Capacitor Lifetime Testing

[onders]

Hi you all.

I am trying to setup a suitable circuit that applies the desired max ripple current to a capacitor which also has the rated DC voltage accross it so that I could realize the lifetime test.

May anoybody help me what kind of circuit to use and the necessary formulas?

Thanks...

[willawake]

Re: Capacitor Lifetime Testing

i would be interested in that also.

[yanz]

Re: Capacitor Lifetime Testing

im waiting for some gurus to arrive here and give some advice..

[davmax]

Re: Capacitor Lifetime Testing

The request to make this cap setup is not too easy. It depends on the type of caps and their value.

Example. Some can take 3.2 amps (Samxon 3300u 6.3V). To add this level of current into 9 milliohm requires an rms voltage of only 28mV. This represents quite a small power of 90 milliwatts. So what is required to produce the 100Khz ripple is a signal generator capable of delivering the power then a high frequency transformer to reduce the voltage down to about 30mV @ 3.2 amps.

The cap on test must be connected to a DC power supply that has an inductor to block 100Khz power from entering the power supply. The 100Khz output from the signal gen via the special transformer must be fed to the cap on test through a capacitor rated to take the high current ripple. The DC voltage is set to the required level. ALL QUITE A CHALLENGE.

[PeteS in CA]

Re: Capacitor Lifetime Testing

Major cap vendors such as UCC use a sine-wave-oscillator-amplifier instrument that costs a couple of kilobucks in a temperature-controlled environment. A more rudimentary approach, based on dav's comments above would be to use a square wave instead of a sine wave - a square wave is easier to produce - a 50% duty-cycle inverter and a step-down (voltage) transformer of suitable turns ratio. The remaining issue (on which I cannot comment further) is controlling the load current. BTW, when caps get warm or hot, their C increases and their Z decreases, both significantly.

[onders]

Re: Capacitor Lifetime Testing

Hi davmax,
Thank you very much.
Is it possible for you to send a basic schematic of the test circuit you suggest?
Will be very happy upon your help.
Thanks.

The request to make this cap setup is not too easy. It depends on the type of caps and their value.

Example. Some can take 3.2 amps (Samxon 3300u 6.3V). To add this level of current into 9 milliohm requires an rms voltage of only 28mV. This represents quite a small power of 90 milliwatts. So what is required to produce the 100Khz ripple is a signal generator capable of delivering the power then a high frequency transformer to reduce the voltage down to about 30mV @ 3.2 amps.

The cap on test must be connected to a DC power supply that has an inductor to block 100Khz power from entering the power supply. The 100Khz output from the signal gen via the special transformer must be fed to the cap on test through a capacitor rated to take the high current ripple. The DC voltage is set to the required level. ALL QUITE A CHALLENGE.

[davmax]

Re: Capacitor Lifetime Testing

I have put together a sketch diagram of the principle. You would need some test equipment to actually measure the level of ripple in the test capacitor. One method is to insert a low value resistor (50 to 100 milliohm) at the output of the step down transformer and measure the 100Khz voltage across it. But this would change the transformer design and increase the power requirement of the signal generator. The 100Khz signal output must be adjustable from 0 to 5 Vrms.

[davmax]

Re: Capacitor Lifetime Testing

I have put a bit more time into a test rig. Getting it to a workable state. See the attachment. Note the 16 ohm 10W resistor is to ensure the amplifier sees some resistive load, depending on the amplifier it still might not like the capacitive load of C1, capacitive loads can cause oscillation. NOTE the turns ratio of the transformer is incorrect on the diagram it should read 7:1 (The 45:1 represents the impedance ratio i.e the Xc of C1 = 0.33 is seen by the amp as in the region of 45 x 0.33 = 15 ohm.).

The current through the device under test (DUT) can be measured across C1 at 0.33volts/ amp, this holds only for 100Khz.

[onders]

Re: Capacitor Lifetime Testing

Hi again Davmax,
Thanks for the schematics.
C1 and the inductor are used to prevent dc and ac voltages to pass accross the sources, is this true?
Also, as I see, most of the datasheets give the ripple current at 120Hz, not 100kHz, is this scheme directly convertable to 120Hz version?
With this rig is it possible to apply ripple currents in the range from 50mA to 3000mA or so while applying dc voltages from 10VDC to 450VDC?
Thank you very much,
Regards...

[arneson]

Re: Capacitor Lifetime Testing

I posted an old cap analyser that should perform this function in a piece of test equipment.
It's in the for sale section, it will quick charge caps at different frequencies and voltages, recondition caps and measure resistance of cap.

[onders]

Re: Capacitor Lifetime Testing

Thank you arneson,
Here there is a bigger picture of your device.

http://imagehost.vendio.com/bin/view...2e4-e19a4&sp=1

Also it is said to be 33$ here:
http://cgi.ebay.com/Pyramid-CRA-2-El...ayphotohosting

And also, looking its properties, it is not suitable for serios lifetime tests I think. It is just to decide for bad caps...

Thank you.