ESR Confusion

[Noidzoid]

As I am very new to all this I do realise I am probably missing a very important point.

I decided to replace a pair of capacitors in an audio amplifier as one of them was bulging at the top and was making a loud buzzing noise. 

Upon removing the suspect one I found that it rattled when shook almost like it contained a large ball bearing.

Before soldering in the new one I thought I would do a side by side test using my newly acquired MESR-100.  To my surprise they measured EXACTLY the same.

As I stated at the beginning of this post I am probably missing something but to my untrained eye it seems the purchase of the MESR-100 was pointless.

BTW the amplifier works great now having only singled out the dodgy capacitor because of how it looked. Had I relied on the meter to make a decision I would still be looking for the fault.

I would be interested to hear comments from anyone more experienced than myself (so almost anyone really ha!)

[wraper]

ESR increases only when most of electrolyte is gone, it may bulge or not depending how tight a rubber seal is. Give it some time and bulging cap which measures fine would die and measure high ESR.

Upon removing the suspect one I found that it rattled when shook almost like it contained a large ball bearing.

Large snap in caps may have a lot of free space inside. Paper/foil roll inside usually is only attached by thin, flat, flexible leads. Depending on how much free space there was inside, and position of that roll, its not surprising it could rattle. You could take that cap apart and look yourself.

was making a loud buzzing noise.

This is something strange. Electrolytic capacitors normally don't buzz regardless if they are fine or faulty.

To my surprise they measured EXACTLY the same.

The same what? And what was the capacitance and voltage rating? Maybe they were equally dead. EDIT: for some reason I thought you compared two old ones, I guess you mean old and new.

[Noidzoid]

Yes, old vs new
6800uF 50v

Big buzzing that was loud with no volume could be amplified through the speakers. Could the capacitors I replaced have been causing something else to buzz? Although I was pretty sure the noise was  coming from the bulging one as when I put my finger on the top of it I could feel it vibrating.

[wraper]

Maybe there was lost/broken connection between actual cap and terminals inside the case, hence it buzzed and rattled. Or broken solder joint (if it's single layer PCB). And maybe connection was restored when you removed or shock it. Anyway, it's something strange. If you still have it, you could peek inside out of interest.

[tautech]

SMPS PSU's are more prone to buzzing if componentry has drifted out of spec but that's not to say linear PSU's won't buzz either. Being their caps are operating at vastly different frequencies, SMPS are more likely to whine in the KHz region whereas linear a hum @ 2x mains frequency.
I'm guessing being old enough for the main filter caps to be shot it's a linear PSU and they usually don't use low ESR caps. The ESR value selected by designers has much to do with the application and this needs be considered when selecting replacements.

[wraper]

6800uF 50v

Are you sure it was actually bulged? Such capacitors usually have a thin black plastic cover over the vent on the top. Often this part is not flat even if there is nothing wrong with capacitor.

[Noidzoid]

Unless I'm misunderstanding the meaning of 'vent' it must have been blocked.

When I pierced the canister there was a hiss like air escaping.

Should there have been fluid inside? There was none.

This is a picture of the capacitor that didn't bulge at the top but I replaced it as they looked like a pair

[wraper]

When I pierced the canister there was a hiss like air escaping.

There was a gas under pressure inside. Such caps usually still retain low ESR. There is no fluid that could flow inside. There is a roll wound of paper and aluminium foil soaked in just enough of electrolyte. You could see fluid coming out only if squeeze this roll.

[Noidzoid]

When I pierced the canister there was a hiss like air escaping.

There was a gas under pressure inside. Such caps usually still retain low ESR. There is no fluid that could flow inside. There is a roll wound of paper and aluminium foil soaked in just enough of electrolyte. You could see fluid coming out only if squeeze this roll.

So it's possible for a faulty cap to have the same low ESR as a new healthy one?

Does this mean my MESR-100 is a pointless gimick/tool?

[wraper]

When I pierced the canister there was a hiss like air escaping.

There was a gas under pressure inside. Such caps usually still retain low ESR. There is no fluid that could flow inside. There is a roll wound of paper and aluminium foil soaked in just enough of electrolyte. You could see fluid coming out only if squeeze this roll.

So it's possible for a faulty cap to have the same low ESR as a new healthy one?

