leakage

[peter g]

hi, how does dc leakage in an electro cap affect the esr and cap value ? does it change either?

 

[Papabravo]

Considering the following spice model of a capacitor - I don't think so. The DC leakage appears to be represented by Rpar, the parallel resistance of the capacitor. This is what will cause an unconnected capacitor with a charge to dissipate that charge over time. Unfortunately the LTspice Help file does not specify a default value for Rpar. It clearly has no effect on the capacitance.

 

[crutschow]

the LTspice Help file does not specify a default value for Rpar.

Since that value varies greatly for different types of capacitors (being relative low for electrolytics and very high for ceramics and film capacitors) I don't see how you could have a representative default value.

 

[Papabravo]

Since that value varies greatly for different types of capacitors (being relative low for electrolytics and very high for ceramics and film capacitors) I don't see how you could have a representative default value.

That is true but there are certainly other items in the LTspice universe that have incomprehensible, arbitrary, and unspecified default values. I don't think this situation is unique by any means. Whenever there are default values it would be helpful to know what they are, can they be ignored, what happens if I do ignore them.

 

[danadak]

If you think about leakage its effectively a resistive kind of component in parallel
with cap....? So that would not affect the value of C one would think.

However esr, being in series, definitely effects the "effective" value of a cap. One
can use the series to parallel transform to find the effective C. And vice versa,
one can use parallel to series to get a proper design value. Not usually used
for electrolytic, one just normally uses more C to offset esr effects. Diminishing
returns I might add.

Series to Parallel Conversion

I am not, however, an expert in any field, so take above with a grain of salt.

I do often look at actual spice models to see how complete they are as often
they are very simple, too simple, for precision analog work. Often incomplete,
not representing the device topology. Good example of this these days are RRIO
OpAmp inputs, their crossover distortion resulting from non ideal switch over
between N and P channel input structures. High res A/D work where this matters
for example.


Regards, Dana.

 

[Kerim]

hi, how does dc leakage in an electro cap affect the esr and cap value ? does it change either?

It surely affects the reading of my capacitance meter.

When my humble capacitance meter reads a relatively higher value than the one written on the electrolytic capacitor, it is a sign for me that the capacitor is made since long, if not defective, for having a relatively low Rpar. I usually try to revive such an old capacitor by repeating the charge-discharge cycle till the LED (in series with the charging resistor) dims considerably at the end of charging.

Kerim

 

[Nigel Goodwin]

As most capacitance meters work by measuring charging times, then a leaky capacitor will read high, as it takes longer to charge it. Likewise, it will discharge faster. But electrolytics do tend to be slightly leaky, it's part of the way they work

 

[Kerim]

As most capacitance meters work by measuring charging times, then a leaky capacitor will read high, as it takes longer to charge it. Likewise, it will discharge faster. But electrolytics do tend to be slightly leaky, it's part of the way they work

Rpar likely depends on the DC voltage on the capacitor (below breakdown). Also, it could vary by repeating the measurement, it usually tends to increase till it reaches a certain high limit.

I wonder if there are meters that measure the leakage resistance.

 

[Nigel Goodwin]

Rpar likely depends on the DC voltage on the capacitor (below breakdown). Also, it could vary by repeating the measurement, it usually tends to increase till it reaches a certain high limit.

I wonder if there are meters that measure the leakage resistance.

Easy enough, just wait until the capacitor is full charged and measure the leakage current that continues to flow - a simple multimeter and a voltage source is all that's required.

 

[Kerim]

Easy enough, just wait until the capacitor is full charged and measure the leakage current that continues to flow - a simple multimeter and a voltage source is all that's required.

Yes, this is the simple accurate practical method. But I forgot that I used doing this in the past when real necessary.
So, I wonder when my short memory (because of age) will let me forget even my name

 

[danadak]

https://product.tdk.com/en/contact/faq/capacitors-0015.html#:~:text=There%20are%20two%20basic%20ways,source%20(See%20Figure%202).

https://assets.testequity.com/te1/Documents/pdf/keysight/capacitance-leakage-measurement-to.pdf

~ What's all this Trapped Charge and Dielectric Compression stuff anyhow? ~

A different veiw of Capacitor leakage - soakage - Q losses - ESR and Dielectric Absorption or DA - Is Capacitor construction and physical compression of the dielectric also part of the problem ?

keith-snook.info

C leakage, and in attached.


Regards, Dana.