Old Capacitor Testing Question

[Phrazemaster]

Hey all, I've got some old Gudeman square metal capacitors, .5uF on the label. When I try to measure capacitance with my Fluke 87, it gives a constant stream of fluctuating values ranging from .530uF to .620uF (roughly). The values zing by so fast and I'm confused as to why this would happen. When I've tested newer capacitors, the values fluctuate slightly on the last decimal, but not like when I'm testing these Gudemans. Note the values do center around .5uF, the rated value.

I'm wondering why this might be the case - does this indicate these are bad? They are hermetically sealed oil caps, and seller stated the "tested good" but I'm skeptical to use them.

I know I know, don't bother with old stuff, just get some new Wima film and be done with it! Well this is for a vintage restoration and ideally using vintage parts would please the client.

Thanks for your thoughts.

Mike

 

[Brian Roth]

Try a different meter? I sometimes have funny measurements with my Fluke 87V. Years ago, I bought one of these:

https://anatekinstruments.com/products/fully-assembled-anatek-blue-esr-meter-besr
Oops, I bought and assembled the kit version:

https://anatekinstruments.com/produ...assembly-besr_kit?_pos=2&_sid=0b62c28d0&_ss=r
It lets me "take another view" of capacitors, plus I found it useful to find a shorted PSU rail capacitor along a string of them on a PCB. The ability to see very low resistances is a bonus.

I've been looking at some of the Atlas branded gizmos as well...I'm a test gear junkie! lol

Bottom line....see if a different instrument yields other results.

Bri

 

[MaxDM]

Use your resistance setting, in the 200K or megaohm setting and make sure there isn't any dc resistance, which would imply leaking.

You can also put some dc voltage across the caps, with one side connected to a resistor to ground, and check there isn't any current going through

 

[living sounds]

You can buy a very inexpensive multi-function tester on ebay that can analyze a capacitor, it gives readings of ESR and Vloss on top of capacitance. It also analyzes many other passive and active parts. Pretty neat.

 

[JohnRoberts]

If they are very old caps that have been sitting unused they may need to be formed in... Perhaps through a series resistance bring them up to rated voltage and let them sit a while.

Of course could be a quirk of the meter.

JR

 

[Bo Deadly]

If you really want to use an old cap you might want to test it a little more than with just a meter. In particular I would look at ESR and leakage. I'm not sure what the standard ESR test is but if you have a function generator and a scope could put a square wave across it (maybe through a diode) see how the square wave charges it up. If the voltage droops back down when the square wave stops contributing current or if the voltage jumps up a little when the square wave first starts to supply current, the amount of droop / jump is a function of ESR. Compare that to a quality film resistor. Too much ESR is not desirable for audio or just about anything as it is effectively like blunting the function of the capacitor. For leakage connect a 1M in parallel, charge up the cap to 10V or so (you can use the square wave through diode again at a really low frequency) and see how quickly the voltage drops. Compare to a good film cap.

 

[emrr]

Those old oil caps seldom fail, and they have a sound. ESR will be high if they’re well used, lower if not. Always higher than anything new.

 

[dom]

ESR is best measured with an LCR meter. Old HP or Stanford research models are pretty cheap these days.

There are lots of ways to measure leakage. The thing to be careful of is not confusing dielectric absorption with leakage. When you start to measure leakage there can sometimes be a long settling tail (dielectric absorption). You need to wait for that to settle before obtaining leakage.
One way is to charge up your cap to a DC voltage with an ammeter in series that can measure in the nA to µA range. An SMU is ideal for this, but a DMM in series with a DC voltage supply can also work. Leakage current is measured directly.
Another way is to charge up the cap to a DC voltage, let it sit there for a while to settle. Then, remove the DC source and record dropping voltage on the cap over time with a high-impedance voltage meter (> 10 GΩ). Compute leakage current with I=C dV/dt.
If leakage is really bad, you can sometimes just measure it with a DMM in high resistance (10 MΩ) mode.

 

[Whoops]

This is a great tester,
it measures Capacitance and ESR very well.
It's pretty unbeatable for the price, one of the most useful tools I ever bough