I decided that it was time to run a test of what happens to the ESR of
an electrolytic capacitor over temperature. As always, there were a
few too many surprises.
http://802.11junk.com/jeffl/Electrolytic-cap-test/
The idea was to test electrolytics that might actually be used in a
computer and which had a history of failing.

I went to considerable trouble to make sure that the lead and contact
resistance did not contribute any errors. The wires were silver
plated and everything possible was soldered. In order to "zero out"
the lead resistance, I crossed the capacitor leads and used pliers to
short them. After some testing at room temperature, I found it to be
highly repeatable. Test setup:
http://802.11junk.com/jeffl/Electrolytic-cap-test/test-setup.jpg
Each run took about 30 minutes.

I ran temp tests on 4 different caps.
Cap A = 1800 uf 6.3VDC. New. 105C. Purchased on eBay.
Cap B = 2200 uf 10VDC. 105C. Removed from Dell with bulging top.
Cap C = 1000 uf 6.3VDC. New. 105C. Unknown origin.
Cap D = 0.22uf 50VDC. New. 85C. Unknown origin.
http://802.11junk.com/jeffl/Electrolytic-cap-test/caps.jpg

Spreadsheet with test results and graphs:
http://802.11junk.com/jeffl/Electrolytic-cap-test/cap-esr-test.xls

Cap A provided the initial surprise. The ESR was so low that the ESR
meter reading was a constant 0.03 ohms above 40C.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-A.jpg
I would guess(tm) that this is a problem with the meter except that
the meter reads my nichrome test wire, with less then 0.03 ohms quite
nicely.

Cap B is a known defective capacitor with a bulging top that had not
quite blown open yet (i.e. no leaks). The starting ESR is well above
normal.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-B.jpg
This produced almost a 10:1 decrease in ESR as Phil Allison predicted.

Cap C was my attempt to find a cap that would give reasonable ESR
values that would not produce the same lower limit problem as Cap A.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-C.jpg
It was better, but also hit bottom at 0.03 ohm. It's difficult to
tell from the curve, but extrapolating from the starting values, it
might have produced a 10:1 decrease in ESR if the meter had
cooperated.

Cap D was another attempt to get reasonable ESR values and which
produced another surprise:
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-D.jpg
Since the ESR was well above the previous attempts, I would have
expected another 10:1 decrease in ESR. Instead, it only decreased
about 1.5:1. This suggests that higher voltage electrolytics have a
more constant ESR, but more testing will be needed to verify that
guess.

So, Phil is mostly correct. ESR does drop 10:1 in some electrolytic
capacitors. Since power dissipation is directly proportional to the
ESR (assuming a constant ripple current), a typical motherboard
electrolytic capacitor will internally dissipate about 1/10th the
power at maximum temperature as it would at room temperature. This
explains an odd phenomenon that I've seen on computers with bulging
caps on the motherboard. They'll sometimes run just fine when hot,
but refuse to turn on function if allowed to cool. Also, I haven't
seen any of this in the capacitor lifetime calculations.


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

"Jeff Liebermann" wrote in message
...
I decided that it was time to run a test of what happens to the ESR of
an electrolytic capacitor over temperature. As always, there were a
few too many surprises.
http://802.11junk.com/jeffl/Electrolytic-cap-test/
The idea was to test electrolytics that might actually be used in a
computer and which had a history of failing.

I went to considerable trouble to make sure that the lead and contact
resistance did not contribute any errors. The wires were silver
plated and everything possible was soldered. In order to "zero out"
the lead resistance, I crossed the capacitor leads and used pliers to
short them. After some testing at room temperature, I found it to be
highly repeatable. Test setup:
http://802.11junk.com/jeffl/Electrolytic-cap-test/test-setup.jpg
Each run took about 30 minutes.

