Testing of large capacitors 2

[Rodneywagr]

There are a number of large capacitors in power supply circuits on our trains. Typical sizes include 3300uF @ 350V; 1uF @ 660V, 10uF @ 800V.

Does anybody have any recommendations for test equipment to check if capacitors are good, or starting to fail.

A capacitance check on a multimeter will measure the capacitance using millivolts, and I am not sure that this is an accurate representation of what happens when these devices have several hundred volts across them.

Are there any other checks that should be made to test capacitors? (DC Resistance?, Equivalent Series Resistance?) How are power factor correction capacitors checked?

 

[jbartos]

Suggestions:
1. Visit

http://www.gallawa.com/microtech/cap_test.html

http://www.evinfo.com/list.html

http://www.pv.infn.it/~servel/atlas/hv/hedgehog/

http://www.novacap.com/NOVACAPTECHNICALBROCHURE.pdf

etc. for more info
2. Also, one may follow industry standards, e.g. IEEE, IEC, etc.

http://www.ieee.org

IEEE Std 18-1992 IEEE Standard for Shunt Power Capacitors

http://www.iec.ch

 

[BillJ]

For the capacitors in the power supply, a good test is to observe the power supply ripple. As electrolytics age, they dry out and capacitance decreases. This will show up as an increase in ripple. A change in the shape of the ripple could indicate leakage or breakdown.

 

[nbucska]

If electrolithic, measure capacitance and series resistance
after several minutes at the operating voltage. <nbucska@pcperipherals.com>

 

[peterb]

We have tested 15 kV capacitor units using a multimeter - this is the test method that was specified by the manufacturer.

 

[etronics]

Testing Caps with just a multimeter is pretty useless. Many times a cap will fail operating undervoltage and test fine with a multimeter or &quot;C&quot; meter. This is true for any capacitor including EHV caps used for PF correction on transmission lines.

I have had great sucess using a capacitor analyzer made by Sencore and no I don't work for the company. It saved my bacon many times over when trouble shooting power supplies, and circuits.

I have seen so many techs/engineers use a multimeter to test caps and conclude that they were good and then could not find anything else wrong with the device under test. It is a very common mistake.

For high power caps I would suggest possibly Hi pot/Doble/megger. I have not tried any of these myself but I'm sure if you contact the manufac. they will be able to point you in the right direction.

Good luck.

 

[nukeman]

Try SENCORE instruments. they have a meter (LC102??) that tests the caps at their rated voltage up to 999V. the meter tests for capacitance, ESR, leakage current, and dielectric absorption.

 

[Rodneywagr]

Thanks for your help.

The Sencore LC103 looks like the instrument for the job.

For those interested in the method used to date, we wired a 40W, 110V lamp in series with the capacitor under test and a variac and isolation transformer/rectifier, and slowly increased the voltage across the capacitor up to working voltage. We would then remove the supply voltage, and measure the time it took for the voltage to drop 50 volts. Anything less than 1 minute was a reject. If the capacitor had an internal flashover, the lamp would limit the current flowing and protect the power supply.

I am sure the proper instrument will provide more accurate results.

Rodney.

 

[torontoele]

for large cap like yours,you need do 2 tests:
1.insulation test,use meg meter.
2.capacity test,apply voltage(from 0 to full rate),read the current and caculate its capacity.
!when apply voltage be sure current in a limited value.
!after test,don't forget discharge the cap.
torontoele

 

[marshg]

Look for an ESR tester. I used one to test caps when I was fixing TVs, VCRs, etc.
Here's how it worked. btw, this was an analog tester.

Short the 2 leads of the tester together. This measurement is the same measurement you should get when testing a capacitor (Don't rely on this method using caps smaller than 10uF). It should appear as if the cap is shorted (even though it's not...check to make sure the cap is not shorted either using an ohm meter). If the value is significantly different, then the cap is bad.

 

[nbucska]

Consult the

http://www.thomasregister.com

<nbucska@pcperipherals.com>