What about home made (ESR) Meter ?

[w2aew]

I didnt notice any other esr schematic this way neither
and trust me I've seen a lot of esr schematics...
seems that w2aew made an original esr meter...
thats mainly why I added his schematic on my repository !

The design I put together is a combination of ideas I gathered years ago when I built it.  The oscillator and DUT circuit were from some of the designs I saw, and the bipolar transistor buffer idea from another.  I followed it up with the AD8032 based peak-to-peak detector (DC Restore + Peak Det).

[Rufus]

I didnt notice any other esr schematic this way neither

Maybe because it isn't an ESR meter. It is an impedance meter. At the testing frequency for biggish caps, especially bad ones, most of the impedance is ESR so it serves.

[kripton2035]

I didnt notice any other esr schematic this way neither

Maybe because it isn't an ESR meter. It is an impedance meter. At the testing frequency for biggish caps, especially bad ones, most of the impedance is ESR so it serves.

all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

[FenderBender]

I meant do you want to scan the schematic by chance?

Sorry, still a bit dense. Why would I want to scan the schematic?

I don't have it, but I can get the original any time from the internet. The hand copy that w2aew posted is good enough anyway.

Richard

Oooo my bad. I mixed you up with w2aew somehow...=P

[jimmc]

...
all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

Using a synchronous (phase sensitive) detector allows the reactive and resistive components to be resolved.
For instance this example from your excellent site http://kripton2035.free.fr/analog%20esr/esr-lowohm-silic.html
Here the switches IC3a & b are driven in phase with the test current form IC3c & d, this will reject the quadrature component caused by the reactance of the capacitor under test.

Jim

[kripton2035]

...
all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

Using a synchronous (phase sensitive) detector allows the reactive and resistive components to be resolved.
For instance this example from your excellent site http://kripton2035.free.fr/analog%20esr/esr-lowohm-silic.html
Here the switches IC3a & b are driven in phase with the test current form IC3c & d, this will reject the quadrature component caused by the reactance of the capacitor under test.

Jim

elektor esr meter uses the same switches system
but I note that they both inject +-5v to the capacitor under test
this imply that you cannot use these kind of esr meters in circuit ?

is it possible to use the same system but with +-100mV test signal ?

there are also on my site the russian go esr meter, which uses a constant current method to measure esr
does this method measure only esr and get rid of the reactance or not ?

[amspire]

...
all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

Using a synchronous (phase sensitive) detector allows the reactive and resistive components to be resolved.
For instance this example from your excellent site http://kripton2035.free.fr/analog%20esr/esr-lowohm-silic.html
Here the switches IC3a & b are driven in phase with the test current form IC3c & d, this will reject the quadrature component caused by the reactance of the capacitor under test.

Jim

The quadrature detector only works well if your test signal is a low distortion sine wave. If you look at the schematic circuit of a handheld LCR meter, there is usually something in the order of an 8th order bandpass filter after the signal generator to wipe out all the harmonics. The final signal applied to the device under test has a distortion of something like 0.01% or less. Really accurate bridges go a step further - they also use a bandpass filter after the device under test, partly so they can cope with distortion caused by the device being tested.

If your test signal is a square wave, the harmonics will make the quadrature detector fairly inaccurate in terms of measuring capacitance.

Also, ESR is most easily measured at high frequencies like 100KHz. Capacitance of large electro's needs to be measured at low frequencies like 120Hz.  To get a useful esr reading at the same frequency that you can measure the capacitance value, you would have to start designing to a true LCR meter quality.  Otherwise, you will have to have two test frequencies to measure the C and ESR.

Nice idea, but is it too much work for results that may often be deceptively wrong? How much time are you going to waste chasing up capacitor problems that are the result of a dodgy meter?

The significance of an ESR test is that with electrolytics, ESR rise is usually the first sign of something going wrong in the capacitor - ESR will rise before capacitance starts to fall. ESR will go up if the electrolyte starts to evaporate, so it is a sign of a capacitor that no longer as a true gas-tight seal. So a  basic ESR meter that purely measures the voltage across a capacitor when fed with 100KHz on its own does a pretty good job. Quadrature switching? Don't know why you would even bother.

Richard.

[jimmc]

In the circuit on your website the open circuit voltage is limited by D1 & 2 to approximately+0.7v and the short circuit current to around + 2 mA (5v/(1.8k +470 +68).
With, say, a capacitor with an ESR of 10 Ohm connected the test voltage will be + 2 mA x 10 Ohm = 40 mV pk-pk; so yes, it can be used for in circuit testing.

To answer your second question no, the only way to separate resistance from reactance is to use a synchronous detector.
(An alternative method would be to use a series inductor to resonate out  the capacitive reactance, but this is not practical.)

[jimmc]

...
all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

Using a synchronous (phase sensitive) detector allows the reactive and resistive components to be resolved.
For instance this example from your excellent site http://kripton2035.free.fr/analog%20esr/esr-lowohm-silic.html
Here the switches IC3a & b are driven in phase with the test current form IC3c & d, this will reject the quadrature component caused by the reactance of the capacitor under test.

Jim

The quadrature detector only works well if your test signal is a low distortion sine wave. If you look at the schematic circuit of a handheld LCR meter, there is usually something in the order of an 8th order bandpass filter after the signal generator to wipe out all the harmonics. The final signal applied to the device under test has a distortion of something like 0.01% or less. Really accurate bridges go a step further - they also use a bandpass filter after the device under test, partly so they can cope with distortion caused by the device being tested.

If your test signal is a square wave, the harmonics will make the quadrature detector fairly inaccurate in terms of measuring capacitance.