I ran temp tests on 4 different caps.
Cap A = 1800 uf 6.3VDC. New. 105C. Purchased on eBay.
Cap B = 2200 uf 10VDC. 105C. Removed from Dell with bulging top.
Cap C = 1000 uf 6.3VDC. New. 105C. Unknown origin.
Cap D = 0.22uf 50VDC. New. 85C. Unknown origin.
http://802.11junk.com/jeffl/Electrolytic-cap-test/caps.jpg

Spreadsheet with test results and graphs:
http://802.11junk.com/jeffl/Electrolytic-cap-test/cap-esr-test.xls

Cap A provided the initial surprise. The ESR was so low that the ESR
meter reading was a constant 0.03 ohms above 40C.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-A.jpg
I would guess(tm) that this is a problem with the meter except that
the meter reads my nichrome test wire, with less then 0.03 ohms quite
nicely.

Cap B is a known defective capacitor with a bulging top that had not
quite blown open yet (i.e. no leaks). The starting ESR is well above
normal.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-B.jpg
This produced almost a 10:1 decrease in ESR as Phil Allison predicted.

Cap C was my attempt to find a cap that would give reasonable ESR
values that would not produce the same lower limit problem as Cap A.
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-C.jpg
It was better, but also hit bottom at 0.03 ohm. It's difficult to
tell from the curve, but extrapolating from the starting values, it
might have produced a 10:1 decrease in ESR if the meter had
cooperated.

Cap D was another attempt to get reasonable ESR values and which
produced another surprise:
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-D.jpg
Since the ESR was well above the previous attempts, I would have
expected another 10:1 decrease in ESR. Instead, it only decreased
about 1.5:1. This suggests that higher voltage electrolytics have a
more constant ESR, but more testing will be needed to verify that
guess.

So, Phil is mostly correct. ESR does drop 10:1 in some electrolytic
capacitors. Since power dissipation is directly proportional to the
ESR (assuming a constant ripple current), a typical motherboard
electrolytic capacitor will internally dissipate about 1/10th the
power at maximum temperature as it would at room temperature. This
explains an odd phenomenon that I've seen on computers with bulging
caps on the motherboard. They'll sometimes run just fine when hot,
but refuse to turn on function if allowed to cool. Also, I haven't
seen any of this in the capacitor lifetime calculations.


--
Jeff Liebermann

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece of
equipment has a power supply that only goes to a standby mode when the owner
of the equipment switches it "off". Then they go on holiday, and switch it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they were
kept warm but went too high on their ESR to fulfill their circuit function,
when thy fully cooled.

Arfa

"Arfa Daily"

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is
very common with other equipment such as TV sets, and indeed switchmode
power supplies in general. I see examples of this most every week where a
piece of equipment has a power supply that only goes to a standby mode
when the owner of the equipment switches it "off". Then they go on
holiday, and switch it off for real, and when they come back and try to
start it up again - nothing. And the cause is always bad electros that
were fine when they were kept warm but went too high on their ESR to
fulfill their circuit function, when thy fully cooled.

** Would you be thinking of low value electros ( like 1 or 2.2 uF at 400V )
that are used to " kick start" SMPS ?

Notoriously short lived, their failure goes un-noticed until the unit is
left unplugged for a time.

Plagued many TVs and most VCRs right through the 80 and early 90s.

The same dopey problem is causing havoc with LED tube lighting right now -
cos the Chinese have yet to discover it.

Parallels the Yellow Glue disaster nicely.



.... Phil

"Phil Allison" wrote in message
...

"Arfa Daily"

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is
very common with other equipment such as TV sets, and indeed switchmode
power supplies in general. I see examples of this most every week where a
piece of equipment has a power supply that only goes to a standby mode
when the owner of the equipment switches it "off". Then they go on
holiday, and switch it off for real, and when they come back and try to
start it up again - nothing. And the cause is always bad electros that
were fine when they were kept warm but went too high on their ESR to
fulfill their circuit function, when they fully cooled.

** Would you be thinking of low value electros ( like 1 or 2.2 uF at
400V ) that are used to " kick start" SMPS ?

Notoriously short lived, their failure goes un-noticed until the unit is
left unplugged for a time.

Plagued many TVs and most VCRs right through the 80 and early 90s.

The same dopey problem is causing havoc with LED tube lighting right
ow - cos the Chinese have yet to discover it.

Parallels the Yellow Glue disaster nicely.



