One last ESR question

[Ivan Vegvary]

Okay, I know that electrolytics are the most likely to fail. However why do the manufacturers of ESR meters discuss ONLY electrolytics? do ceramics etc. not have an "equivalent series resistance"? Can they also be checked?
Thanks for putting up with me and your previous answers.

Ivan Vegvary

 

[M Philbrook]

In article <c21126e3-1005-4dd8-80e2-9fbf63edb6d9@googlegroups.com>,
ivanvegvary@gmail.com says...

Okay, I know that electrolytics are the most likely to fail. However why do the manufacturers of ESR meters discuss ONLY electrolytics? do ceramics etc. not have an "equivalent series resistance"? Can they also be checked?
Thanks for putting up with me and your previous answers.

Ivan Vegvary

You can check any of them if you wish but keep in mind
that the value of the cap to start with has to be up to
minimum size before the meter can even get a valid reading.

If you pick a meter that has a high base freq you'll do beter.

We have one that has a 100khz test scale and can be used for
low value caps. We use that scale mostly when checking poly type
caps.

Electro's are the most noted problematic type caps when it
comes to ESR issues, but poly wraps can also give you some fits
and the types that have crimped terminations internally.

Jamie

 

[Jeff Liebermann]

On Wed, 22 Apr 2015 19:03:41 -0700 (PDT), Ivan Vegvary
<ivanvegvary@gmail.com> wrote:

Okay, I know that electrolytics are the most likely to fail.
However why do the manufacturers of ESR meters discuss ONLY
electrolytics? do ceramics etc. not have an "equivalent series
resistance"? Can they also be checked?

You can check any type of capacitor. However, electrolytics are one
the few types of capacitors that have a liquid electrolyte, which is
subject to decomposition, contamination, leaking, outgassing, and
corrosion. The other liquid dielectric is the wet slug tantalum,
which is considered a health and fire hazard. Solid dielectric caps,
such as ceramic, film, mica, etc don't have these problems.
Unfortunately, they have other problems. Film caps fall apart
internally. MLCC (multilayer ceramic caps) have a very thin brittle
dielectric which easily cracks when stressed or thermal shocked. Some
film caps are hygroscopic and change value when the case is cracked.
Most of these failures are catastrophic and are easily seen with a
capacitance meter. Only the electrolytics are able to change ESR
without a corresponding change in capacitance. Therefore, the ESR
meter targets electrolytic capacitor failures, while a capacitance
meter works well enough for the other capacitor types.

Note that an ESR meter is also used for battery testing, but those
tend to be rather specialized. For example, capacitor ESR meters run
at 100KHz, while battery testers run at 1KHz.
<http://batteryuniversity.com/learn/article/how_to_measure_internal_resistance>


--
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

 

[N_Cook]

On 23/04/2015 03:03, Ivan Vegvary wrote:

Okay, I know that electrolytics are the most likely to fail. However why do the manufacturers of ESR meters discuss ONLY electrolytics? do ceramics etc. not have an "equivalent series resistance"? Can they also be checked?
Thanks for putting up with me and your previous answers.

Ivan Vegvary

The common failure mode of non-electrolytics eg tantalum and MLCC is
they go ohmic , so just DVM-R test picks them up. Tantalum because of
microscopic "plate" separation and shorting across, then MLCC due to
metal migration into microscopic cracks , usually created at soldering ,
especially PbF temps. Incidently for electrolytics , you should check
for capacity with a capacitance meter also, if ESR checks out ok

 

[Phil Allison]

Ivan Vegvary wrote:

Okay, I know that electrolytics are the most likely to fail.
However why do the manufacturers of ESR meters discuss ONLY
electrolytics?

** Cos they are the only kind of caps the meters they sell are useful for.

do ceramics etc. not have an "equivalent series resistance"?
Can they also be checked?

** It never varies, so there is no point.

Look it up on the maker's spec sheets if you need to.



..... Phil

 

[Phil Allison]

N_Cook wrote:

Ivan Vegvary wrote:

Okay, I know that electrolytics are the most likely to fail. However why do the manufacturers of ESR meters discuss ONLY electrolytics? do ceramics etc. not have an "equivalent series resistance"? Can they also be checked?
Thanks for putting up with me and your previous answers.


The common failure mode of non-electrolytics eg tantalum and MLCC is
they go ohmic , so just DVM-R test picks them up. Tantalum because of
microscopic "plate" separation and shorting across, then MLCC due to
metal migration into microscopic cracks , usually created at soldering ,
especially PbF temps.

