UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Components and Circuits (https://www.vintage-radio.net/forum/forumdisplay.php?f=40)
-   -   Electrolytic capacitor fails after 1 hour in service! (https://www.vintage-radio.net/forum/showthread.php?t=141831)

Al (astral highway) 1st Dec 2017 4:31 pm

Electrolytic capacitor fails after 1 hour in service!
 
Hi folks,

I don't want this to turn into a 'aren't components poor quality compared to the good old days' rant, but...

I feel nevertheless surprised that a capacitor from what I thought was a reputable manufacturer (Mods am I allowed to say, or is this defamatory, albeit with a defence of justification??)failed in less than an hour in service.

It was a 63V 1,000uF rated item used as a decoupling capacitor straight across the Vcc-ground pins of a a gate driver chip, in parallel with a 100nF poly component. The supply voltage was just 18V.

I know this is a 'noisy', environment and the component is under great stress, but I was surprised. It failed catastrophically for no apparent reason. Fortunately I had S/c protected the power supply but it could have been very damaging.

On inspection of the item when I had isolated it from the circuit, it performed normally up to 12V and then became progressively leaky, until it was passing 4A at a supply voltage of 30V.

I don't know what I'm expecting by posting this, that isn't encouraging complaints and sharing similar stories, but it's so surprising to me that I want to put it out there.

jjl 1st Dec 2017 4:37 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Al

Do you know the ripple current rating of your failed capacitor? In this sort of application it is possible that you exceeded the specified rating.

John

Nuvistor 1st Dec 2017 4:42 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Couple of thoughts.
1. Is it counterfeit?
2. Is it worth contacting supplier/ manufacturer for their response? A reputable company usually want to know about early failures.

Appreciate neither of the the above may be possible.

3. Has the replacement worked ok?

Al (astral highway) 1st Dec 2017 4:43 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by jjl (Post 995975)
Al

Do you know the ripple current rating of your failed capacitor? In this sort of application it is possible that you exceeded the specified rating.

Good point, John. I didn't look that up. I did take every precaution - very short leads, radial component mounted right across the pins of the device to reduce parasitic inductance...

But the active device can switch in 15nS, and was supplying 2A pulses to a gate, so it is a stressful role.

Nuvistor 1st Dec 2017 4:51 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Just checked the Vishay spec and they rate ripple current at 100Hz, at 10khz the multiplier is 1.3.
http://www.vishay.com/docs/28371/101102phrst.pdf

Boom 1st Dec 2017 4:53 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Some years back I bought a hi end hi-fi and it immediately blew the tweeters when turned on. Replacements went the same way. The woofers stayed fine.

After much head scratching I traced the trouble to the Dolby decode chip. It was putting out an almighty 10kHz (or so) tone.

I'd be tempted to put a 'scope on the cap' to make sure there is nothing there that shouldn't be.

David G4EBT 1st Dec 2017 5:11 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
I think that rippe current is rarely considered and rarely given by makers, which is why I tend to stick to Rubycon electrolytics from Farnel, because they're 105C and have high ripple ratings despite their compact physical size. Furthermore, comprehensive data is available so you aren't buying a 'pig in a poke' where he only factors you know about are the capacitance and working voyage. You can check whether they meet the desired ripple rating spec. EG:

http://www.farnell.com/datasheets/23...377.1373580746

I use the compact high voltage ones for stuffing smoothing/reservoir caps, as I can check that despite their diminutive size, the ripple rating is well up to the task with a good margin of safety.

Hope that's of interest.

Al (astral highway) 1st Dec 2017 5:12 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
It’s a decoupling cap sitting right on a chip power switching at 320KHz. Perhaps it is indeed greatly derated in this application.

Argus25 1st Dec 2017 5:13 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by astral highway (Post 995972)
On inspection of the item when I had isolated it from the circuit, it performed normally up to 12V and then became progressively leaky, until it was passing 4A at a supply voltage of 30V.

Al,

This sort of thing doesn't surprise me as after seeing electrolytic caps fail in multiple ways so many times. In theory at least most of the high frequency ripple should have been bypassed by the 100nF. I guess for the replacement you could put a 1uF poly across it might help a little.

It is almost acting like a lower voltage part where the leakage is ok up to a certain voltage and suddenly goes up over that. It is possibly a miss labelled part or very bad quality clone part.

I'd also be cautious applying 30v x 4A or 120W to the the capacitor for more than a second or two, it will heat up pronto and explode its rubber seal and spray its guts all over the room.

A 63V rated cap seems like a very safe margin for an 18V supply, probably not much point in going to a 160V one, sometimes the higher voltage ones have a slightly higher ESR which is not helpful. And as noted a high ripple current rated one would help.

What does it read on the ESR meter ?

Al (astral highway) 1st Dec 2017 5:38 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by Argus25 (Post 995988)
This sort of thing doesn't surprise me as after seeing electrolytic caps fail in multiple ways so many times. In theory at least most of the high frequency ripple should have been bypassed by the 100nF.

Hey Hugo, it is somewhat reassuring that this is familiar territory to you in relation to electrolytics!

Quote:

Originally Posted by Argus25 (Post 995988)
I guess for the replacement you could put a 1uF poly across it might help a little.

Ahah, good thinking, thank you.

Quote:

Originally Posted by Argus25 (Post 995988)
I'd also be cautious applying 30v x 4A or 120W to the the capacitor for more than a second or two, it will heat up pronto and explode its rubber seal and spray its guts all over the room.