Also, ESR is most easily measured at high frequencies like 100KHz. Capacitance of large electro's needs to be measured at low frequencies like 120Hz.  To get a useful esr reading at the same frequency that you can measure the capacitance value, you would have to start designing to a true LCR meter quality.  Otherwise, you will have to have two test frequencies to measure the C and ESR.

Nice idea, but is it too much work for results that may often be deceptively wrong? How much time are you going to waste chasing up capacitor problems that are the result of a dodgy meter?

The significance of an ESR test is that with electrolytics, ESR rise is usually the first sign of something going wrong in the capacitor - ESR will rise before capacitance starts to fall. ESR will go up if the electrolyte starts to evaporate, so it is a sign of a capacitor that no longer as a true gas-tight seal. So a  basic ESR meter on its own does a pretty good job.

Richard.

We are talking at cross purposes here, I agree with all you've said.

I'm sorry I did not make it clear that I was talking of measuring only the resistive component and rejecting the reactive part.
For this the presence of odd harmonics would not be a problem and so a square wave is OK.

As I've said before ESR meters are not precision component analysers but a very useful diagnostic tool.
They do not measure capacitance.

I would say that a measurement accuracy of +10% is more than adequate for all practical purposes.

Jim

[amspire]

The point I was making is that if you are measuring at 100KHz, the reactive part is negligible. If you are testing a 1000uF capacitor, you are looking for an ESR better then 50 to 100mohms and the reactive part is 1.6 mOhms. For a 50mOhm ESR in a 1000uF cap, difference between a quadrature detector reading and a plain rectifier will be less then 0.05%.

For capacitors with better then 50mOhm ESR, you are really pushing it to get any accurate reading with two wire measurement.

So all a quadrature/Switching detector is achieving is a technically cool way to rectify AC.  Any other sort of rectifier will work as well, but I have no problem with the quadrature detector as a rectifier.

Richard.

[kripton2035]

ok with all you say
but if you measure always at 100KHz, when you measure the mains capacitor in a smps, its often this one that is bad
and it is only excited with 100 to 120 Hz signal, he never gets 100KHz of its all life except when I test it with the esr meter...
so
also when you see comparison tables of suggested esr values regarding capacitor value
it should be nice to get the capacity together with the esr and indicate directly if these two values are recommended
or not

what do you think of this ?

[jimmc]

The point I was making is that if you are measuring at 100KHz, the reactive part is negligible. If you are testing a 1000uF capacitor, you are looking for an ESR better then 50 to 100mohms and the reactive part is 1.6 mOhms. For a 50mOhm ESR in a 1000uF cap, difference between a quadrature detector reading and a plain rectifier will be less then 0.05%.

For capacitors with better then 50mOhm ESR, you are really pushing it to get any accurate reading with two wire measurement.

So all a quadrature/Switching detector is achieving is a technically cool way to rectify AC.  Any other sort of rectifier will work as well, but I have no problem with the quadrature detector as a rectifier.

Richard.

Once again I am in total agreement.
My original post about synchronous detectors was in reply to the question in a previous post...

all esr meters are impedance meters aren't they ?
how do you make the difference between an impedance meter and an esr meter
apart the fact that esr meters are impedance meters testing at 100KHz ?

Jim

[alm]

ok with all you say
but if you measure always at 100KHz, when you measure the mains capacitor in a smps, its often this one that is bad
and it is only excited with 100 to 120 Hz signal, he never gets 100KHz of its all life except when I test it with the esr meter...
so

Because in normal circuits you want the resistance to be small compared to the reactance (capacitance). In ESR meters, you want the ESR to be large compared to the capacitance, especially if you don't use phase information to distinguish resistance from reactance.

also when you see comparison tables of suggested esr values regarding capacitor value
it should be nice to get the capacity together with the esr and indicate directly if these two values are recommended
or not

what do you think of this ?

It would be nice, but capacitance is hard to accurately measure in-circuit, which would limit the value of this.

[kripton2035]

another question, as you seems very qualified on the subject !

can we consider the esr for a given capacitor is (almost?) the same value at 10KHz and at 100KHz ?
I speak about the esr only, excluding the reactance
thanks.

[alm]

Resistance is not a function of frequency, so the ideal ESR meter would give identical results regardless of frequency, since it would measure ESR independent of capacitance and ESL. Of course this falls apart if the capacitor does not match the C-R-L model.

[jimmc]

The 'perfect' capacitor in series with a 'perfect' resistor is only a model, of a real world capacitor.
At  high frequencies where the reactance is low (compared with the ESR) it is works pretty well, however as the reactance rises other losses start to become significant leading to an apparent rise in ESR.

For an example see http://www.niccomp.com/help/techinfo/NRE-HL102M16V-ESR-Temp-Freq0909.pdf
(Remember that the reactance of 1000uF at 1kHz is 0.16 ohm.)

Jim

[kripton2035]

effectively on the graphic, the esr changes 3 times below 10KHz
but it stays quite the same from 10KHz to 1MHz
thanks for this !

[FenderBender]

Hey guys. Old thread...I know. I asked one of our users,  w2aew, if he'd scan me the schematic of his ESR meter which you can find here on youtube:

I couldn't 100% see the schematic, so here it is.

[kripton2035]

nice idea - I 'll update my esr repository with that schematic. - thanks.

[ramin110]

Hi guys,
I want to build my own ESR meter based on Alan Design but in his design, he used ad8032 IC and I can't find it in my country so I find the ad8572 one and as soon as I plug it to the power supply it fried in an ms so I want to find a common replacement IC for this IC.
Can You help me? Thanks
my design link:
https://www.eevblog.com/forum/beginners/esr-meter-105123/