** A DMM ohmage test is no use unless the cap is removed from circuit.

So the way you find the bad one is with a bench PSU good for a few amps.

The smoking cap or caps are then the bad ones.

Incidently for electrolytics , you should check
for capacity with a capacitance meter also, if ESR checks out ok

** Absolute bollocks !!!!!!!!!!!!!!!!!

If the ESR reading is good then the uFs are ALL there.




..... Phil

 

[John-Del]

On Thursday, April 23, 2015 at 4:08:09 AM UTC-4, Phil Allison wrote:

** Absolute bollocks !!!!!!!!!!!!!!!!!

If the ESR reading is good then the uFs are ALL there.




.... Phil

Bullshit.... Just because *you* haven't see one doesn't make it true. While a tiny percentage, I have seen several electros that check well within the realm of normal ESR and are decidedly well down in value.

 

[RobertMacy]

On Wed, 22 Apr 2015 21:56:14 -0700, Jeff Liebermann <jeffl@cruzio.com>
wrote:

..snip....
You can check any type of capacitor. However, electrolytics are one
the few types of capacitors that have a liquid electrolyte, which is
subject to decomposition, contamination, leaking, outgassing, and
corrosion. The other liquid dielectric is the wet slug tantalum,
which is considered a health and fire hazard. Solid dielectric caps,
such as ceramic, film, mica, etc don't have these problems.
Unfortunately, they have other problems. Film caps fall apart
internally. MLCC (multilayer ceramic caps) have a very thin brittle
dielectric which easily cracks when stressed or thermal shocked. Some
film caps are hygroscopic and change value when the case is cracked.
Most of these failures are catastrophic and are easily seen with a
capacitance meter. Only the electrolytics are able to change ESR
without a corresponding change in capacitance. Therefore, the ESR
meter targets electrolytic capacitor failures, while a capacitance
meter works well enough for the other capacitor types.

Note that an ESR meter is also used for battery testing, but those
tend to be rather specialized. For example, capacitor ESR meters run
at 100KHz, while battery testers run at 1KHz.
http://batteryuniversity.com/learn/article/how_to_measure_internal_resistance

Or, you can use your soundcard and do any spectrum from 10Hz to around
91kHz, and actually get plots of Re(Z) amd Im(Z) vs Freq. I used my card
nad found a 'tracking' relationship between reactive impedance and the
loss in an aluminum cap! My whole career, that was something I never knew
existed [paid attention to]! But, found verified in an AppNote from AVX.
Very educational to 'play' with such an inexpensive instrument.

 

[Jeff Liebermann]

On Thu, 23 Apr 2015 05:35:06 -0700, RobertMacy
<robert.a.macy@gmail.com> wrote:

Or, you can use your soundcard and do any spectrum from 10Hz to around
91kHz, and actually get plots of Re(Z) amd Im(Z) vs Freq. I used my card
nad found a 'tracking' relationship between reactive impedance and the
loss in an aluminum cap! My whole career, that was something I never knew
existed [paid attention to]! But, found verified in an AppNote from AVX.
Very educational to 'play' with such an inexpensive instrument.

ESR is normally measured at 100KHz. A 192KHz sound card won't do
100KHz.

I'm more into the RF chacteristics of passive components which
requires a VNA (vector network analyzer). For example:

"Measuring Capacitor Parameters Using Vector Network Analyzers"
<http://electronics.etfbl.net/journal/Vol18No1/xPaper_05.pdf>
Notice on the first page that the authors redefine ESR as the residual
resistance at resonance. Sigh.

Oops. gotta run...

--
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

 

[RobertMacy]

On Thu, 23 Apr 2015 08:21:00 -0700, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Thu, 23 Apr 2015 05:35:06 -0700, RobertMacy
robert.a.macy@gmail.com> wrote:

Or, you can use your soundcard and do any spectrum from 10Hz to around
91kHz, and actually get plots of Re(Z) amd Im(Z) vs Freq. I used my card
nad found a 'tracking' relationship between reactive impedance and the
loss in an aluminum cap! My whole career, that was something I never
knew
existed [paid attention to]! But, found verified in an AppNote from AVX.
Very educational to 'play' with such an inexpensive instrument.

ESR is normally measured at 100KHz. A 192KHz sound card won't do
100KHz.

I'm more into the RF chacteristics of passive components which
requires a VNA (vector network analyzer). For example:

"Measuring Capacitor Parameters Using Vector Network Analyzers"
http://electronics.etfbl.net/journal/Vol18No1/xPaper_05.pdf
Notice on the first page that the authors redefine ESR as the residual
resistance at resonance. Sigh.

