Qualitative analysis of Capacitors?

[eccerr0r]

Anyone actually have gear to do qualitative analysis on capacitors?

I mean, have a bake oven and a high frequency high current "zapper" that will torture test capacitors until they fail?

Granted capacitors tend to vary from lot to lot, and even unit to unit, but perhaps some of the "bad" cap manufacturers got their act together and make some semi-decent caps now?

[momaka]

Re: Qualitative analysis of Capacitors?

My friend baked the caps on quite a few Xbox 360 boards while he was setting his reflowing machine. Those 360 boards have good caps on them, though (Rubycon, Panasonic, Nichicon, etc.) so the results weren't all that spectacular.

What I noticed is that some caps will bulge right after being overheated while others would take a week or two to bulge after they were overheated. And what's even more interesting is that most of those caps that bulged right after being overheated, I could push their tops back down, meaning there was only temporary pressure buildup. Checking them with a capacitance and ESR meter after they bulged and a few weeks after pushing their tops down didn't reveal any significant changes in their properties. In fact, I have used a few such caps to recap a motherboard of mine. They tested fine for capacitance and ESR for over a month after I pushed their tops down, so they are likely going to be fine.

However, for the caps that bulged a few weeks after they were overheated - on those I usually couldn't push the tops back down, meaning some of the electrolyte must have broken down and there is a permanent gas buildup in the cap which will lead to eventual failure. On those, I noticed the ESR was slightly degraded.

I don't know if the above is a unique case or not, since I only tested that with about 20 caps. But it seems like if there is no permanent pressure build up in the cap and if the cap didn't leak any electrolyte, then it might work just fine.

[mockingbird]

Re: Qualitative analysis of Capacitors?

There are some interesting posts by Pete in CA regarding this in this thread:

At my, now, previous employer I compared HD with HE for a very different usage than what folks here do. I was not impressed. Then again, I was subjecting the things to 7X-10X (not a typo!) the rated ripple current, not anything like P/S or VRM usage. Generally, Nichicon PW, PA & HE, Panasonic FC & FM, Rubycon ZL, and UCC LXZ and KZE series did best in my torture test. That was over 3 years ago, BTW.

[PeteS in CA]

Re: Qualitative analysis of Capacitors?

Very briefly, I had a tester that used an LM317 to create an adjustable constant current (datasheet app) that then fed a square wave push-pull inverter. The transformer ratio stepped down the voltage and stepped up the current. The output side of the transformer was loaded by the cap being tested.

As I said in what mb quoted, the ripple currents were literally 7X-10X the parts' rating. The tests were open air, room temp. With a few of the smallest caps I tested (12.5mm diameter) the current fused the lead wires. Under that torture, the best parts often ran for hundreds or over a thousand hours before venting. Most of the deterioration in impedance happened toward the very end of the parts' lives.

[KD0CAC]

Re: Qualitative analysis of Capacitors?

I've used a Hipot tester to test large vacuum caps like Jennings , a couple of the Hipot testers have 3,000 AC , 4,000 or 5,000 AC/DC volts ,
http://www.surplussales.com/vaccumvarcaps/VVC6.html

[Paddan1000]

Re: Qualitative analysis of Capacitors?

I'm resurrecting this thread because I'm very interested in reading some reports where they compare caps of different brands and origins by having them endure standardized torture tests.
Most reports of bad capacitors seem to be anecdotal and usually involving 10+ years old capacitors made during the great capacitor plague.
If anyone know of any studies, please post a link.

[Paddan1000]

Re: Qualitative analysis of Capacitors?

I just desoldered a 10V 3300uF Teapo capacitor from a FSP PSU that I bought in 2006 and have abused daily for about 5-6 years in a badly ventilated computer way too power hungry for it's rated wattage. It showed no bulging or leakage. The capacitance was 3600uF (9% above rating) and the ESR somewhere between 0,05 and 0,00.
Clearly Teapo (and the Capxon that also were present in the PSU) can't be as bad as people say if they survived all those years in a budget PSU without losing their specifications.

[tron]

Re: Qualitative analysis of Capacitors?

