Jellybean electrolytic capacitors

Thanks, JL, that goes up on my wall. Got the equations that fits the graph? Cheers, Harry

We're not going to teach ac-power-line design, excepting SMPS types. But we may teach signal-transformer design, including high-frequency power stuff.

What're lams, laminations? You mean buy transformers?

John, thanks very much, that's a spectacular graph, with a 1000:1 VA range and an over 200:1 weight range. What kind of transformers (specific types?) did you measure for the data points? Do you still have the raw data? We're using Igor for our graph plotting.

Good lead, then I can get them through Newark.

I'm willing to go for a few, but I imagine they're so overpriced that could be an issue. If you'll suggest a few candidates and provide link info, I'll pursue them and try to include the detailed measurements. I'll also post my measurements.

Yeah, laminations. Transformer companies tend to use really cheap silicon steel and just skim the edge of saturation. If you get in trouble on saturation, they can usually go to some better material and get decent margins in the same physical size.

I did a few custom transformers that we use, and looked up some others in catalogs, just to get an idea of the slopes.

I still like to use graph paper. You just whack the little dots down... no typing numbers. There is no raw data, you just plot it.

John

See:

It's not a flamed out tantalum, but an example of what happens when a cheapo ATX power supply goes over-voltage. The paper mess was from two electrolytic capacitors. The charred PBC near the purple electrolytic is probably what inspired the over-voltage. Incidentally, literally every board and drive in the computer was also blown.

Incidentally, speaking of tantalums, have you run into any with sky high ESR? I have. See the description next to the photo of the two tantalum caps at:

My favorite capacitor experiment was (someone) dropping a metal bar across the terminals of an oil drum capacitor. I forgot the capacitance, but the voltage rating was at least 10Kv. A bank of these were used to vaporize a wire to fire off a big Hi-V switch. It took 1 or 2 days to charge the oil drum capacitor, and a few milliseconds to temporarily wreck my hearing, spray molten aluminum slag everywhere, and panic everyone in the building. Sorry, no photos of the devastation.

There are about 30 assorted videos on YouTube under "exploding capacitor" and "capacitor bomb".

All are electrolytics. No tantalums. Well, time to make an exploding tantalum video.

This has nothing to do with the discussion, but is kinda interesting. Exploding a potato with a defibrillator cazapitor:

Same thing with apples:

I've seen more than my share of those looking that bad or worse. In more than one instance we had to replace the entire set of caps on the pcb. I've had not only the teardrops, but multiple smt tantalums too, 50V parts being used as 5V bypass, kablowee! That was on a pcb in our $750,000 NMR machine.

I seem to be using Panasonic's FK series a lot, if for no other reason than that they are cheaply available from Farnell.

robert

C'mon, that's straight lines on a log-log plot. The equations can be left as an exercise to the reader.

robert

We did an analysis of field returns on our VME modules. We're besting Bellcore mtbf calculations by about 3:1, and half the failures were tantalum caps.

We rarely use them any more. That's too bad, because they are like Saabs: great gadgets, between fires.

John

Gotcha, Y= mX+b Thanks, Harry

If my assertion is correct, it would be a good side note. However, it's getting harder to find axial electrolytics these days so it might not be necessary. United Chemi-Con and Panasonic don't make them anymore. Digikey only carries BC Components for axial Al electrolytics. In the late 80s, I was looking for electrolytics for high current pulse applications. The esr difference between radial and axial was quite visible.

Forgot to add my favorite cap into the fray. Although not jellybean, they aren't much more expensive. United Chemi-Con KZE series. DigiKey and Mouser stocks these parts. These are low impedance 105 deg parts. I use these for high current pulse applications and replacing bum capacitors in modern electonics with switching supplies that die in a year or two.

We have zillions of these at

Not necessarily in order: Xicon, Panasonic, Nichicon.

The link goes to your eBay seller page, now showing 368 items, was there a specific item?

I saw your 330uF 200V snap-in caps. As it happens a recent design of mine calls for 330uF 250V caps, so that got my attention. But I like the extra 80-volt safety margin of a 250V cap over my design's nominal 170V working voltage. I selected a Chemi-Con KMH type. It's 1" dia x 1.2" high, much shorter than your 1" dia x 2" high 200V version. It appears modern caps are better than old ones.

Xicon is sold only by Mouser, and a couple of other outfits that buy stuff from Mouser and resell it. It's their "house brand" and applies to whatever Taiwanese or (more lately) Chinese stuff comes over in the shipping container. House brands are of dubious utility in textbook tables even though they might be good deals on good stuff.

Tim.

What usually gets past purchasing for through-hole parts: Panasonic KS, Panasonic M, Nichicon VR

These are all good basic parts but not the best in terms of ESR or capacitance per cubic inch.

Surface-mount involves radically different dimensions and form factors in some cases.

In making a textbook table, it would be worthwhile to show a scatter plot of guaranteed ESR vs dollars spent, all for the same capacity/ volt capacitor (e.g. 1000uF@16V). Did I just volunteer?

Tim.

I've got lots of aftermath photos. I guess what Win would really like is a photo of one "caught in the act". IOW about one millisecond after lift-off, orange flames spewing out the bottom.