Oops. gotta run...

a Soundcard that can do 192kS/s, like Creative Lab's old EMU1212, will
give you over 120 dB dynamic range to approx 89kHz. That ratio is 89% of
100kHz and as people have noticed there's not a lot of difference being
slightly off in frequency by small margin. How much 'error' between 100kHz
and 89kHz would one expect when measuring esr? Well, unless here's a high
Q resonance in there.

 

[Michael Black]

On Wed, 22 Apr 2015, Ivan Vegvary wrote:

Okay, I know that electrolytics are the most likely to fail. However
why do the manufacturers of ESR meters discuss ONLY electrolytics? do
ceramics etc. not have an "equivalent series resistance"? Can they also
be checked? Thanks for putting up with me and your previous answers.

You do realize that the only reason anyone uses electrolytics is because
that's the only way to get larger values of capacitance into a small
package? That's the only value, and if something better came along,
electrolytics would be dropped. Likewise, everytime an electrolytic is
specified, it's not because a "polarized" capacitor is needed, that's
fallout from the way electrolytics work. You can't get non-polarized
electrolutics (well not as one might want), so you are stuck. But no
circuitry demands polarized capacitors.

That said, the fuss over ESR and electrolytics is because of the way
electrolytics are made (to get that large capacitance in a small package).
They aren't as good at handling higher frequencies, and because they use
that electrolyte (which can be vented if things get too hot), their
ability to handle those higher frequencies can decay over time.

In the old days, you'd see electrolytics in power supplies, and bypassing
the the cathodes of audio amplifiers. So they were only having to deal
with 60Hz (in the power supply) and audio frequencies in the audio
amplifiers, relatively low frequencies. If the electrolytic went bad in
the power supply, you'd hear hum in the radio (or see hum manifested some
other way), you'd notice the decline of the capacitor because you could
hear the 60Hz coming from the wall, and when it wasn't filtered, you'd
hear it. The cathode bypasses, if the electrolytics declined there, (but
they weren't having to deal with much power), you'd lose audio gain.

When solid state devices came along, there was a big rise in the use of
electrolytics. Tubes are high voltage, low current devices, so relatively
small value electrolytics were used. Transistors and ICs were low
voltage, higher current devices, so the values went up, putting more
demand on the electrolytics (and causing new methods to come along; large
value electrolytics are now much smaller than forty and fifty years ago).
Things might be a bit more complicated if an electrolytic went bad, you
might get motorboating (feedback) when a supply line wasn't as well
filtered, but still the electrolytics were doing the same thing, at 60Hz
or audio frequencies.

Then along came switching supplies. Lots of electrolytics running at
significantly higher frequencies. Since they were large value capacitors,
they were expected to handle quite a bit. But electrolytics can have
large capacitance and not be so good at higher frequencies, by the nature
of their manufacture. Forty years ago, National would tell us to bypass
three terminal regulators with electrolytics and .1uF ceramics, the
ceramics to better bypass the higher frequencies.

Switching supplies meant electrolytics had to do more, and the higher
frequencies they were dealing with put more strain on that.

That's the rise of ESR meters, suddenly electrolytics were plentiful,
being stretched to their limits, and in circuits where a bad capacitor
might not be so obvious as in the days of tubes, when you'd hear the hum
if the electrolytic was bad.

ESR meters rose up because suddenly this was a problem, but also it was
easier testing the electrolytics (especially in circuit) than trying to
figure out those newfangled switching supplies with all their feedback and
the like.

But other capacitors are a different matter. They are relatively small
value capacitors, far better at handling the higher frequencies. There
were paper capacitors which caused trouble, but those haven't seen new use
in decades (they had large inductance because of the way they were made,
and also they could go bad for their own reasons). A ceramic capacitor is
about the best higher frequency bypass capacitor you can get, though
packaging matters since the higher in frequency you go, the more the leads
count towards inductance. But ceramic capacitors don't go bad.

It's the same with other types of capacitors. They don't have electrolyte
to dry up, and since the small values are already better to handle higher
frequencies, that's not a problem.

ESR by definition is about testing electrolytics at higher frequencies.
SInce small value capacitors are better able to handle those, ESR doesn't
come into play.

This is the problem with that "Bad Caps" forum, a bit of knowledge makes
the members think they know a lot, and then they try to extend that bit of
knowledge to other capacitors, where it doesn't apply.

Some older capacitors (like from the fifties) need replacing, but that's a
different thing.

