device should be furnished with characterization curves similar to the ones provided by diodes and transistors. In the absence of that for passives, the next best thing is to (1) try and go by with the spot numbers of their datasheets - usually the D factor at 120Hz and sometimes the Z at a higher frequency - and (2) with a complete understanding of the circuit to evaluate if ESR or other parameters can exert an influence. I could not yet find a reliable table of ESR (or D) ratings around, but as PA4TIM and others have said there are simply too many factors to create a reliable one.

Another excellent reference is the discussion below, in particular the posts by the resident expert free_electron.

https://www.eevblog.com/forum/beginners/esr-meter/

[The Electrician]

Yes I know, you are right, as the capacitance increases Z comes closer to R. A bit a bad example. Now beginners think it never makes a difference because they do not understand complex numbers and math.

I don't think it's a bad example at all.  If they carefully read what I said they won't think it never makes a difference.  They don't need to understand complex numbers to understand my conclusion.  They can see what I'm talking about in the image showing |Z| and ESR over a wide frequency range with understanding complex numbers.  I said that ESR and |Z| at 100 kHz are the same for typical aluminum electrolytics, and that means no film capacitors, no tantalum, no polymer, no MLCC capacitors, etc.  In the next few posts after the one linked above, I explain that there a few exceptions, but reply #14 in that thread: https://www.eevblog.com/forum/projects/impedance-lcr-esr-meters/msg459471/#msg459471 shows a range of capacitor sizes that more than span the range of cheap ESR meters so that within the allowable range of capacitor size for those ESR meters, such capacitors still have |Z| and ESR the same at 100 kHz.  Low cost ESR meters do an adequate job of measuring ESR at 100 kHz for typical aluminum electrolytics.

But something else, as far as I know the 100 kHz in datasheets is measured with an impedance analyser using a true AC sinewave. The signal from my IET is a sinewave, as goes for all my other LCR meters and bridges. But many ESR meters use a squarewave with an DC offset.

True enough, but it doesn't matter if they correctly measure |Z|.  In the referenced thread, I discussed what errors might occur due to using a square wave excitation voltage.  I showed that if ESR is constant from 100 kHz to higher frequencies, square wave excitation doesn't lead to an error.   I gave an example (axial lead capacitors) https://www.eevblog.com/forum/projects/impedance-lcr-esr-meters/msg459434/#msg459434 where square wave excitation could lead to an error in the measurement due to the higher ESL of axial lead capacitors; the reading was in error by a factor of 2.  A factor of 2 error won't prevent the user from detecting a failed capacitor's greatly elevated ESR; in case of a failed capacitor the ESR will be many times larger than it should be.

The low cost ESR meters aren't precision instruments anyway.  I wouldn't use one to collect data to make a table of capacitor ESRs.

My own ESR meter reads the same value with a 12kHz  squarwave as my IET at 100 kHz(sinewave), but not around 100 kHz, then they differ.

                                                                                                   ^^^^^                                  ^^^^^
I don't understand this; both frequencies pointed to above are the same.  That doesn't make sense.  Did you intend for the first one to be 10 kHz rather than 100 kHz?

If you are measuring a capacitor whose ESR is constant at frequencies above 100 kHz (to at least 1 MHz), then you should get the same result with square wave excitation as with sine wave.  Why do you suppose your own meter gets a different result at 100 kHz?

But Tigre does not understand my point.
- in situ measurement CAN be usable, but the problem is you need to know if there are components parallel that influence the readings. In many circuits there are at least several 1 to 100 nF caps, for instance on the Vcc pins of IC's. So a reading can be spot on or way off. As the reading is to high you replace a good cap, so no problem and a rare occasion, but the change the reading is to low is bigger and that makes in situ measurements handy as an aid and not a replacement for real troubleshooting with scope and multimeter.

- The table thing. After I told my caps are good he comes with yet another table that shows the impedance and according that table they are bad but he says they are good . The ESR is in my case on its own already higher as the reactance. But according the datasheet they are still within the new-specs. However, if I was believing this table I would have replaced them and then still the secutest would not have worked. I knew they must be good because ripple voltage was good but I did it just to see how usable an in situ measurement was.

So you get measurment data that could be way off, and the dangerous part is that the biggest chance is, the cap is worst as measured.
you can not really check the value because the tables are to different from each other and based on "nothing" without data about the caps and meter.