... Phil



I was thinking more the typical 47uF 35v ones that hang across the supply
pin to the controller IC, and effectively reduce the otherwise high
impedance of the startup supply for a brief time, in order that it can
supply enough current to get the chip started so that the low impedance
maintenance supply from the secondary side can get going and take over. I
see so many caps in that position that are high ESR and stop the supply from
starting, that I order them fifty at a time. And one particular supply that
I repair lots of for a company, has the Yellow Glue syndrome as well, and
where is it ? Yep, all around the base of that same 47uF cap ...

Yes, also used to see a lot of the high voltage 1uF's o/c, but not so much
these days. I was also thinking of secondary side 1000 and 2200uF caps.
Again, when kept warm, they often work ok, despite showing signs of bulging,
but go very high ESR when cold. In TV sets, this often leads to them taking
longer and longer to start, until finally, they won't start at all. This can
be because of excess ripple on the 5 volt rail, upsetting the system control
micro, or excess ripple on the 12 or 24 volt rails, upsetting things like
the backlight inverter, or rails of completely the wrong voltage if the
whole supply is regulated from the bad rail. I have seen 5 volt rails
increase to 10 or more volts because of this, with LSIs elsewhere in the
equipment being destroyed as a result ...

Arfa

"Arfa Daily"
"Phil Allison"

** Would you be thinking of low value electros ( like 1 or 2.2 uF at
400V ) that are used to " kick start" SMPS ?

I was thinking more the typical 47uF 35v ones that hang across the supply
pin to the controller IC,

** Like the UC3842N ?

see so many caps in that position that are high ESR and stop the supply
from starting, that I order them fifty at a time.

** So you are doing warranty work on home studio powered speakers that run
24/7 ?


And one particular supply that I repair lots of for a company, has the
Yellow Glue syndrome as well, and where is it ? Yep, all around the base
of that same 47uF cap ...

** The PSU board in the " Alesis M1 Active" - right ?

Where the problem is not due to bad electros at all.

But a 2W resistor that almost touches the cap and heats it to a high temp
for 8,760 hours per year.

Wot a piece of garbage.


.... Phil

"Phil Allison" wrote in message
...

"Arfa Daily"
"Phil Allison"

** Would you be thinking of low value electros ( like 1 or 2.2 uF at
400V ) that are used to " kick start" SMPS ?

I was thinking more the typical 47uF 35v ones that hang across the supply
pin to the controller IC,

** Like the UC3842N ?

Yes, that series of IC. There are quite a few UC38xx chips about



see so many caps in that position that are high ESR and stop the supply
from starting, that I order them fifty at a time.

** So you are doing warranty work on home studio powered speakers that run
24/7 ?

I do work for several hifi shops, and see quite a few powered subs, some of
which have switchers in them, and suffer from caps, but I also do a fair bit
of work on other items that are powered 24/7 like DVD players / home cinema
systems and set top boxes.


And one particular supply that I repair lots of for a company, has the
Yellow Glue syndrome as well, and where is it ? Yep, all around the base
of that same 47uF cap ...

** The PSU board in the " Alesis M1 Active" - right ?

The actual one that I was thinking of is from a commercial hot drinks
machine. The company that sends me them, has them rented out into workplaces
and takeaways and educational establishments all over the country, so
there's never any shortage of them for repair. It has a small vertical
sub-pcb that has the switch mode controller and driver on it. The 47uF is
laid over on its side, and then fixed to the board with the dreaded yellow
glue. The actual mix of faults is about 50 / 50. If the glue has started to
go brown, chances are that's the primary problem, and the cap is actually
ok, reading normally on the ESR meter. If the glue hasn't gone yellow, then
most times, the cap reads bad. But to be honest, whether it's actually cap
or glue, I always just go ahead and remove the cap, clean up the board, and
then fit a new cap. They are so cheap, it's not worth doing anything else
and risking it bouncing a couple of months later.

Where the problem is not due to bad electros at all.

But a 2W resistor that almost touches the cap and heats it to a high temp
for 8,760 hours per year.

Yes, I've never understood this. It seems to be almost a 'design rule' that
any electrolytic should be mounted as close to a heat source on the PCB as
possible ... Anyone in the repair business knows that this is going to
lead to problems after not too long a life. If I suspect a cap problem on
any piece of circuitry, the first ones that I check are those near to big
resistors and heatsinks. How come this has never managed to get back to the
designers, and the supervisors who approve the designs ? Seems to be common
across all manufacturers.



