Post by thread:[OT] Short life on electrolytic caps

At 02:28 PM 3/8/00 +0200, you wrote: >Hello Everyone, > >I have just (with somewhat of a shock) found out that electrolytic caps are >rated for only 2000 to 7000 hours of operation. These ratings seem to be at >high temperatures, but nonetheless they make me feel very uncomfortable >about using them. Is this some unspoken rule that somehow passed me by? That's usually 4% failures at that life. Usually consider life to double for every 10'C lower. >I would like to hear some comments on what caps are usually used in >differenent situations.. That could fill a book, or at least a chapter.. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Spehro Pefhany "The Journey is the reward" spam_OUTspeffTakeThisOuTinterlog.com Fax:(905) 271-9838 (small micro system devt hw/sw + mfg) =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Of course, an electrolytic capacitor stored at the drawer can survive longer than 7000 hours. When in operation is that it suffers all possible torture. A big torture is temperature, so, strategies are in order to locate the capacitor away from heat sinks, transformers, and so on. If your power supply has a blower, make sure the caps are in direct hit with the air flow *BEFORE* the air gets warm from heat sinks and other hot components. Imagine two 10,000 µF capacitors rated for 16V, one is working with 15V, another with only 3V, which one do you think is receiving more torture and could get early problems? The one with 15V for sure. Using electronic components at marginal values is only good for fuses, and this is an interesting point. It is common to find power supplies that consume 1 Amp at the output with a protection fuse of 1.5 or even 2A, since nobody wants that fuse to blow without any reason, isn't it? But then, in the same power supply you can find a 10V capacitor filtering 9Vdc... or even a LM317 regulator supplying 1.4A, it means, everything is wrong, right? To work at 9Vdc, the capacitor should stand minimum 16V and the regulator should stand at least twice the output current. Chances are that installing a 50Vdc capacitor to filter 9Vdc output will make this capacitor lasting longer than the regulator, the power transformer, the wires, the metal case, and so on. According to your numbers, 7000 hours is barely 10 months of continuous operation. I have some electronic devices around that are powered continuously for more than 10 years... it is more than 86 thousand hours... Wagner Roland Andrag wrote: {Quote hidden}> > Hello Everyone, > > I have just (with somewhat of a shock) found out that electrolytic caps are > rated for only 2000 to 7000 hours of operation. These ratings seem to be at > high temperatures, but nonetheless they make me feel very uncomfortable > about using them. Is this some unspoken rule that somehow passed me by? > > I would like to hear some comments on what caps are usually used in > differenent situations.. > > Thanks > Roland

>Imagine two 10,000 5F capacitors rated for 16V, one is working with 15V, >another with only 3V, which one do you think is receiving more torture >and could get early problems? The one with 15V for sure. Not necessarily. It is the heat generated by the ripple current through the effective series resistance which is the main killer. This heat causes the electrolyte to "boil off" and evaporate. there is a disadvantage to the situation in the quote above. If a capacitor is used at too low a voltage below its rated voltage, the electrolyte fails to maintain the oxide film on the surface of the metal plates, and the effective leakage resistance drops alarmingly, making the device look more like a resistor than a capacitor. How low is too low below rated voltage? This can be a wet finger in the air thing to guestimate, but it often is not a big problem, despite all I have just said. Using the wrong capacitor as a filter capacitor in a power supply is most likely going to be the cause of failure of a capacitor in a system, unless it fails on the early slope of the bathtub curve.

I should agree with you, but unless practical experiences show me different results, I always would choose a 10000uF @ 50V capacitor instead a 10000uF @ 5V for a 5V application, at least I would have a considerable lower ESR, lower resistance, lower generated heat. I already saw few large caps exploding (protective caps and all), it is never a nice picture. I never thought about the reducing oxide film effect with lower voltages, it also means that the oxide is reduced (or even doesn't exist) when the equipment is powered off for quite some time? I really don't know much about it. Wagner Alan Pearce wrote: {Quote hidden}> > >Imagine two 10,000 uF capacitors rated for 16V, one is working with 15V, > >another with only 3V, which one do you think is receiving more torture > >and could get early problems? The one with 15V for sure. > > Not necessarily. It is the heat generated by the ripple current through the > effective series resistance which is the main killer. This heat causes the > electrolyte to "boil off" and evaporate. > > there is a disadvantage to the situation in the quote above. If a capacitor is > used at too low a voltage below its rated voltage, the electrolyte fails to > maintain the oxide film on the surface of the metal plates, and the effective > leakage resistance drops alarmingly, making the device look more like a resistor > than a capacitor. > > How low is too low below rated voltage? This can be a wet finger in the air > thing to guestimate, but it often is not a big problem, despite all I have just > said. Using the wrong capacitor as a filter capacitor in a power supply is most > likely going to be the cause of failure of a capacitor in a system, unless it > fails on the early slope of the bathtub curve.

