Mica Capacitor Stumper

[W2VW]

 Working on a tank circuit for some sand state devices on 160. I needed about 20,000uf so I found some of those big old white CDE mica jobs with the aluminum ends.

.02 MFD rated for 22 amps at 1000 kc. I figured it would get warm at the power level involved. It did.

I reconfigured the thing to a bunch in series parallel but it was not working properly. Went back to the original configuration to try some different tests but this time I grabbed a different mica cap with the same .02 value.

The cap didn't work. It made me nuts for a while because I didn't consider that possibility. I had even measured it first. After a few hours of "troubleshooting" I took the cap out and measured it again. Still read OK. Another cap was placed into service and the thing works FB.

What gives? Anyone seen this before?

[DMOD]

I needed about 20,000uf so I found some of those big old white CDE mica jobs with the aluminum ends.

Did you mean you needed 20,000 pico Farads? 0.001 uF = 1,000 pF, 0.02 uF = 20,000 pF.

What do you mean it didn't work? Did it short, not resonate?

Phil - AC0OB

[WV Hoopie]

Have to ask, when you say mica, is it a mica or a silver mica?

I've never had any luck testing silver micas on a T0-6A. They test good on value and insulation resistance, but do not work. So when I have a circuit that doesn't want to go along with the program, the toobs & resistors test good, and has silver micas. All the silver micas get shot gunned! Replace them all (with new ones because of silver migration).

Craig,

[KM1H]

Silver migration disease (aka SMD) is becoming a huge problem to vintage gear restorers. Ive blown up several at the 500-1000W level that tested good on a Solar CE.

The old WW2 regular TX micas arent much better having absorbed years of moisture. Ive blown them apart in a 160M shunt fed vertical years ago.

One way to test is with a GDO, calculate the coil necessary to be near the frequency of interest. A sig gen and scope is an alternative.

Carl
KM1H

[W2VW]

I needed about 20,000uf so I found some of those big old white CDE mica jobs with the aluminum ends.

Did you mean you needed 20,000 pico Farads? 0.001 uF = 1,000 pF, 0.02 uF = 20,000 pF.

What do you mean it didn't work? Did it short, not resonate?

Phil - AC0OB

Yup 20,000 uf. Sandstate stuff.

[W2VW]

Silver migration disease (aka SMD) is becoming a huge problem to vintage gear restorers. Ive blown up several at the 500-1000W level that tested good on a Solar CE.

The old WW2 regular TX micas arent much better having absorbed years of moisture. Ive blown them apart in a 160M shunt fed vertical years ago.

One way to test is with a GDO, calculate the coil necessary to be near the frequency of interest. A sig gen and scope is an alternative.

Carl
KM1H

Thanks Carl. The thing is 70 years old. I suppose if it worked a little better it may have exploded.

Next question is anything besides heat and drift to look for with the others I have in use?

[k4kyv]

I suppose the end of vintage gear will come with widespread component failure.  Carbon resistors, wax paper caps, tubes, mica caps, HV electrolytics, transformer insulation, are all becoming increasingly problematic, and these components are getting harder and harder to find and fewer and fewer are still being manufactured.

[w3jn]

Check 'em on a LCR bridge, paying close attention to D or Q factor.

I have a box of those brown bakelite HV micas, threw about half away for DC leakage and/or poor D factor.

[WB6NVH]

Just to clarify, I believe at 22 amps he's referring to a "bucket mica" capacitor as used in broadcast transmitters.  The ones with the big aluminum end caps.

Those can have their issues just like the smaller ones.  They may measure the correct value but D and Q are in the "fail" range, or they break down as soon as any RF current tries to pass through them.

This issue follows Murphy's law precisely and brings back fond memories - - 0200 on a Sunday at an AM broadcast station after a lightning strike trying to find a "close enough" bucket mica in the transmitter hut scrap pile to sub into the T-network at the tower base tuner, but everything which should work, does not, because they are all bad too...sigh...

[Opcom]

Check 'em on a LCR bridge, paying close attention to D or Q factor.

I have a box of those brown bakelite HV micas, threw about half away for DC leakage and/or poor D factor.

Is there a good tutorial to be recommended on D and Q factors? I have an old military checker (no instructions) with lots of knobs that has settings for those, but I am not well informed.

Q, is "Q", right? What is "D"? Is it the same as dissipation factor? Having messed only with things << 1KW on a regular basis, I have not had component issues and not had to worry about these items. I do have >KW stuff but it is the exception and long term projects, so not much hands-on.

[w3jn]

D=1/Q

HP for some reason likes to display D factor on its LCR bridges.  Boonton and GR use Q.

Here's a good writeup on capacitor Q, D, ESR, losses, etc

http://www.radio-electronics.com/info/data/capacitor/esr-df-loss-tangent-q-tutorial-basics.php

[W2VW]

D=1/Q

HP for some reason likes to display D factor on its LCR bridges.  Boonton and GR use Q.

Here's a good writeup on capacitor Q, D, ESR, losses, etc

http://www.radio-electronics.com/info/data/capacitor/esr-df-loss-tangent-q-tutorial-basics.php

Thanks Johnny.

So HP = 1/Boonton?

[WA1GFZ]

D is actually more useful information. When pulse lightning hits some dielectrics the cap may not be damaged in all cases, but during the pulse the value of the cap could change a lot. 50% less in the case of XR7 ceramic.   

[w3jn]

Thanks Johnny.

So HP = 1/Boonton?

It does appear that way 

Git yer bad self a digital RLC bridge.  HP 4261A is a good lil' unit, paid $50 or so for mine.  One of the most used pieces of test equipment in the shop.  It only tests at 120 hz and 1 KHz tho'.  Still, that's good enough to find bad electrolytics whose ESR has gone to hell, and give you an indication that things have gone south in RF capacitors.