Wot a piece of garbage.


Many are

Arfa

... Phil

On Sun, 7 Jul 2013 10:44:57 +0100, "Arfa Daily"
wrote:

"Jeff Liebermann" wrote in message
.. .
This
explains an odd phenomenon that I've seen on computers with bulging
caps on the motherboard. They'll sometimes run just fine when hot,
but refuse to turn on function if allowed to cool.

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece of
equipment has a power supply that only goes to a standby mode when the owner
of the equipment switches it "off". Then they go on holiday, and switch it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they were
kept warm but went too high on their ESR to fulfill their circuit function,
when thy fully cooled.

Arfa

It's all too common with servers and machines that are powered on all
the time. Last week, I had to deal with an office computer like that.
Powered on 24x7 and nothing wrong. However, it was full of dust, so I
thought it could use a cleaning. I turned it off, blew out the dust
(outside), and it refused to turn on. When I tore apart the ATX power
supply, there were three obvious bulging caps.

While I had the cover off, I decided to run a little test. I gave the
PCB a blast from a heat gun to get it up to temperature and found that
it will turn on. I have several cheap ATX testers, one of which has a
built in load for the +5V and +12V lines. I connected an oscilloscope
to these lines to see if there was any ripple. Both were fairly clean
(I forgot the numbers) but as the power supply cooled down, the +5V
noise started ominously increasing. The PG (power good) light was
oddly thrashing. When I turned it off, it didn't want to turn on
again. It was too old and ugly to repair, so I just installed a
replacement.


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

On 7/7/2013 8:14 AM, Jeff Liebermann wrote:
On Sun, 7 Jul 2013 10:44:57 +0100, "Arfa Daily"
wrote:

"Jeff wrote in message
...
This
explains an odd phenomenon that I've seen on computers with bulging
caps on the motherboard. They'll sometimes run just fine when hot,
but refuse to turn on function if allowed to cool.

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece of
equipment has a power supply that only goes to a standby mode when the owner
of the equipment switches it "off". Then they go on holiday, and switch it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they were
kept warm but went too high on their ESR to fulfill their circuit function,
when thy fully cooled.

Arfa

It's all too common with servers and machines that are powered on all
the time. Last week, I had to deal with an office computer like that.
Powered on 24x7 and nothing wrong. However, it was full of dust, so I
thought it could use a cleaning. I turned it off, blew out the dust
(outside), and it refused to turn on. When I tore apart the ATX power
supply, there were three obvious bulging caps.

While I had the cover off, I decided to run a little test. I gave the
PCB a blast from a heat gun to get it up to temperature and found that
it will turn on. I have several cheap ATX testers, one of which has a
built in load for the +5V and +12V lines. I connected an oscilloscope
to these lines to see if there was any ripple. Both were fairly clean
(I forgot the numbers) but as the power supply cooled down, the +5V
noise started ominously increasing. The PG (power good) light was
oddly thrashing. When I turned it off, it didn't want to turn on
again. It was too old and ugly to repair, so I just installed a
replacement.


It's not just the power supplies that have that problem.

We have a selective soldering machine, CNC controlled, using a pc with a
no longer available video card.

Ran fine for several years when the machine was on 24 hours per day.
Then when work slowed, and the machine was off for several days, the
video would act up, then settle down and work fine. Time to run properly
took longer and longer, till one day, it wouldn't display at all.

Factory had no replacement boards available, and in fact had quit using
them because of problems. We replaced all the electrolytic caps with
ones close to the original value. Had bags full! And the board has
performed correctly ever since.

Paul

Arfa Daily wrote:

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece of
equipment has a power supply that only goes to a standby mode when the owner
of the equipment switches it "off". Then they go on holiday, and switch it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they were
kept warm but went too high on their ESR to fulfill their circuit function,
when thy fully cooled.


Has anyone tried one of these $18 ESR meters?

http://www.ebay.com/itm/190859294488

"Michael A. Terrell" wrote in message
m...

