Using Models to Predict Capacitor Leakage Current

[oldway]

You consider that an abused capacitor is a damaged capacitor...I doe not think it is so easy....

For example an abused X2 capacitor is not imediately damaged...It needs a lot of abuse to be damaged and the damage he suffered is only a reduction of it's capacitance.

So, the first thing to determine is what is a damaged or bad capacitor ?

For me, as user of capacitor tester, a bad capacitor is one who will not works correctly in the circuit it will be used...

What do you consider as bad capacitor ?

[oldway]

What are the most important characteristics of a capacitor that we must check to be sure it will work correctly in circuit ?
1) capacitance
2) leakage current at rated voltage : this is in reality two
 different characteristics: quality of dielectric + dielectric strength of isolation (ability to sustain the rated voltage without breakdown)
3) ESR  (not always important)

A 20V tester does obviously not test the dielectric strenght of a capacitor with rated voltage higher than 20V.

[bugi]

What are the most important characteristics of a capacitor that we must check to be sure it will work correctly in circuit ?
1) capacitance

I guess depending on the capacitor type, that should be "capacitance at the working voltage and frequency", since both parameters could affect the effective capacitance significantly.

[oldway]

What are the most important characteristics of a capacitor that we must check to be sure it will work correctly in circuit ?
1) capacitance

I guess depending on the capacitor type, that should be "capacitance at the working voltage and frequency", since both parameters could affect the effective capacitance significantly.

The most important characteristics of a capacitor are those indicated on the capacitor, that's:

- non electrolytic capacitors: capacitance (nothing more is specified), tolerance, rated voltage
-electrolytics : capacitance, rated voltage, working temperature (85, 105C°)

That are the first characteristics to be checked to know if the capacitor is bad or good.

Without testing if a 400V capacitor is able to sustain 400V applied voltage, without breakdown, you miss one of the most important information you need to check to know if the capacitor is good or bad.

Leakage current is only a secondary characteristic....that's why it is not specified on the marking of the capacitor.

[bugi]

The most important characteristics of a capacitor are those indicated on the capacitor, that's:
...

That would be quite an oversimplification.

Afaik, the markings on a cap are mainly for identifying it, not for characterizing it, although in many cases the markings are conveniently obvious and enough for the latter, too. What really is important is defined by the circuit the cap is in (or schematics etc.).

E.g. some caps have no markings on them at all, would that mean no parameter is important?

I do not know of any electrolytic cap to have ESR marked on them, but in many circuits (especially in various power supply types) the ESR can be more important than the capacitance (as in, the allowable range of values can be tighter), even so that ESR must be within some range, instead of being less than some value, otherwise the circuit won't be stable. (Imho, a stupid design idea, but I guess the goal is to save 0.1 cents per device, ignoring any future costs of more difficult repairs or even full replacement of the device).

Also, consider e.g. ceramic caps, where some types have a huge reduction in effective capacitance as one gets higher and higher in the voltage, and the nominal capacitance is stated at some low voltage (e.g. TDK rates capacitance of ceramics even down at 0.5+/-0.2Vrms, another at 0V+/-?), instead of at rated voltage. If one accidentally replaces C0G with a X5R or such types in a circuit that works at higher voltage, things can fail to work. (I'd have given an example for the reduction, but of course, now that I want to find such graph, I can not spot any in the datasheets, and just couple weeks ago I seemed to see them in every single one... )

Of course, if the capacitance is out of spec already at a low voltage, it will be bad cap, but hey it was the same for the leakage current. According to you, test has to reveal all the issues, even the high voltage ones, or it is of 100% unusable. Right?

[oldway]

Oversimplification? You are kidding ? 

A 2 transistors LV tester is not even capable of making this oversimplificate and basic tests.....

What are looking for ?

There are Keysight testers to make the tests you want to do....but they cost an arm...

Nothing to do with a condensator tester LV ou HV.

If you need more tests than the oversimplificate and basic tests, simply replace the condensator by a new one, it is cheaper than to buy a 10K Keysight instrument.

If you are speaking about measuring characteristics for new products development, nobody working as engineer will be satisfied with predicting leakage current, he will measure it in the circuit.