Does this mean my MESR-100 is a pointless gimick/tool?

It's not pointless. It just means that at the moment you measured the cap it was performing just like a good capacitor. When capacitor measures well, it means it measures well at this moment, it could die in a few months. That's why people recap some monitor PSU, they usually change most of the caps even if most among them still measure fine and only a few failed. As I said, probably the hum problem wasn't because of the capacitor itself, maybe it was a bad soldering or bad contact which did go away when you removed a PCB. ESR meter is very handy when you have a PCB with a lot of caps which you need to check. Also often you repair a device, fix it by replacing something except capacitors. Device performs well or almost well, you check ESR and may find some which are bad. Without ESR meter, you'd never know.

[Noidzoid]

So a capacitor showing low ESR could be faulty, but a capacitor showing high ESR is definitely faulty. Correct?

[wraper]

So a capacitor showing low ESR could be faulty, but a capacitor showing high ESR is definitely faulty. Correct?

Yep. Although visually faulty does not mean necessarily electrically faulty (unless leaks current or is completely shorted). It could be electrically fine at the moment (hence measures fine), building pressure inside (bulging) and will electrically fail in the near future.

[Noidzoid]

Ok. I won't be throwing my MESR-100 out then.

Thanks for taking the time to explain.

[Doctorandus_P]

Some audio capacitors have a plastic disc on the top and that disk may bulge from mechanical stress (it is compressed by the crimping foil around the cap) but that does not mean the capacitor itself is faulty.

I one cut a hole in such a disk and the aluminimum top of the Elco was perfectly flat and it also measured OK on an ESR meter (which I bought later).

[james_s]

Like any tool, ESR meters have limitations. They are not magic, they still rely on you knowing how to interpret what they are telling you but they are useful tools. Increased ESR is not the only symptom of a failing capacitor but it is one common symptom.

[bitseeker]

Capacitors have many properties and there are many tests that can be used to determine their well being. ESR is but one type of test to measure one property. Others may include, but are not limited to, capacitance, leakage, and Q factor. Take a peek at the data sheet for several types of capacitors and you'll see what I mean.

So, just because two capacitors have the same ESR doesn't necessarily mean they're equally good or bad. There may be other problems that aren't detected by an ESR measurement. However, if you didn't have an ESR meter, then you'd have even less information available to make a determination.

[Terry01]

You'll find the ESR meter you have is better suited for measuring smaller smd caps, sort of. It measures @ 100kHz in series which is way too high for a 6500uf electrolytic. Also @ 100 kHz paralel measurment is better than series which is why i say "sort of".  The cap you measured should be tested closer to @ 120 Hz in series or there about. Have a read at this and it'll give you a better idea of what I mean. Get your head around the graph and you'll understand how i mean.

 www.techni-tool.com/site/ARTICLE_LIBRARY/BK%20Precision%20-%20How%20to%20Use%20an%20LCR%20Meter.pdf

[wraper]

You'll find the ESR meter you have is better suited for measuring smaller smd caps, sort of. It measures @ 100kHz in series which is way too high for a 6500uf electrolytic. Also @ 100 kHz paralel measurment is better than series which is why i say "sort of".  The cap you measured should be tested closer to @ 120 Hz in series or there about. Have a read at this and it'll give you a better idea of what I mean. Get your head around the graph and you'll understand how i mean.

 www.techni-tool.com/site/ARTICLE_LIBRARY/BK%20Precision%20-%20How%20to%20Use%20an%20LCR%20Meter.pdf

There is nothing wrong with measuring ESR at 100kHz for large capacitors unless your device measures capacitance at the same time and you need capacitance measurements as well.

Get your head around the graph and you'll understand how i mean.

That graph is not about ESR.

[Terry01]

You'll find the ESR meter you have is better suited for measuring smaller smd caps, sort of. It measures @ 100kHz in series which is way too high for a 6500uf electrolytic. Also @ 100 kHz paralel measurment is better than series which is why i say "sort of".  The cap you measured should be tested closer to @ 120 Hz in series or there about. Have a read at this and it'll give you a better idea of what I mean. Get your head around the graph and you'll understand how i mean.

 www.techni-tool.com/site/ARTICLE_LIBRARY/BK%20Precision%20-%20How%20to%20Use%20an%20LCR%20Meter.pdf

There is nothing wrong with measuring ESR at 100kHz for large capacitors unless your device measures capacitance at the same time and you need capacitance measurements as well.