Arfa Daily wrote:

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is
very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece
of
equipment has a power supply that only goes to a standby mode when the
owner
of the equipment switches it "off". Then they go on holiday, and switch
it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they
were
kept warm but went too high on their ESR to fulfill their circuit
function,
when thy fully cooled.


Has anyone tried one of these $18 ESR meters?

http://www.ebay.com/itm/190859294488


I recently bought one of the similar ones that has component testing on it
as well as ESR

http://www.ebay.co.uk/itm/1810723297... 4.m1439.l2648

and I have to say that it works very well. ESR function agrees nicely with
my Bob Parker

Arfa

Arfa Daily wrote:

"Michael A. Terrell" wrote in message
m...

Arfa Daily wrote:

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is
very
common with other equipment such as TV sets, and indeed switchmode power
supplies in general. I see examples of this most every week where a piece
of
equipment has a power supply that only goes to a standby mode when the
owner
of the equipment switches it "off". Then they go on holiday, and switch
it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they
were
kept warm but went too high on their ESR to fulfill their circuit
function,
when thy fully cooled.


Has anyone tried one of these $18 ESR meters?

http://www.ebay.com/itm/190859294488


I recently bought one of the similar ones that has component testing on it
as well as ESR

http://www.ebay.co.uk/itm/1810723297... 4.m1439.l2648


Don't you love that they never bother to mount the LCDs in these
designs?


and I have to say that it works very well. ESR function agrees nicely with
my Bob Parker


The original, or the newer one? I have the original, bought as a
kit.

"Michael A. Terrell" wrote in message
m...

Arfa Daily wrote:

"Michael A. Terrell" wrote in message
m...

Arfa Daily wrote:

I haven't had a chance yet to fully examine your results links, but I
thought that I would just wade in to say that your last observation is
very
common with other equipment such as TV sets, and indeed switchmode
power
supplies in general. I see examples of this most every week where a
piece
of
equipment has a power supply that only goes to a standby mode when the
owner
of the equipment switches it "off". Then they go on holiday, and
switch
it
off for real, and when they come back and try to start it up again -
nothing. And the cause is always bad electros that were fine when they
were
kept warm but went too high on their ESR to fulfill their circuit
function,
when thy fully cooled.


Has anyone tried one of these $18 ESR meters?

http://www.ebay.com/itm/190859294488


I recently bought one of the similar ones that has component testing on
it
as well as ESR

http://www.ebay.co.uk/itm/1810723297... 4.m1439.l2648


Don't you love that they never bother to mount the LCDs in these
designs?


and I have to say that it works very well. ESR function agrees nicely
with
my Bob Parker


The original, or the newer one? I have the original, bought as a
kit.


The original, also bought as a kit. Did you see the other post that I did
today that detailed a comparison between the Bob P and the little Chinese
job on a faulty cap that I had cause to replace ?

Arfa

"Cap D was another attempt to get reasonable ESR values and which
produced another surprise:
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-D.jpg
Since the ESR was well above the previous attempts, I would have
expected another 10:1 decrease in ESR. Instead, it only decreased
about 1.5:1. This suggests that higher voltage electrolytics have a
more constant ESR, but more testing will be needed to verify that
guess.
"
It might not be so much of a surprise. Do you happen to know what frequency the ESR meter uses ? Thing is, if they choose the frequency too high there will be some error due to ESL in the larger cap values. Too low and the ESR of a 0.22uF will inherently read high because of the Xc component.

In this case you might not be getting empirical data but rather a more "usable" figure for servicing as that is the intended application of most ESR meters.

Granted, the higher ESR of a lower value such as a 0.22 might be well acceptable for use in real circuitry because of course they would not expect it to handle much ripple current. Well not so much current for the obvious reason. If a low value like that is ever used for filtering it is more likely a local cap near a linear regulator to stop oscillation. It would never be the main filtering on a power supply line unless it was extremely low current, and unless they wanted the source to rise and drop fast, a higher value cap would likely be chosen most of the time anyway, like a 1 uF or something, or anything they are already using a bunch of to get a better price.