There is no great capacitor plague,
there are quality caps, and
not so good quality caps
(Toepo, CapXon, ....not good)

plenty of bulging(Capxon)/
losing capacity(Teapo)
examples among them

[Rulycat]

Re: Qualitative analysis of Capacitors?

I find that with cheap capacitors like Teapo the main problem isn't that they never make them well... it's that they don't make them consistently to begin with.

[Heihachi_73]

Re: Qualitative analysis of Capacitors?

It really is just hit and miss with the big Taiwanese brands. Teapo, CapXon, OST, and even Fuhjyyu can have caps work perfectly for years on end without being out of spec, yet other seemingly identical caps can fail after a year or two. I'll still take a Taiwanese brand over a Chinese brand any day.

I am even thinking that the companies themselves are outsourcing or have separate factories - why do Teapo caps have completely different bungs for example? I've seen Teapo bungs with many rings within rings (bullseye), I've seen ones with a single circle in the middle, and I've seen others which are split in the middle and stick out like a Rubycon bung rather than having the cap sit flat against the PCB. All of these I have come across were earlier than 2005, so it predates the G-Luxon merger.

[momaka]

Re: Qualitative analysis of Capacitors?

I am even thinking that the companies themselves are outsourcing or have separate factories - why do Teapo caps have completely different bungs for example? I've seen Teapo bungs with many rings within rings (bullseye), I've seen ones with a single circle in the middle, and I've seen others which are split in the middle and stick out like a Rubycon bung rather than having the cap sit flat against the PCB. All of these I have come across were earlier than 2005, so it predates the G-Luxon merger.

Yeah, I've noticed that too with Teapo. Indeed makes you wonder if there are different factories that make them. I have a big bag of Teapo SC 6.3V 1000uF 8mm caps pulled from HP DC5000 motherboards. I tried to see if there is a pattern among the failed ones and which ones (as in, by which bung design) failed the most. Couldn't really find a pattern, though. There were failures from almost each kind. Indeed they are just inconsistent.

[Wester547]

Re: Qualitative analysis of Capacitors?

^ I strongly believe that G-Luxon gained large control over the Teapo's stock once they merged at the end of 2005/beginning of 2006. I have heard too many reports of Teapo failures -after- 2005 and not -before-... makes me wonder if 8mm Teapos and below are really G-Luxon and 10mm Teapos and above regular Teapos. But 1000uF, 6.3V Teapos have a rather low ripple rating and high ESR rating to begin with, especially if they measure 8x11 and not 8x15.

I just desoldered a 10V 3300uF Teapo capacitor from a FSP PSU that I bought in 2006 and have abused daily for about 5-6 years in a badly ventilated computer way too power hungry for it's rated wattage. It showed no bulging or leakage. The capacitance was 3600uF (9% above rating) and the ESR somewhere between 0,05 and 0,00.
Clearly Teapo (and the Capxon that also were present in the PSU) can't be as bad as people say if they survived all those years in a budget PSU without losing their specifications.

Depends on their date code. If they were manufactured in 2005 it could be predate the G-Luxon merger. However, 3300uF Teapos are rather overspec'd for the job to begin with.

[mockingbird]

Re: Qualitative analysis of Capacitors?

makes me wonder if 8mm Teapos and below are really G-Luxon and 10mm Teapos and above regular Teapos.

This.

[Heihachi_73]

Re: Qualitative analysis of Capacitors?

Doesn't Teapo still use the G-Luxon brand? They show both Teapo and 'Luxon' in their datasheets, so if anything, all G-Luxon caps this side of 2005 are Teapo anyway. Maybe it's time to open up a few and see what the innards look like, along with the electrolyte used.

[lti]

Re: Qualitative analysis of Capacitors?

I think the ones made by G-Luxon still use the X-shaped vent.

[eccerr0r]

Re: Qualitative analysis of Capacitors?

I wonder what conditions they use to test capacitors? When a company specs 5000 hours at 105C, how do they run the capacitors? I wonder if small (say, few tens of millivolts), frequent (1MHz) discharges is more stressful than large (say, near full rail) but less frequent (perhaps 10s of Hz) discharges?

This is still a good 7 months worth of testing which is a very long test... but still would be interesting.