Get your head around the graph and you'll understand how i mean.

That graph is not about ESR.

When would a large electrolytic be used @ 100kHz? What would happen if it was? I have a good machine that measures @ both 100 or 100k so which is correct?

If that graph is not for ESR did they just make it for a laugh or does BK not know nothing?

[wraper]

If that graph is not for ESR did they just make it for a laugh or does BK not know nothing?

BK is fine, something wrong on your side. READ again, LCR measurement guidelines, not ESR. LOW ESR capacitor ESR usually is specified at 100kHz, be it small or big capacitor. General purpose capacitor ESR may be specified either at 120/100Hz or 100kHz. In both cases capacitance is usually specified at 120Hz or 100Hz. In practice there is no need to check ESR at 120Hz when checking components, unless you want compare apples vs apples with figure in a datasheet.
Here you go, up to 18000uF with ESR specified at 100kHz. http://www.chemi-con.com/upload/files/5/1/74811667552d6c4d41a84c.pdf

When would a large electrolytic be used @ 100kHz?

Say SMPS.

If your LCR meter allows it, measure the impedance
at the application frequency if you can. Seeing the
impedance magnitude in ohms helps you use your
intuition—and the phase angle tells you quickly if
you have a pure reactance or a mixture of resistance
and reactance.

[BravoV]

When would a large electrolytic be used @ 100kHz? What would happen if it was? I have a good machine that measures @ both 100 or 100k so which is correct?

Most low ESR cap datasheets stated that the capacitance was measured at the factory at 120 Hz, while the ESR (most datasheet called "Impedance") was measured at 100 Kilo Hertz.

Some well known solid polymer also specified at 100 KHz "up to" 300 KHz, although not the common practice.

So the correct frequency is measure them "according" to the manufacturer's specification, just check the datasheet.

[Terry01]

If that graph is not for ESR did they just make it for a laugh or does BK not know nothing?

BK is fine, something wrong on your side. READ again, LCR measurement guidelines, not ESR. LOW ESR capacitor ESR usually is specified at 100kHz, be it small or big capacitor. General purpose capacitor ESR may be specified either at 120/100Hz or 100kHz. In both cases capacitance is usually specified at 120Hz or 100Hz. In practice there is no need to check ESR at 120Hz when checking components, unless you want compare apples vs apples with figure in a datasheet.
Here you go, up to 18000uF with ESR specified at 100kHz. http://www.chemi-con.com/upload/files/5/1/74811667552d6c4d41a84c.pdf

When would a large electrolytic be used @ 100kHz?

Say SMPS.

If your LCR meter allows it, measure the impedance
at the application frequency if you can. Seeing the
impedance magnitude in ohms helps you use your
intuition—and the phase angle tells you quickly if
you have a pure reactance or a mixture of resistance
and reactance.

Nope! I am fine, no problems here. English is my 1st language so no help needed thanks.

Just for fun I did a couple caps @ 100 Hz, 120 Hz, 1 kHz, 10 kHz and 100 kHz and just between just 120 Hz and 1 kHz the ESR measurement is half or more "less" at the higher frequency. Before you say meter is wrong or some other waffle I did a few.

A big electrolytic may well/defo is! used in a SMPS, but it won't be working @ 100 kHz or the rest of the circuit wouldn't like it very much. Not at all in fact. Nice try but come on. Your getting deeper.

That's all I won't reply no more.

[wraper]

Just for fun I did a couple caps @ 100 Hz, 120 Hz, 1 kHz, 10 kHz and 100 kHz and just between just 120 Hz and 1 kHz the ESR measurement is half or more "less" at the higher frequency. Before you say meter is wrong or some other waffle I did a few.

There is nothing wrong with measurements or your meter. Nobody said you'll get exactly the same figures. You just confirmed that measuring at 100kHz is completely valid to estimate condition of the capacitor. Read again:

In practice there is no need to check ESR at 120Hz when checking components, unless you want compare apples vs apples with figure in a datasheet.