But on the upside, the ESR meter is cheap and makes it possible for beginners to repair some consumer gear without much knowledge.  (because there caps are number 1 and often so bad the ESR is very high). It can be a handy help to save time for pro's to0. They often know upfront the problems with TV's and know the ESR values they must see. The secutest is something like that. If I check diodes I know there are two that always measure a short in the diode function of the DMM.

I completely agree that the repairman needs to determine if there are other capacitors in parallel with a capacitor being measured in circuit, or any other factors that make it necessary to remove a capacitor to properly diagnose its health.

I also agree that the various tables are a general guide only.  As a repairman gains experience, he will learn what a good ESR is for a particular capacitor.  Better than using tables is to just compare to the ESR of a known good capacitor.    It's the capacitors that are marginal that are troublesome.

When I used to repair TVs many years ago, what was known in the business as a "callback" was a big money loser.  If there was any question about a capacitor's health, I replaced it; it wasn't worth the cost of a new capacitor to take a chance on having to deal with an unhappy customer and repair the equipment again.

[tigr]

There is also an interesting discussion.
https://www.eevblog.com/forum/beginners/capacitors(c-esr-losses-tables-meters-etc)/

[tigr]

Difficulties in estimating the values of ESR, happen with such capacitors, which are not present in the tables.
2,2uF/10v

[PA4TIM]

One last try.

Measurements out of circuit:

The good way:
-Take a LCR meter that gives you D (=tan d).
- Look up the datasheet
- lookup D (=tan d = DF = 1/Q = related to loss angle)
- and the frequency they measured that, often 100 or 120 Hz (it was 1 kHz for a long time)
- lookup the max value of D, most times 200%
- Take a real LCR meter, set it at the stated frequency
- Measure C, is it within specs ? often +/- 20% when they are new, how low the circuit tolerates is an other question.
- A real LCR meter or bridge will give you D
- If both are within specs the cap is good. (if you want to go all the way you can measure leakage. (often stated for 2 or 3 minutes, this is not the working voltage leaking test, that is a test you do for safety.)
 
The easy way, but not always a good way
- replace the caps by the same brand an type new ones (nothing wrong with this)
- or with a suitable replacement with the same specs
- do not kill the pcb by using crappy desolder tools
- this did not help ? Do some real trouble shooting with things like a scope and multimeter.

I only repaired 3 TV's and non of them had dead electrolytics. (a shorted 1 nF 1kV cap, a dead 4 MHz Xtal and a shorted mosfet) but my satellite receiver needed over 40 new caps.)

The hard way:
- fire up your ESR indicator
- measure the cap in circuit
- spend an afternoon looking for tabels that gives the ESR for 2 uF and works for your indicator (so a Z or an ESR table)
- Then decide witch table could be the correct one.
- You do not know so start a bunch of new topics to get the same answer in all of them...or roll a dice
- before going mental you just replace the cap to be sure

3rd way:
- Fire up your impedance meter (a sweeping one or a simple DIY self ESR meter that does not measure ESR but Z, how convenient    )
- look up datasheet, note impedance Z at 100 kHz.
- is measured Z is withing the stated specs, the cap is good.

4th way, involves some math:
- fire up your real ESR meter
- measure the ESR at 100 kHz
- look up the datasheet and note the impedance at 100 kHz
- calculate the reactance of the cap (based on C measured at 100 kHz ) 1/(2pifC) for 2,06 uF = 0.773 ohm
- calculate Z ; in your case based on 1,2 ohm and 0.773 ohm makes: 1.427 ohm
- if Z is within the stated specs from the manufacturers datasheet the cap is good.

5th way,
- take a scope
- probe the power rails
- if there is to much ripple, replace the cap.
- how high the ripple is allowed to be is stated in the service manual of the thing you repair.
- if it is not stated you must estimate that. Not to hard, measure on the Vcc pins of some components.
You need electronic knowledge for that but you need that for repair anyhow (only swapping maybe-dead caps and random components  is not really repair)

The very bad way:
- every way that measures component in circuit 
You can do that if you want to quickly find the real dead caps in something that is so dead you can not start it. Then remove and measure them using one of the ways above. Then start the normal trouble shooting.