NB: capacitores with low ESR are identified by a marking on the capacitor.
For exemple, here CJ = low ESR with specifications to -40°C
http://www.nichicon.co.jp/english/products/pdfs/e-cj.pdf

[GEOelectronics]

Oldway said: An HV tester can also have other applications like testing high voltage zener diodes, "


You are correct that HV must be used to perform certain tests, and I add that capacitors vary by type radically. One type I test is vacuum capacitors. There is a special tester for this, it uses to 50kV high frequency AC. If the vacuum is intact, no sparks inside. If there is air inside there are sparks. CRTs tested same way.

Leakage is not always "simply resistance" that can be measured with low Voltage.

"Meggers" have been around since the early days for this same reason.

Capacitors that use liquid or paste dielectric are another can of worms.

Geo

When I think of capacitors it may be a very different mental picture of what the next person thinks of. Notice how the High Voltage, very ultra low leakage types are always handled and shipped with the leads shorted. These types can accumulate a really high charge from the air and hold it for a surprisingly long time.
\edit-30 May 2018 Capacitor illustrations added for clarity.

Geo>K0FF

[oldway]

Any electrician knows that an insulation resistance measurement must be done in high voltage with a megger and that measuring the insulation resistance with a multimeter is useless because this gives a false (to high) value ....

Similarly, to test a capacitor, it is necessary to measure its insulation resistance (or its leakage current) at its nominal voltage, not under low voltage.

NB: At and up the breakdown voltage, model of the leakage current changes drastically and became highly non linear.

That's why values of insulation resistances measured with a megger and with a multimeter are totally different IF THE INSULATION IS BAD AND DOES NOT SUSTAIN THE TEST VOLTAGE.

[joeqsmith]

When I think of capacitors it may be a very different mental picture of what the next person thinks of. Notice how the High Voltage, very ultra low leakage types are always handled and shipped with the leads shorted. These types can accumulate a really high charge from the air and hold it for a surprisingly long time.
\edit-30 May 2018 Capacitor illustrations added for clarity.

Geo>K0FF

I like the lamp.

[joeqsmith]

All for caps were then allowed to reform for a full 10 hours in addition to what had been done. 

[BravoV]

All for caps were then allowed to reform for a full 10 hours in addition to what had been done.

Joe, with that cool logging there, is there possible to log the leakage during the reform ?

[joeqsmith]

All for caps were then allowed to reform for a full 10 hours in addition to what had been done.

Joe, with that cool logging there, is there possible to log the leakage during the reform ?

I don't see why not.

[BravoV]

All for caps were then allowed to reform for a full 10 hours in addition to what had been done.

Joe, with that cool logging there, is there possible to log the leakage during the reform ?

I don't see why not.

Great  , really curious whether it will be linear or "else", which interest me, thanks and expecting to see the results. 

Also there are Japanese capacitor manufacturers app note that stated temperature also plays role in this reforming period, and most stated they did it at high temperature.

[joeqsmith]

We had always reformed them at room temp. I can tell you that the curve was not linear for these caps.  I was watching them off and on during the first few cycles.   

[BravoV]

Great, looking forward to see the "curve".

[joeqsmith]

Great, looking forward to see the "curve".

You asked if it was possible to which I responded I didn't see why not.  I think you are reading too much into my answer to your question. 

Don't you have a UT61E and power supply where you can run your own test?   Join the fun.  Let's see what you come up with! 

[BravoV]

Great, looking forward to see the "curve".

You asked if it was possible to which I responded I didn't see why not.  I think you are reading too much into my answer to your question. 

Don't you have a UT61E and power supply where you can run your own test?   Join the fun.  Let's see what you come up with!

LOL .. my mistake. 

Actually I did , like this below for polymer cap, its new old stock Sanyo OSCON purple cap, leakage curve logged for about 1.5 hours below, the vertical axis 1 Volt = 0.01 uA, and its started to decrease at 30 minutes.



Its just I don't HV DC psu, really interested at those mains rated electrolytic caps, like popular 350 or 400 V ones.

[joeqsmith]

LOL .. my mistake. 

Actually I did , like this below for polymer cap, its new old stock Sanyo OSCON purple cap, leakage curve logged for about 1.5 hours below, the vertical axis 1 Volt = 0.01 uA, and its started to decrease at 30 minutes.



Its just I don't HV DC psu, really interested at those mains rated electrolytic caps, like popular 350 or 400 V ones.

Added your link to that 2015 data:
https://www.eevblog.com/forum/beginners/is-there-any-practical-or-quick-n-dirty-method-to-measure-capacitor-leakage/

I take it you are not comfortable working with something like this.   Let me see if I can locate some parts that have not been used for at least a couple of years. 