The issue with measuring ESR at low frequencies it that you need a full blown LCR meter, as capacitance plays a big role in measurement. When measuring at 100kHz you can basically ignore the capacitance and assume that total impedance = ESR.

[wraper]

A big electrolytic may well/defo is! used in a SMPS, but it won't be working @ 100 kHz or the rest of the circuit wouldn't like it very much. Not at all in fact. Nice try but come on. Your getting deeper.

Big electrolytic capacitors are used in SMPS not for big capacitance which drops at high frequencies but for large ripple current they can handle. That's why they usually can be replaced with smaller value polymer capacitors which have much higher ripple current rating at the same capacitance.
As you can see from the graphs, electrolytic caps are completely usable at 100 kHz.

[Gyro]

If that graph is not for ESR did they just make it for a laugh or does BK not know nothing?

BK is fine, something wrong on your side. READ again, LCR measurement guidelines, not ESR. LOW ESR capacitor ESR usually is specified at 100kHz, be it small or big capacitor. General purpose capacitor ESR may be specified either at 120/100Hz or 100kHz. In both cases capacitance is usually specified at 120Hz or 100Hz. In practice there is no need to check ESR at 120Hz when checking components, unless you want compare apples vs apples with figure in a datasheet.
Here you go, up to 18000uF with ESR specified at 100kHz. http://www.chemi-con.com/upload/files/5/1/74811667552d6c4d41a84c.pdf

When would a large electrolytic be used @ 100kHz?

Say SMPS.

If your LCR meter allows it, measure the impedance
at the application frequency if you can. Seeing the
impedance magnitude in ohms helps you use your
intuition—and the phase angle tells you quickly if
you have a pure reactance or a mixture of resistance
and reactance.

Nope! I am fine, no problems here. English is my 1st language so no help needed thanks.

Just for fun I did a couple caps @ 100 Hz, 120 Hz, 1 kHz, 10 kHz and 100 kHz and just between just 120 Hz and 1 kHz the ESR measurement is half or more "less" at the higher frequency. Before you say meter is wrong or some other waffle I did a few.

A big electrolytic may well/defo is! used in a SMPS, but it won't be working @ 100 kHz or the rest of the circuit wouldn't like it very much. Not at all in fact. Nice try but come on. Your getting deeper.

That's all I won't reply no more.

(@Terry01) Just for the avoidance of any confusion. The only way to confirm the health of a capacitor (or any other component come to that) is to measure it against the specification set out in its datasheet. If the datasheet specifies measurement at 100 or 120Hz then you should confirm the ESR at that frequency, however if the datasheet specifies a measurement 100kHz then the value needs to be measured at that frequency. Electrolytic capacitors are manufactured in many different families, some 'higher' ESR ones are intended for simple mains transformer/rectifier applications. Low ESR one are designed for high frequency SMPS uses, where the other ones would be ineffective and would quickly overheat.

If you adhere to the datasheet figures specified for that specific part then there can be no confusion and no dispute.

[wraper]

Low ESR one are designed for high frequency SMPS uses, where the other ones would be ineffective and would quickly overheat.

It's less about frequency and more about ripple current. General purpose capacitors can work at the same frequencies completely fine, it's just they won't survive high ripple current for long. Say if you place several general purpose caps in parallel instead of one LOW ESR, they'd work just fine. A few decades ago, there were SMPS but nobody heard about such thing as LOW ESR capacitors.

[Terry01]

If that graph is not for ESR did they just make it for a laugh or does BK not know nothing?

BK is fine, something wrong on your side. READ again, LCR measurement guidelines, not ESR. LOW ESR capacitor ESR usually is specified at 100kHz, be it small or big capacitor. General purpose capacitor ESR may be specified either at 120/100Hz or 100kHz. In both cases capacitance is usually specified at 120Hz or 100Hz. In practice there is no need to check ESR at 120Hz when checking components, unless you want compare apples vs apples with figure in a datasheet.
Here you go, up to 18000uF with ESR specified at 100kHz. http://www.chemi-con.com/upload/files/5/1/74811667552d6c4d41a84c.pdf

When would a large electrolytic be used @ 100kHz?

Say SMPS.

