I've run across something I've never encountered before. In testing a non-polarized electrolytic, all checks are ok except one - the value is double or better in each and every one I tested.

I needed to replace a blown yoke return cap in a Zenith BW set. Easy - I've got the replacement Zenith 4.7uFd cap in my parts stash, so just to be safe, I thought I'd run it across my Sencore LC102. ESR, leakage at full voltage and dielectric absorbtion are A-OK. Value? 4.7ufd it should be, but measures 10.9ufd. Wait - lemme check another I have in the used parts stash. Same thing - double the value at 10.7uf. Yes, I know the value can be high for electrolytics (Sprague TVL's are +80/-20 tolerance), but these are clearly labled with the tolerance - 10%.

Opening up a new (packaged) NTE 2.2Ufd non-polarized cap (actually a United Chemicon by the logo on the cap) it too measures 100% higher, and the tolerance on the cap is stated as 20% (M after the value, corresponds with the 20% rating on the United Chemicon datasheet)

Ok, so let me check against a standard cap I have - 10uF (actually 10.0138uFd by the cal label) - dead on. Next, a common Nichicon 4.7uf 100V cap - 5.8uf, well within the 20% tolerance.

Now here's what I think is happening: Non-polarized electrolytics are made by placing two polarized electrolytics of the double the desired value in series. I'm surmising the Sencore LC102 can't discern that they are in series and somehow reads things high. My EDC cap checker (value only, but spot on with other caps) also checks all my non-polarized electros at 200% or higher.

Anybody out there have any non-polarized electros they'd be willing to check for value?

BTW, I put the first cap I checked in the Zenith - it's now got sweep and working fine. All voltages (per Tekfax) all check out just fine as well. Cap runs cool, so no issues with the double value.

Any help is appreciated. Although my original problem is solved, I'm still wondering.:scratch2:

Cheers,

I don't have one right now, but I'm pretty sure the one I tested about a year ago with an LC101 was on value.

Try a leakage test in both directions a couple of times to reform both caps. If you only did the test in one direction, the cap that was reversed biased will not have a very thick dielectric layer formed and will have a much higher capacitance. Then you'll end up measuring close to the value of the smaller cap.

John

Just tried that - no joy. Less than 50uA leakage either way. Value still x2 and the same as before. I found an old 10uF Non-polar cap on an Acer CRT monitor board - it too reads nearly twice the value at 19.8uf. Two crossovers I've got yielded some 8uF caps - old (1976 dated codes) but they read 12uf each. Tolerance isn't stated, so I've no quantitative means to determine validity in checking those. Odd thing is, all the polarized caps and even a 5uf 50V film cap check dead on or within tolerance.

It isn't my LC102, becuse my EDC checker reads the same. The EDC meter is 1980's technology that was a throwaway from NASA Langley. I have a Paco cap tester that uses an eye tube - you adjust the tube for minimum opening of the eye and read the value from the dial. It has 8VAC open circuit across the leads, and reads the Non-polarized caps high as well. Regular caps check fine.

I think I'll dig around the web some more. All the text books mention very little about non-polarized caps.

Cheers,

If you build up a non polar cap by putting two same value electrolytics in series (+to+ or -to-) how do your meters read? :scratch2:

jr

For grins, I just tried the experiment with a couple of 220uf caps that measured close to the same value. I got about 110uf with them connected in series either way. I used a Tektronix DMM850 and an old Heathkit C-3 to make the measurements. The C-3 read a little low (about 100uf).

If you build up a non polar cap by putting two same value electrolytics in series (+to+ or -to-) how do your meters read? :scratch2:

jr

For grins, I just tried the experiment with a couple of 220uf caps that measured close to the same value. I got about 110uf with them connected in series either way. I used a Tektronix DMM850 and an old Heathkit C-3 to make the measurements. The C-3 read a little low (about 100uf).

Ditto herre. Two 10uf electros connected in series back-to-back (-++- or +--+) yield the proper division by two, albeit about 10% high at 5.6uf. New and used single unit Non-polarized caps reading 2x the value is odd.

We've got an LC53 at work - guess what's going with me tomorrow?? There's also a college professor friend of mine I'll contact - his doctoral thesis was on dielectrics and electrostatics, a good read if you can stomach the math and physics. He worked with NASA while a graduate student - awesome labs!

Cheers,

I just tested a 10% Xicon 22uF, 50V NP electrolytic with my LC101. It tested out at 25uF.

John

Got a buttload of non-polarized caps from work and tried them. Way more than half read okay, so I apparently have a handful of out of tolerance caps. Thanks for all the input.

Cheers,