Michael

 

Topresponding to get above all the BS, all capacitors have ESR. All capacitors have ESL as well as leakage. It is a matter of what is important and what is not. to understand this you need to know what these capacitors are for.

There are small caps used to tune RF circuits and all that, and they usually have a very low leakage and ESL. The ESR is not al that critical in that type of circuit. It is important, but there are aot mor eimportant things like ESL and the accuracy of the capacitance, in percent. Not like electrolytics where anytihng above is good until it hits ridiculous.

Another type of cap is used in signal. it is sometimes like the RF cap, like in tone control circuits fro example, they set the frequency. Or any frequency discrininating network. The capacitance is many times more critical than ESR. However ESL can still be a problem.

Electrolytics are used a different way. First of all, their value is almost never critical. Seruiously, in some cases you can put a 100 uF where a 1,000 uF was and it will work fine. Actually tis is especially true of olders lytics.

Speaking of which, this ESR comes on but all the cpacitance is still there only aplies to newer ones. The older ones actually would lose capacitance. In fact I got one at workk right now. Filter cap in an old tube amp. It has a perfect sawtooth across it, absolutely none of that characteristic hump from the rectifier conduction. It has lost capacitance. Old paper job.

Like with replacing caps, electrolytics ? The value almost does not matter. Almost. It is all switched mode now, anyting over 0.01 uF would work.

If it canhandle the ripple current.

now there are some thihngs where they want the sources to decay in an orderly fashion at power down or power removal. Well OK, but in those applications they should be using better grade caps.

I'm sure they are.

LOLOLOLOLOLOLOL

 

I'll have to write this some other time............

 

[Phil Allison]

John-Del wrote:

Phil Allison wrote:


** Absolute bollocks !!!!!!!!!!!!!!!!!

If the ESR reading is good then the uFs are ALL there.



Bullshit....

** Loss of capacitance and high ESR derive from the same cause - the gradual loss of electrolyte. But ESR rises fist and capacitance much later.

If ESR is good, then there is plenty of electrolyte and so C must be good also.


> Just because *you* haven't see one doesn't make it true.

** Very few people have ever seen an example that geos the other way, bout the same number who have seen real Martians.

My advice is solid, only a raving nut case would REMOVE every electro from circuit in order to test the C value when the * in circuit* ESR reading was good.

Do you ?



.... Phil

 

[Jeff Liebermann]

On Thu, 23 Apr 2015 10:46:27 -0700, RobertMacy
<robert.a.macy@gmail.com> wrote:

On Thu, 23 Apr 2015 08:21:00 -0700, Jeff Liebermann <jeffl@cruzio.com
wrote:

On Thu, 23 Apr 2015 05:35:06 -0700, RobertMacy
robert.a.macy@gmail.com> wrote:

Or, you can use your soundcard and do any spectrum from 10Hz to around
91kHz, and actually get plots of Re(Z) amd Im(Z) vs Freq. I used my card
nad found a 'tracking' relationship between reactive impedance and the
loss in an aluminum cap! My whole career, that was something I never
knew
existed [paid attention to]! But, found verified in an AppNote from AVX.
Very educational to 'play' with such an inexpensive instrument.

ESR is normally measured at 100KHz. A 192KHz sound card won't do
100KHz.

I'm more into the RF chacteristics of passive components which
requires a VNA (vector network analyzer). For example:

"Measuring Capacitor Parameters Using Vector Network Analyzers"
http://electronics.etfbl.net/journal/Vol18No1/xPaper_05.pdf
Notice on the first page that the authors redefine ESR as the residual
resistance at resonance. Sigh.

Oops. gotta run...

a Soundcard that can do 192kS/s, like Creative Lab's old EMU1212, will
give you over 120 dB dynamic range to approx 89kHz.

Most of the 192 KHz cards will digitize to 24 bits. That's a
theoretical 144 dB:
<http://en.wikipedia.org/wiki/Audio_bit_depth#Quantization>

These daze, the next big thing are 32 bit 384 KHz sound cards:
<https://hifiduino.wordpress.com/2011/12/28/musiland-dragon/>
<http://www.antelopeaudio.com/en/products/Zodiac-Gold-DAC>
<http://www.speeddragon.com/index.php?controller=Default&action=ProductInfo&Id=572>
<http://www.amazon.com/dp/B00A2QL1CI/?tag=fox003-20>
These are audiophile grade toys, so please have your loan advisor
available before pricing.

That ratio is 89% of
100kHz and as people have noticed there's not a lot of difference being
slightly off in frequency by small margin. How much 'error' between 100kHz
and 89kHz would one expect when measuring esr? Well, unless here's a high
Q resonance in there.