[BravoV]

PA4TIM, save your energy, cause Tigr is well known to bombard this forum not only this thread, with photos with his various "elite" esr meters and caps, and I'm guessing he has problem with English.

[JoeO]

PA4TIM, save your energy, cause Tigr is well known to bombard this forum not only this thread, with photos with his various "elite" esr meters and caps, and I'm guessing he has problem with English.

I agree with BravoV.  tigr is also trying to show off how much he "knows".  He is trolling people wasting their time. 

[tigr]

BravoV
Do you have any complaints about my measurement results?

[BravoV]

BravoV
Do you have any complaints about my measurement results?

Complaints what ? Its your measurements, your equipments, your caps , why should I have any problem with that ? 

My post is merely for PA4TIM as I know he is a really generous & knowledgeable guy here around with tons .. yeah, literally tons in weight of 1st tier test equipments, and seeing his posts are getting ignored again and again, just pity him. 

Also probably to The Electrician too, another really helpful, resourceful and knowledgeable guy around here "especially" in capacitor I believe.

[tigr]

No problem. All good. Good luck.

[David Hess]

The good way:
-Take a LCR meter that gives you D (=tan d).
- Look up the datasheet
- lookup D (=tan d = DF = 1/Q = related to loss angle)
- and the frequency they measured that, often 100 or 120 Hz (it was 1 kHz for a long time)
- lookup the max value of D, most times 200%
- Take a real LCR meter, set it at the stated frequency
- Measure C, is it within specs ? often +/- 20% when they are new, how low the circuit tolerates is an other question.
- A real LCR meter or bridge will give you D
- If both are within specs the cap is good. (if you want to go all the way you can measure leakage. (often stated for 2 or 3 minutes, this is not the working voltage leaking test, that is a test you do for safety.)

It is possible for a capacitor to pass all of these tests and still be bad.

I came across a z-axis amplifier in a Tektronix 7904 which could not meet its transient response specifications after calibration which had almost no effect.  The natural suspect was the Sprague 30D series 5uF 150V 8105 capacitor used for low ripple current decoupling of the output stage's bias supply but testing after removal revealed proper capacitance, low frequency dissipation, and leakage.  The capacitor was still bad however since replacing it solved the problem.  I assume it suffers from high impedance at higher frequencies which a network analyser would reveal in a comparison with a good capacitor.

[The Electrician]

The good way:
-Take a LCR meter that gives you D (=tan d).
- Look up the datasheet
- lookup D (=tan d = DF = 1/Q = related to loss angle)
- and the frequency they measured that, often 100 or 120 Hz (it was 1 kHz for a long time)
- lookup the max value of D, most times 200%
- Take a real LCR meter, set it at the stated frequency
- Measure C, is it within specs ? often +/- 20% when they are new, how low the circuit tolerates is an other question.
- A real LCR meter or bridge will give you D
- If both are within specs the cap is good. (if you want to go all the way you can measure leakage. (often stated for 2 or 3 minutes, this is not the working voltage leaking test, that is a test you do for safety.)

It is possible for a capacitor to pass all of these tests and still be bad.

I came across a z-axis amplifier in a Tektronix 7904 which could not meet its transient response specifications after calibration which had almost no effect.  The natural suspect was the Sprague 30D series 5uF 150V 8105 capacitor used for low ripple current decoupling of the output stage's bias supply but testing after removal revealed proper capacitance, low frequency dissipation, and leakage.  The capacitor was still bad however since replacing it solved the problem.  I assume it suffers from high impedance at higher frequencies which a network analyser would reveal in a comparison with a good capacitor.

Here's an example of the very sort of thing you're describing.  Without a sweep with an analyzer the difference between the defective cap and a good one wouldn't be discovered.  It's too bad you didn't get to do a sweep of your anomalous capacitor!

https://www.eevblog.com/forum/projects/capacitor-measurements-on-an-impedance-analyzer/msg178923/#msg178923

[oldway]

PA4TIM, save your energy, cause Tigr is well known to bombard this forum not only this thread, with photos with his various "elite" esr meters and caps, and I'm guessing he has problem with English.

I agree with BravoV.  tigr is also trying to show off how much he "knows".  He is trolling people wasting their time.

All his posts are about ESR....He probably also sleep with his ESR meter.... 