Actually, if you want to get persnickety about it (that is not a bad thing) you could run the test with a sine wave generator and a scope or voltmeter.. It's not difficult, and you could isolate the ESL component completely by using a set of frequencies that keep the Xc to a minimal value.

And now, I might just have to do that myself because I would like to know how temperature affects ESL. I would bet it does, and depending on the frequency applied might throw off the ESR measurement as would Xc. You are talking from 1,800 down to 0.22, that is like a one to eight thousand ratio. There must be some effect but I am not going to figure it out right now, especially not knowing the frequency applied to the DUT.

Incidentally, years ago I designed an ESR meter but was too lazy to actually build it. My scheme to get a wider range of accurate ESR values for different value caps was to apply a square wave to the DUT. I had two scales discriminated for high and low frequency and the square wave was positive going only so it could detect a short easily. I chose a low amplitude like a couple hundred millivolts to prevent the short/leakage detect from triggering if the polarity was wrong, or if there were semis in circuit that would conduct. Of course if Schottky rectifiers were used it might still false trigger, but then it was just a matter of the user reversing the leads.

I used a pretty low inpedance so that it would read fast, and made it so discharging wasn't necessary. My main concern was mass cap testing at the time as I knew plenty of techs that just check all the caps regardless of the symptom and where they were in the circuit. I don't do that.

In fact I don't even replace all the "bad" ones either, as some are more critical than others. Some are very uncritical. The engineer might use a 100 uF where a 22 would work just fine, but they get a price break using the 100, on all 100s in each unit of course. Basically if there is no AC across the cap when it is working and it is not leaking electrolyte, it does not need to be changed.

Hillbilly ? Fine, but market conditions have made me this way. The only thing that will stop me from ever pushing down the bottom line is recalls, or safety issues.

On Sun, 7 Jul 2013 14:34:58 -0700 (PDT), wrote:

"Cap D was another attempt to get reasonable ESR values and which
produced another surprise:
http://802.11junk.com/jeffl/Electrolytic-cap-test/Cap-D.jpg
Since the ESR was well above the previous attempts, I would have
expected another 10:1 decrease in ESR. Instead, it only decreased
about 1.5:1. This suggests that higher voltage electrolytics have a
more constant ESR, but more testing will be needed to verify that
guess.

Do you happen to know what frequency the ESR meter uses ?

100KHz. That's also one of the frequencies at which ESR is specified
on the data capacitor sheets.

http://en.wikipedia.org/wiki/ESR_meter

Thing is, if they choose the frequency too high there will be
some error due to ESL in the larger cap values.

It's not an issue, but let's see. I had about 30cm of #18 wire leads
connected to the capacitor under test.
http://www.consultrsr.com/resources/eis/induct5.htm
That's about 0.43uH. At 100Khz, that's about 25 ohms. That would
wreck any measurement were it not for a feature in the Bob Parker ESR
meter that eliminates connection and cable resistance and reactance.
When I first press the on/off button, it reads the uncalibrated
reactance, including the what is contributed by the leads and
connections. As I vaguely recall, it was about 25 ohms. I then short
the capacitor leads, and push the on/off button again. The meter now
reads zero. When I unshort the capacitor leads, the meter now reads
the relative reactance, which is just the ESR, and without the cable
and connector contribution.

Too low and
the ESR of a 0.22uF will inherently read high because of the
Xc component.

Bingo. That explains part of what happened. At 100Khz, 0.22uf is 7.2
ohms. What the meter is reading is the vector sum of the 7.2 ohm
reactance and a fairly small ESR value. The result is the vector sum
of the capacitive reactance and the ESR. To get rid of this error, I
could run the test again, this time measuring the capacitance of this
capacitor. I could then vector subtract the calculated capacitive
reactance from the measured value leaving only the ESR.

More of the same:
http://www.eevblog.com/forum/projects/esr-vs-capacitive-reactance/msg38653/?PHPSESSID=4fb1bfbcb84bbc7bc034a67c3ec6dca2#msg386 53

In this case you might not be getting empirical data but rather a
more "usable" figure for servicing as that is the intended
application of most ESR meters.