8-o Exactly! I was just sweeping the voltage up steadily from my power supply and it became leakier and leakier. It's a one-shot experiment and I took the precaution of covering it with a ceramic cup (as in my empty tea mug). I don't have an ESR meter, so I rely for these things on the current being displayed on my DC power supply.

I do suspect it's a clone. It's a very high rep name on there, and I don't even want to mention it.

I feel pleased with my instinct to suspect the cap. rather than the driver chip or the MOSFET it was driving. It would have been expensive to bin the chip, if I'd jumped to the conclusion in error that it had developed a fault condition...

kalee20 1st Dec 2017 7:47 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
This is surprising to me! Even if you exceed the ripple rating, I would expect a lifetime of more than an hour - and the failure mode to be gradual drying out and rising ESR.

Just a thought - it was the right way round wasn't it? 18V is about what I'd expect to gradually destroy an electrolytic, climaxing in an hour or so.

HamishBoxer 1st Dec 2017 8:33 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Talking of caps ,remind me where can I find suitable 4.7uF 500Volts for stuffing smoothing caps?

Thank you

MrBungle 1st Dec 2017 8:39 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Out of interest see there any negative transients or positive spikes? Each of those can knacker even the best of electrolytic capacitors in no time at all. Sticking an X2 rated cap across it as well might help.

Al (astral highway) 1st Dec 2017 8:47 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by kalee20 (Post 996022)
Just a thought - it was the right way round wasn't it? 18V is about what I'd expect to gradually destroy an electrolytic, climaxing in an hour or so.

I did place it with the correct polarity.

Nuvistor 1st Dec 2017 9:01 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by HamishBoxer (Post 996033)
Talking of caps ,remind me where can I find suitable 4.7uF 500Volts for stuffing smoothing caps?

The smallest I could find was 10uf 500vw, is that too big?
https://uk.rs-online.com/web/p/alumi...itors/7683911/

Al (astral highway) 1st Dec 2017 9:06 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by David G4EBT (Post 995986)
I think that rippe current is rarely considered and rarely given by makers, which is why I tend to stick to Rubycon electrolytics from Farnel, because they're 105C and have high ripple ratings despite their compact physical size.


That's useful info, David. Thanks.

On closer inspection, the chip switches on even faster than I thought, in less than 10nS, pushing at least 2A pulses, in my case at around 300KHz. It would be unsurprising if even 1cm of lead length on the cap leads was a parasitic inductor. But sourcing a suitably robust replacement seems like a wise plan.

I might put a bi-directional TVS diode across the supply rail.

HamishBoxer 1st Dec 2017 9:27 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Thank you Frank,they are for a Murphy A40c.

Radio Wrangler 1st Dec 2017 9:53 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
If you were going to fake a capacitor, you would pick a premium brand to mark on it because it would sell for more yet cost no more to make.

If you were going to fake a capacitor, you would consider putting a cheapie of the right capacitance inside, or re-marking a cheapie so that on a quick check with a capacitance meter it would show a non-suspicious value and so would get passed on further along the supply chain and wouldn't get spotted until it was too late to do anything about it. A lower voltage rating would be unlikely to be noticed until some time after the part went into its final application.

A cheapie would be cheap because of a low quality maker, and also because of a lower voltage rating I'd expect it to be somewhat smaller.

What was inside the one which blew?

Where did you get it from?

You didn't happen to buy a few at the same time, did you?

If the capacitor wasn't a fake, but was a rogue which slipped through the quality control system, then a reputable maker would be interested in diagnosing the failure.

If the capacitor was a fake, then a reputable manufacturer would be interested in inspecting it to see if there are any clues as to its source, and also for taking action against whoever supplied it.

I don't suppose it matters whose name was on the thing.

My designs use a great many 63v electrolytic capacitors, with many tens of thousands out in the field running at 40-50v and large, fast current pulses in applications where failure would certainly be noticed and reported back. So far I've seen no failures, but the supply chain is very carefully managed. The maker is not a well-known brand so maybe the risk of faking is lower.

David

kalee20 1st Dec 2017 10:05 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by MrBungle (Post 996035)
Out of interest see there any negative transients or positive spikes? Each of those can knacker even the best of electrolytic capacitors in no time at all.

If there were any negative transients, then bearing in mind the capacitor was right on the driver chip, I'd expect the chip itself to have either died, or clamped the negative transient to a volt or so.

Nuvistor and Argus have a point, is it a clone or counterfeit? Or maybe it's just a single bum capacitor?

I've used beefy gate driver chips to drive chunky power MOSFETS, I can't see anything wrong in what you have done. A 1,000μF 63V capacitor should be well up to the job, both in terms of having a low ESR to give a stiff supply to the chip, and also being able to cope with the ripple.

A quick calculation says that if you are taking 2A gulps of current, lasting 15nsec, at a repetition rate of 300kHz (so 3.3μsec apart), you have an RMS value of 135mA. And this shouldn't fry a capacitor of this sort of size.

Maarten 1st Dec 2017 10:45 pm

Re: Electrolytic capacitor fails after 1 hour in service!
 
Quote:

Originally Posted by MrBungle (Post 996035)
Out of interest see there any negative transients or positive spikes? Each of those can knacker even the best of electrolytic capacitors in no time at all. Sticking an X2 rated cap across it as well might help.

An X2 capacitor may be severly underrated or otherwise unsuitable for decoupling noisy DC power supply lines. It was designed to satisfy a very specific set of conflicting requirements found in mains filters and the like.


All times are GMT +1. The time now is 2:04 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.