Let's be serious, that exaggeration is not healthy for the beginners who can believe ESR is more important than it is really.

[David Hess]

It's too bad you didn't get to do a sweep of your anomalous capacitor!

I kept the capacitor for testing when I have the proper equipment.

[marcuswilson007]

I don't understand why anyone would remove capacitors on an old piece of gear to test them, if I removed a capacitor I would not refit it, just replace it.
If one capacitor is out of spec, they should all be replaced.

[rsjsouza]

I don't understand why anyone would remove capacitors on an old piece of gear to test them, if I removed a capacitor I would not refit it, just replace it.
If one capacitor is out of spec, they should all be replaced.

Not all capacitors are bad on a board just because one or two show some fault - they may belong to entirely different circuits that impose stresses in different manners. Also, not all equipment that has some mileage on it is doomed to have all its capacitor replaced. At last, some capacitors are quite expensive, which needs to be considered on the overall economical viability of the repair. I have one example on my bench: a mid 1990s A/V receiver that has monstruous filter capacitors in excellent shape and some smaller ones that are showing signs of wear - by your logic, a $2.00 repair can easily become $50.00, which surpasses the resale value of the equipment.

[David Hess]

I don't understand why anyone would remove capacitors on an old piece of gear to test them, if I removed a capacitor I would not refit it, just replace it.
If one capacitor is out of spec, they should all be replaced.

If I unsolder a capacitor then I will likely replace it whether it tests bad or not.

For aluminum electrolytic capacitors, I may or may not replace all or specific ones depending on the circumstances.  If a decoupling capacitor tested bad, then I would change all of the decoupling capacitors.  If an output capacitor tested bad, I would replace all of the output capacitors.  If an input capacitor tested bad, then I would replace all of the input capacitors.  These three catagories tend to all wear out at the same time within their catagory.

[ipman]

Sorry for digging the topic grave, but really need to ask something related to this one.
I am trying to troubleshoot a SMPS inside a blender (based on LNK304) which is behaving like bad caps. The suspects are two 1 microfarad capacitors at 450v. One of them has an ESR of 31 and the other one 37 ohm. They are used for AC rectification (here in EU 230V).
I have nothing to compare against, if somebody has a reference I really appreciate.
P.S. I've changed places, its slightly better now, but still not working. I suspect those caps to be bad, but i have nothing similar now and obtaining new ones is a PITA for the moment.
Thank you!

[mariush]

Seems like a bit on the high side. I would expect less than 10-15 ohm

[PA4TIM]

Search the datasheet for your cap. They state D (often as tan d) Then measure the capacitance and D at 120 Hz. If those values are within limits and the cap does not leak DC he is good. ESR is often not stated but they give impedance at 100 kHz. The reactance is low at 100 kHz so you can use that value as if it was ESR.
But often they give no limit for the impedance.

Measure desoldered. 

However the best test is measuring the ripple with your scope. If that is OK the caps are good. Easiest way is replacing them with new ones that have the same specs (D, C, Vdc and max allowed current ) you need to correct the values for the current pulses in the datasheet for the operating frequency (they state the correction factor in the datasheet)

[ipman]

I cannot find the datasheet for that capacitors, as usual in cheap ones.
Of course, i can find the datasheet for ones manufactured by Panasonic and compare, but this is not 100% relevant.
I don't have a scope now unfortunatelly.
I removed them both from circuit and measured them with a Atmega based tester, which showed allmost exact capacitance value, but this is not helping everytime.
My problem is that i am now in a new country and lacking language skills makes living more dificult, imagine searching for capacitors ... that's why i am searching for help here.

[perieanuo]

Sorry for digging the topic grave, but really need to ask something related to this one.
I am trying to troubleshoot a SMPS inside a blender (based on LNK304) which is behaving like bad caps. The suspects are two 1 microfarad capacitors at 450v. One of them has an ESR of 31 and the other one 37 ohm. They are used for AC rectification (here in EU 230V).
I have nothing to compare against, if somebody has a reference I really appreciate.
P.S. I've changed places, its slightly better now, but still not working. I suspect those caps to be bad, but i have nothing similar now and obtaining new ones is a PITA for the moment.
Thank you!

Check resistors,inductance, IC.I don't think it's those capacitors in question


Envoyé de mon iPad en utilisant Tapatalk