New Kemet PP Capacitor Measures Low

[Smoky]

I'm matching capacitors tonight and I found this new Kemet .22uf 160v 5% PP film cap measuring way out of value.

It is one out of ten that I purchased. I have two Sencore meters that measure accurately below 1uf and they both measured ~4730pf.

This capacitor passes its leakage test but fails its value test.

The other nine capacitors measured .227, .227, .218, .222, .217, .223, .224, .216, .218uf. So as you can see, there are matched pairs in just this group of nine. More than enough for my vintage equalizer repair.

This Kemet capacitor is of the F461 series.

This is my first new film capacitor that I have found to be off. Thank goodness too that I measure every capacitor that I put in circuit since I don't have the best trouble-shooting skills yet.

[bob91343]

It appears that you found a bad apple in an otherwise good batch.

[Smoky]

I sent Mouser the picture of the capacitor while being tested. They said they will let me know if Kemet would like to have it back

These are PP caps sold in bulk, so someone sticks their hand into a box of 3,000 and pulls out 10. One of the 10 turns out to be bad. That's not what I would call "good-odds" IMO.

The cap looks perfect.

Like I said above, from day-one of starting this hobby a little over two years ago, I've been able to use a very accurate LC-53 capacitance meter. Now I have an LC-102 too. What great machines. Especially when you're needing matched pairs for audio equipment.

[bob91343]

While I agree that one bad out of ten is lousy, you indeed have too small a sample to judge.  Most of the time the rejection ratio runs around 2% so you'd need a larger sample to resolve that.

They don't test every part.  That would run the cost too high for many.

[wraper]

They don't test every part.  That would run the cost too high for many.

They do test every single part. Probably twice during production. Even every single 0.01 cent SMD resistor gets tested.

[Tomorokoshi]

I sent Mouser the picture of the capacitor while being tested. They said they will let me know if Kemet would like to have it back
...

If possible send one example that tests to the correct value.

I had a different capacitor issue with a different capacitor manufacturer some time ago. I was able to collect quite a bit of testing information across new, NOS, and used parts. I sent examples in so they could check against my measurements.

It was an interesting process to make the report and then get their results some time later. Working with engineering or manufacturing is much more interesting than working with sales.

[TimFox]

Since your bad apple is about 50x too small, it might suggest something actually broken inside the part, rather than a normal part with a low value.  What size case would a good 4.7 nF part have in this series?  If the same dimensions as the 220 nF part, then it is more likely a mis-labeled 4.7 nF component due to a manufacturing error.

[Smoky]

Here's a shot of the data sheet showing the 160VDC PP capacitor section. The cap I purchased has a red dot next to it. It also has a 7.5mm lead spacing. There doesn't appear to be a .0047uf capacitor with the 7.5mm spacing listed.

It's funny that it passes the leakage test but not the value. So something's gone haywire somewhere

[Gyro]

Most of the time the rejection ratio runs around 2% so you'd need a larger sample to resolve that.

At 2% rejection ratio, they wouldn't be staying on anyone's approved supplier list. You're out by at least two orders of magnitude.

[TheMG]

It's funny that it passes the leakage test but not the value. So something's gone haywire somewhere

If I had to guess, I'd say there might be a hairline crack where the lead connects to the metallized end of the capacitor inside (in other words, it's gone open-circuit and the reading you get is the small capacitance across the broken connection in series with the actual capacitor). That would cause the capacitance to go way way down while having no effect on leakage.

[magic]

I would believe it if it were 4.7 picofarad, not nano.
If anything, part of the internal foil could have disconnected, reducing total capacitance.

But the capacitance being only 0.4% off from a standard value looks like a bit too much coincidence. I think they somehow managed to take a 4.7nF puppy and package it as 22nF

Be sure to tear it down and compare with a good one if they don't want it back

Any detectable difference in weight?

[Kleinstein]

I could be a 4.7 nF unit from a high voltage series that somehow got it's way to the wrong box before printing the value and type. A different weight may give a hint, but it could still be very similar as there is not much difference in density between PP and epoxy.

It is 1 out of 10 that is bad. However it is this one bad capacitor that made you report the finding. the normal 10 out of 10 good would not be worth a post.
Bad electronic components are rare - unless you by from Ebay or Ali and get fakes.