[BravoV]

Added your link to that 2015 data:
https://www.eevblog.com/forum/beginners/is-there-any-practical-or-quick-n-dirty-method-to-measure-capacitor-leakage/

I take it you are not comfortable working with something like this.   Let me see if I can locate some parts that have not been used for at least a couple of years.

Thanks.

That thread started as question and ended with the test, the more accurate link at post #10 ->

https://www.eevblog.com/forum/beginners/is-there-any-practical-or-quick-n-dirty-method-to-measure-capacitor-leakage/msg722332/#msg722332


I'm not afraid of working with high voltage, its I'm just an enthusiast and I don't own any HV DC PSU with CC feature.

Bit off topic if you don't mind, actually I'm looking at this circuit for quite sometimes now for candidate of simple DIY, its just it doesn't have the adjustable CC.



Source -> EDN: Regulate a 0 to 500V, 10-mA power supply in a different way

[PA4TIM]

For leakage tests or reforming you do not need a psu with CC. Just use a resistor. That is needed for safety as a cap  goes  short.  I never logged reforming. What I do with real old ones I want to survive is doing it in steps, for example first 10V, if the leakage then is under 50 uA, I go to 50V, wait until leakage is low enough and  then I increase the voltage again etc. The charging is not interesting, you want to see the current decrease while the voltage stays the same. 

[BravoV]

I never logged reforming.

You should try it as you have far better tools and also you show you're interested in capacitor all this time, its interesting at least for me.

As my experience with polymer cap, 1st I thought it will be just boring with stable or at least predictable leakage, cause I was "assuming" that polymer type is well known to be better compared to common electrolytic ones, but the fact like my chart, its proven that its also has the "curve" as well, even at it's rated low voltage at 16 Volt.

Not claiming an expert, especially with limited testings, but as my chart shows, even on polymer cap, especially the old one, it "needs time" to wake up, and there is no literature about this particular type of cap when old as far as I'm aware of.

[joeqsmith]

I would guess that most of the higher quality brands will provide a process for reforming, just like measuring leakage.  Some will call out the resistor to use based on the voltage range (using CV).   I've seen notes about an hour per year.   So a cap that has sat unused for 5 years would require 5 hours to reform.   

For example, pages 15 & 18 may be of interest.
https://de.tdk.eu/download/185386/5f33d2619fa73419e2a4af562122e90c/pdf-generaltechnicalinformation.pdf

I've only ever been involved with alum elec reforming.  Are there papers about reforming other types?  I've never looked.

[GEOelectronics]

When I think of capacitors it may be a very different mental picture of what the next person thinks of. Notice how the High Voltage, very ultra low leakage types are always handled and shipped with the leads shorted. These types can accumulate a really high charge from the air and hold it for a surprisingly long time.
\edit-30 May 2018 Capacitor illustrations added for clarity.

Geo>K0FF

I like the lamp.

You might like the tube (Eimac 8171, also called 4CX10000D) we built a tube tester for that used components that were state of the art in 1975. No that was a HVDC power supply- 5kV @ 5A. \ Filament ran 75 Amps, HV was ramped up via a pair of 220V 28A ganged and inductor-balanced Variac a.k.a. "The Big Wheel".The transformer was a pole-pig backwards.

Geo>K0FF

[joeqsmith]

I would put both of those lamps in my office, no problem.   The 5KV at 5A continuous, no thanks..... Well maybe... 

[joeqsmith]

.... really interested at those mains rated electrolytic caps, like popular 350 or 400 V ones.

For your reforming current curve,  I found a United Chemi-Con part number EKMH451VQT221MB25T.  This is NOS, 220uF, 450V, alum elec, sitting for at least five years.   

Datasheet can be found here:
https://media.digikey.com/pdf/Data%20Sheets/United%20Chemi-Con%20PDFs/KMH%20Series.pdf

They spec 1.13 Ohm @ 120Hz, measured 0.266 
DF  max is 0.15, measured 0.036
C +/-20% @ 20C 120Hz, measured 206.9

I=0.02CV (uA ) or 3mA, whichever is smaller, after 5 minutes at 20 deg C.
Where I = Max. leakage current (uA), C = Nominal capacitance (uF) and V = Rated voltage (V)

So that's 1.98mA after 5 minutes.  This datasheet mentions a 30 minute reforming cycle at the rated voltage.  Two pictures show the leakage after 15 seconds, 1 , 2 and 5 minutes.   This part is already well below the specified leakage and could be put into service.