Well, it's certainly usable for large value capacitors, where the
capacitive reactance is orders of magnitude smaller than the ESR.

you could run the test with a sine wave generator and a scope or
voltmeter. It's not difficult, and you could isolate the ESL
component completely by using a set of frequencies that keep
the Xc to a minimal value.

Yep, a vector impedance meter or network analyzer would do the trick.
However, I'm not sure how well they would perform with extremely small
values of ESR and if they could be used to test caps in circuit.

And now, I might just have to do that myself because I would like
to know how temperature affects ESL.

My guess(tm) is that it would be dominated by the lead inductance.

I would bet it does, and depending on the frequency applied might
throw off the ESR measurement as would Xc. You are talking from
1,800 down to 0.22, that is like a one to eight thousand ratio.
There must be some effect but I am not going to figure it out right
now, especially not knowing the frequency applied to the DUT.

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

"Jeff Liebermann = Idiot "


It's not an issue, but let's see. I had about 30cm of #18 wire leads
connected to the capacitor under test.


** You should used a twisted pair of leads, figure 8 twin lead or even
co-ax, which form transmission lines at high frequencies and series L all
but vanishes. The residual before zeroing is then less than 1ohm.

My god you are a ****wit .


.... Phil

On Mon, 8 Jul 2013 13:59:06 +1000, "Phil Allison"
wrote:

"Jeff Liebermann = Idiot "


It's not an issue, but let's see. I had about 30cm of #18 wire leads
connected to the capacitor under test.

** You should used a twisted pair of leads, figure 8 twin lead or even
co-ax, which form transmission lines at high frequencies and series L all
but vanishes. The residual before zeroing is then less than 1ohm.

Good idea. So, I tried it.
http://802.11junk.com/jeffl/Electrolytic-cap-test/coax.jpg
I get between 0.3 and 0.6 ohms residual reading. The problem is that
the BNC connectors are not the best DC connection possible. Wiggle
the connectors and the reading changes. That's why I soldered almost
everything on my insulated hookup wire version.



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

On Sun, 07 Jul 2013 19:01:54 -0700, Jeff Liebermann
wrote:

It's not an issue, but let's see. I had about 30cm of #18 wire leads
connected to the capacitor under test.
http://www.consultrsr.com/resources/eis/induct5.htm
That's about 0.43uH. At 100Khz, that's about 25 ohms.

Ooops. That should be 0.38uH and 0.24 ohms.

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

I think to be properly persnickety you are going to have to change the vertical scale on those graphs from "ESR" to "impedance".

I have a bunch more to say on this but it is late and I am half drunk. Off work tomorrow...

I wonder what kind of readings you might see out of things like motor start or run caps, or speaker crossover caps.

In fact the run cap in my AC unit took a **** last year and I had to replace it, after that the summer electric usage went down. I wonder if I could go around selling cap jobs to homeowners and claim honestly that it will save them money in the next few seasons.

It would also prevent unscrupulous assholes from selling them a whole new unit. At first I priced new condensing units and it is hard as hell to get them with R22 these days. And then there is some doubt about how well it would work with the original evaporator and cap tube or expansion valve, which might mean tearing the whole damn thing down.

Cap is about twenty bucks, charge a hundred for the job, lasts five years and almost guaranteed to save them that in the next five seasons ? Now that would be some serious math.

I am lazy when it coes to that. I would like easy ways do calculate Xc, Xl and do vector math. there is probably a way out ther eI just haven't found it. In fact I would be happy with a reactance/resonance table. Damn, with a computer it could be zoomable and just eliminate all this.....

You know in the old BASIC days I had programs for that. Also for calculating new displacements of engines bored out, carberator requiremnets, all that cool ****. But I did it once and minded my syntax, and didn't have to do it again.

Until I fell behind on programming. Well I was never really ahead, I just made it do a few things I wanted it to do because I used to design things here abnd there and build them.

I got a bone to pick with time, I lack my ambition, something took it ! Call the electron police.....

On Mon, 8 Jul 2013 01:21:37 -0700 (PDT), wrote:

I think to be properly persnickety you are going to have to change
the vertical scale on those graphs from "ESR" to "impedance".

