Cap Quiz - what is it?

[waltk]

So I have these two caps that look very similar.  One is marked with "104K X7R", the other is marked only with "104S".  They are clearly both .1uF caps, and the first is one is clearly a ceramic cap (the X7R gives it away).  How would you tell what dielectric material was used for the second one (assuming you just found it in a parts bin with no other indication what it was)?

(I already have an answer, but I wanted to hear some other ideas - so it's a quiz  )

[frequencycentral]

the other is marked only with "104S".

The S suffix indicates that this is the sport version, higher performance, corners like it's on rails, but more expensive to insure.

[anchovie]

"S" as a tolerance code means -20%/+50%, which I've only ever seen on ceramics.

[waltk]

The S suffix indicates that this is the sport version, higher performance, corners like it's on rails, but more expensive to insure.

Good guess! A very plausible answer.

"S" as a tolerance code means -20%/+50%, which I've only ever seen on ceramics.

The "S" could be a tolerance code, however, the "S" tolerance code is not applied consistently by manuafacturers.

The mystery cap looks almost exactly like the known-ceramic cap, and seems a bit too small to be a film cap with this value.

Here's another hint:

The mystery cap in question had a measured value of .10005 uF and a dissipation factor of .0054.

[flintstoned]

Silver mica

[waltk]

OK.  Thanks for playing along with my quiz.  I have about 5000 of these mystery caps, and I really wanted know what they are.  So I used the destructive method - a few seconds on the belt sander, and the top rubs right off...



They still look kind of similar, so you need to crunch them up with pliers to see the rest.  The ceramic X7R cap has a solid (ceramic) core as the dielectric.  A little hard to tell from the picture, but it's obvoiusly a hard crunchy thing.



The mystery cap is a metalized film cap.  As you're crunching it up, you pull apart all the little filmy layers.  The "S" marking is just just some unknown manufacturer-specific code.

[Toney]

 Wow! That looks interesting, thanks.
Belt sander- Hope you wore a mask!

[Galego]

Ok. Now put them back together. 

[waltk]

Ok. Now put them back together.

No problem.  I did this same test on a couple thousand of these, put them all back together, and they turned out just fine.  See?  .



But seriously...  I was thinking about your suggestion (joke).  The film is wrapped around inside the cap with the radial axis of the wrapping in parallel with the leads.  Suppose you have some of these .1uF caps, but need a .075uF cap to tweak a circuit.  If you skimmed off just enough of the top to reduce the capacitance to .075uF, you will have manufactured your own cap!!! (assuming it still works at all)  What do you think the chances are of actually making this work?

Turns out it works fine.  This was my first attempt...

Before cap surgery:


After cap surgery:



Then I just dipped it in a little epoxy to make a new top for it - and voila - I now have an endless supply of caps that are variable to any value I want (less than the original value).



Edit:  I think this one of coolest things I've personally discovered (although I'm sure I'm not the first).  I should make a kit with a bunch of these caps and some sandpaper - I'll call it "the adjustable capacitor kit".   ...caveat - you need a good meter so you know when you've reached your target value while sanding.

[CynicalMan]

You can do the same thing with resistors:
www.all-electric.com/schematic/res_trim.htm

[waltk]

You can do the same thing with resistors:
www.all-electric.com/schematic/res_trim.htm

Wow - thanks for the link.  I was thinking it wouldn't work with a metal-film resistor because the metal would be too thin.  I tried a quick test, and it worked.

I started with a 2.4K metal film resistor (that measured 2393 ohms), and with a couple quick strokes of a file, I had resistor of exactly 2472 ohms.
(Some of you will recall that this is exactly what is suggested for R.G.'s geranium transistor tester.)

[amptramp]

It turns out you can freeze a liquid crystal display, cut it to size and seal the end before anything thaws.  We had one consultant try to sell us this idea, but it turns out he stole the idea from another company.

[RedHouse]

After cap surgery:

OH MAN! that is sooo.... One flew over the cookoo's nest!

Word-up dude, I'd refrain from component decapitation pics on the internet if I were you.  

[waltk]

OH MAN! that is sooo.... One flew over the cookoo's nest!
Word-up dude, I'd refrain from component decapitation pics on the internet if I were you.

Yes, thanks for the warning - but nurse Ratched has already wacked my pp for this infraction of the rules.

[candidate]

Any increased potential for noise with those lids on the caps?

[waltk]

Any increased potential for noise with those lids on the caps?

Don't see why there would be.  The caps are justed coated with epoxy in the original manufacturing process.  As far as I know, a new epoxy coating would be just as non-conductive and wouldn't provide any more or less magnetic shielding.  I think as long as you didn't short any layers of the dielectric when you are sanding them off, and the ESR of the modified cap was comparable to the original, you would get similar performance.

When I think about the function of caps, I just consider the capacitance, voltage rating, and ESR - If these are all comparable, I can't think of how you would get more or less "noise" in a circuit.