Good point. Close enough. However, for a few centi-bucks extra, a
384 KHz sound card will probably work up to about 150 KHz.


--
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

 

[mike]

On 4/22/2015 11:23 PM, N_Cook wrote:

On 23/04/2015 03:03, Ivan Vegvary wrote:
Okay, I know that electrolytics are the most likely to fail. However
why do the manufacturers of ESR meters discuss ONLY electrolytics? do
ceramics etc. not have an "equivalent series resistance"? Can they
also be checked?
Thanks for putting up with me and your previous answers.

Ivan Vegvary


The common failure mode of non-electrolytics eg tantalum and MLCC is
they go ohmic , so just DVM-R test picks them up. Tantalum because of
microscopic "plate" separation and shorting across, then MLCC due to
metal migration into microscopic cracks , usually created at soldering ,
especially PbF temps. Incidently for electrolytics , you should check
for capacity with a capacitance meter also, if ESR checks out ok

I'm a time-domain kinda guy.
If you stick a current pulse into a cap and look at the voltage
on a scope, the step gives you a measure of ESR and the exponential
gives you a measure of the capacitance.

If testing in a circuit, you can often find which cap is bad in
a cap/inductor/another cap circuit.

 

[John-Del]

On Thursday, April 23, 2015 at 11:05:14 PM UTC-4, Phil Allison wrote:

** Very few people have ever seen an example that geos the other way, bout the same number who have seen real Martians.

The next time I get one, I'll mail it to you.

My advice is solid, only a raving nut case would REMOVE every electro from circuit in order to test the C value when the * in circuit* ESR reading was good.

Do you ?

I don't really use the ESR meter much anymore, just for morbid curiosity after the fact really. Scopes are much faster. Plus, there are circuits that will fool ESR meters into showing perfectly good caps when they are actually close to being exhausted. Anybody who's ever rebuilt a Mitsubishi format or DM board with the half dozen dc-dc converters on board has seen this.

 

[JW]

On Thu, 23 Apr 2015 14:13:03 -0400 Michael Black <et472@ncf.ca> wrote in
Message id: <alpine.LNX.2.02.1504231344110.30442@darkstar.example.org>:

<Snip>
>But ceramic capacitors don't go bad.

Good post, but I disagree with that. I've seem 'em short out. Just last
week I removed one from a -12VDC rail on an HP 4195A network/spectrum
analyzer. The cap had caught fire.

 

[Chuck]

On Fri, 24 Apr 2015 07:55:58 -0400, JW <none@dev.null> wrote:

On Thu, 23 Apr 2015 14:13:03 -0400 Michael Black <et472@ncf.ca> wrote in
Message id: <alpine.LNX.2.02.1504231344110.30442@darkstar.example.org>:

Snip
But ceramic capacitors don't go bad.

Good post, but I disagree with that. I've seem 'em short out. Just last
week I removed one from a -12VDC rail on an HP 4195A network/spectrum
analyzer. The cap had caught fire.

They short, go leaky, open up, and NPOs in rf circutry can begin to
drift in value. In the 20 years I've used an ESR meter, I've seen 2
electrolytic capacitors which checked good for esr and capacitance be
defective. Never found out what the failure mode was, but they were
in the video amplifier section of Mitsubishi televisions back in the
90s. Symptom in both cases was no video.

---
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com

 

[Jeff Liebermann]

On Fri, 24 Apr 2015 07:55:58 -0400, JW <none@dev.null> wrote:

On Thu, 23 Apr 2015 14:13:03 -0400 Michael Black <et472@ncf.ca> wrote in
Message id: <alpine.LNX.2.02.1504231344110.30442@darkstar.example.org>:

Snip
But ceramic capacitors don't go bad.

Good post, but I disagree with that. I've seem 'em short out. Just last
week I removed one from a -12VDC rail on an HP 4195A network/spectrum
analyzer. The cap had caught fire.

Yep. I have a small pile of Apple Mac Mini computahs that won't power
on. The problem is usually one of the MLCC (mult-layer ceramic
capacitor) caps on the bottom of the PCB.
<http://blog.helpmymac.ru/?p=3585>
They're available in values up to about 1 uF and are commonly used in
place of electrolytics where space is a premium. For automotive
applications, AVX has "Flexiterm" (soft termination) packages that
allow the end terminations to flex a little instead of cracking and
shorting the ceramic:
<http://www.avx.com/docs/Catalogs/flexiterm.pdf>
--
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