Nope. The correct measurement would be dissipation factor:
http://en.wikipedia.org/wiki/Dissipation_factor
which is a conglomeration of the resitive part of the impedance (ESR),
and the capacitive reactance part (1/2PiFC). For this exercise, I'm
only interested in the part that actually produces heat, which is ESR.
The reactive parts generates no heat and can therefore be ignored.

I have a bunch more to say on this but it is late and I
am half drunk. Off work tomorrow...

Alcohol should best be used for removing solder flux residue.

I wonder what kind of readings you might see out of things like
motor start or run caps, or speaker crossover caps.

Dunno. I don't work on speakers and the motor caps are really obvious
when they need replacement.

I am lazy when it coes to that. I would like easy ways do calculate
Xc, Xl and do vector math. there is probably a way out ther eI
just haven't found it. In fact I would be happy with a reactance/resonance
table. Damn, with a computer it could be zoomable and just eliminate all this.....

In the bad old daze, I used nomograms for all that. These days, I
just Google for a suitable Javscript based online calculator. I also
have reactance and resonance calcs programmed into my HP41CX and other
HP programmable calculators. However, the online version is easer to
read and use.

Until I fell behind on programming. Well I was never really ahead,

I don't have much trouble programming for myself and on my own
projects. However, I don't like programming and am not very good at
it. To avoid dissapointing friends and customers that want me to
program, I plea ignorance.

I got a bone to pick with time, I lack my ambition, something
took it ! Call the electron police.....

Sorry, but the electron police had its funding redirected by the
government so that the NSA can built the ultimate telephone directory.


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

Do you happen to know what frequency the ESR meter uses ?
Thing is, if they choose the frequency too high there will be
some error due to ESL in the larger cap values.

** Not true for Bob's design.

Even large can electros with bolts on top have ESL values of about 20nH =
much the same value for nearly all electros.

At 100kHz, this results in a reactance of 0.01 ohms = the resolution of
Bob's ESR meter.


Too low and the ESR of a 0.22uF will inherently read high because of the
Xc component.


** Bobs meter is NOT suitable for low value caps ( under 1uF ) or any cap
type except aluminium electros.

In this case you might not be getting empirical data but rather a
more "usable" figure for servicing as that is the intended
application of most ESR meters.

** Most low value electros have high ESRs ( like 20 or 30 ohms) and this
saves the day. Tantalums are an exception, but as they do not have an issue
with rising ESR it is irrelevant to the purpose of Bob's meter.


..... Phil

On Sun, 07 Jul 2013 00:52:47 -0700, Jeff Liebermann
wrote:

The data is rather interesting, and somewhat revealing.
It occurred to me to wonder what the actual leakage current, and
capacitance was for similiar conditions in each case.

peter

On Tue, 09 Jul 2013 10:00:06 +1000, Peter wrote:

On Sun, 07 Jul 2013 00:52:47 -0700, Jeff Liebermann
wrote:

The data is rather interesting, and somewhat revealing.
It occurred to me to wonder what the actual leakage current, and
capacitance was for similiar conditions in each case.

peter

I was going to do that, just to see what happened. However, it was
midnight and I had just enough time to write it up before I fell over.
I need to measure the capacitance over temperature in order to remove
the capacitive reactance from the ESR measurements on the small value
cap (Cap-D).

I don't consider leakage current to be significant at the voltages
found in computer power supplies and motherboards, but testing might
prove otherwise.

Application Guide Aluminum Electrolytic Capacitors:
http://electrochem.cwru.edu/encycl/m...4-appguide.pdf
DC Leakage Current (DCL)

DC Leakage Current is the DC current flowing through
the capacitor with the rated voltage applied. The value
of leakage current depends on the voltage applied, the
charging period and capacitor temperature.

DCL Method of measurement

Measure leakage current at 25 °C with the
rated voltage applied through a protective
resistance of 1000 Ohm in series with the capacitor
in the measuring circuit. Five minutes after the
application of voltage, the leakage current is
not to exceed the maximum value indicated in
the specification.

The 5 minutes delay is to reduce the effects of forming current and
dielectric absorption. There are some typical graphs of leakage
versus temperature and voltage on Pg 2.191. I guess the easy way to
measure current is to just measure the voltage across the 1000 ohm
resistor and compute the current.

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