If your LCR meter allows it, measure the impedance
at the application frequency if you can. Seeing the
impedance magnitude in ohms helps you use your
intuition—and the phase angle tells you quickly if
you have a pure reactance or a mixture of resistance
and reactance.

Nope! I am fine, no problems here. English is my 1st language so no help needed thanks.

Just for fun I did a couple caps @ 100 Hz, 120 Hz, 1 kHz, 10 kHz and 100 kHz and just between just 120 Hz and 1 kHz the ESR measurement is half or more "less" at the higher frequency. Before you say meter is wrong or some other waffle I did a few.

A big electrolytic may well/defo is! used in a SMPS, but it won't be working @ 100 kHz or the rest of the circuit wouldn't like it very much. Not at all in fact. Nice try but come on. Your getting deeper.

That's all I won't reply no more.

(@Terry01) Just for the avoidance of any confusion. The only way to confirm the health of a capacitor (or any other component come to that) is to measure it against the specification set out in its datasheet. If the datasheet specifies measurement at 100 or 120Hz then you should confirm the ESR at that frequency, however if the datasheet specifies a measurement 100kHz then the value needs to be measured at that frequency. Electrolytic capacitors are manufactured in many different families, some 'higher' ESR ones are intended for simple mains transformer/rectifier applications. Low ESR one are designed for high frequency SMPS uses, where the other ones would be ineffective and would quickly overheat.

If you adhere to the datasheet figures specified for that specific part then there can be no confusion and no dispute.

I agree 100%.

The data sheet is the way to go..... but I doubt very much BK thought we'll make a graph to help users get the best results from our instruments but we'll feed them a load of shite just to confuse them in the instructions. I get it's more a "rule of thumb" thing than "it applies to EVERY capacitor".

What happens if for some reason you can't get the data sheet for the part you need to test? You don't take a measure because you can't use your head and think. You must stick to the data sheet 1000%! Come on??

There's no confusion here.

[Gyro]

While that "rule of thumb" might have some relevance to signal level components, BK themselves clarify this in the text:

Measurement frequency: Since the reactance is a function of frequency, your choice of measurement frequency should reflect the usage of the component. For capacitors, larger values (tens to hundreds of μF or more) are often used in power supply filtering applications, so should be tested at twice the line frequency. Smaller capacitors (fractions of a µF) tend to be used at higher frequencies, so should be tested at 1kHz or more.

In the case of SMPSs the 'line', ie. switching frequency is normally 50kHz or more, so their words are completely in line with 100kHz testing of these higher capacitance parts. You can't simply rely on that straight line graph without paying attention to the accompanying text.


The trouble is, these days, there are several "rules of thumb". A capacitor with, say, 0.5R may be perfectly fine for a gentle rectified mains transformer based supply, it would be totally ineffective (Faulty) in an SMPS application with high frequency, high peak currents.

Sadly there is no non-application-specific simple single rule of thumb.

[wraper]

BTW if you think that in linear power supply capacitors work only at double of the mains frequency, you are wrong. They work at frequency which is consumed by the load. Say if it is an audio amplifier, expect up to 20 kHz.

[Terry01]

While that "rule of thumb" might have some relevance to signal level components, BK themselves clarify this in the text:

Measurement frequency: Since the reactance is a function of frequency, your choice of measurement frequency should reflect the usage of the component. For capacitors, larger values (tens to hundreds of μF or more) are often used in power supply filtering applications, so should be tested at twice the line frequency. Smaller capacitors (fractions of a µF) tend to be used at higher frequencies, so should be tested at 1kHz or more.

In the case of SMPSs the 'line', ie. switching frequency is normally 50kHz or more, so their words are completely in line with 100kHz testing of these higher capacitance parts. You can't simply rely on that straight line graph without paying attention to the accompanying text.


The trouble is, these days, there are several "rules of thumb". A capacitor with, say, 0.5R may be perfectly fine for a gentle rectified mains transformer based supply, it would be totally ineffective (Faulty) in an SMPS application with high frequency, high peak currents.

Sadly there is no non-application-specific simple single rule of thumb.

Eh?? 

Are we looking at the same thing?

You've got it all the paddy way round about neeb! I think you need a pair of spectacles!

[Gyro]

  I have no idea what you're talking about, what I said seems perfectly clear and the quote is from the BK document you linked.