<x-flowed>Back in the B+ days of tubes and such,,, if a cap was really old,,, you would put a big value resistor in series with it and 'Reform' it,,, maby over night,,, or over a weekend before putting it back in service after a long storage time. My gut level reaction,,, the oxide film at 5 volts would work fine at five volts,,, but over time,,, the device may not like going back to 50 volts all at once, without a current limited reforming time. At 12:09 PM 3/8/00 -0500, you wrote: > I never thought about the reducing oxide film >effect with lower voltages, it also means that the oxide is reduced (or >even doesn't exist) when the equipment is powered off for quite some >time? I really don't know much about i 'Grif' N7IVS </x-flowed>

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 02:28 PM 3/8/00 +0200, Roland Andrag wrote: >Hello Everyone, > >I have just (with somewhat of a shock) found out that electrolytic caps are >rated for only 2000 to 7000 hours of operation. These ratings seem to be at >high temperatures, but nonetheless they make me feel very uncomfortable >about using them. Is this some unspoken rule that somehow passed me by? Don't let it scare you TOO much. Jim Williams has a nice formula for this, that is darned accurate, but in general terms, every 10 degrees C you take off the cap, doubles it's life. A typical garbage electrolytic is 2000 hours at 85C, so roughly 4000 hours at 75, 8000 at 65, 16000 at 55, 32000 at 45, 64000 at 35 (where I generally figure my designs will operate) This is correlated by my repair experience were most failures are traceable to failed electrolytics, especially in anything that runs hot. Other parts fail too, but ask anyone who repairs computer monitors or terminals, or power supplies, or TVs (all use electrolytics at elevated temperatures, and moderate to high currents) -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMbC+YFlGDz1l6VWEQIYNQCfYpN5FYhpnD8Vvw6TTJMvrM1QWVMAoP6X GMaApcmr4swRu4EBJ6CbITKP =J8FT -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 12:09 PM 3/8/00 -0500, Wagner Lipnharski wrote: \ I never thought about the reducing oxide film >effect with lower voltages, it also means that the oxide is reduced (or >even doesn't exist) when the equipment is powered off for quite some >time? I really don't know much about it. >Wagner True, though you usually see it only on very old caps. It may be a situation where ALL will eventually show the problem, but some show it sooner. In repairing old radios (paper cap days) you want to apply voltage very slowly, and let the caps re-form if they will, to preserve the original parts. I do this with all high voltage caps I use, with a current limited 0-400V bench supply. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMbE04FlGDz1l6VWEQLL6QCfTOLg1CdNJ8a3R7lYR+alqUmMohwAoIDY HeM1f75A97fFykAp+CHqVyP8 =xSk/ -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 08:05 AM 3/9/00 +1100, Paul B. Webster VK2BZC wrote: >David VanHorn wrote: > >> running an electrolytic at voltages very much lower than it's rating >> can cause the forming process to somewhat reverse itself, resulting in >> a lower effective capacitance value. > > Lower? That's what they told me. "They go to sleep, and stop being capacitors". -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMbsj4FlGDz1l6VWEQK0LgCg/31FSLA8PebJHXYs0j1rmhf5REkAn1Ja 5a1ynn8pBGR3KDwyUT30wXlK =kMpb -----END PGP SIGNATURE-----

Lots of opinions on this I see. Here's what I understand (*but then, I may be wrong :-) ) I base my design decisions on this information :-) 1. Electrolytic capacitors are lifetime rated at their maximum design temperature. As noted by others, lifetimes will approximately double for every 10 degrees Celsius / kelvins drop below this temperature. eg Lifetime = (Manufacturers stated lifetime) x 2^( (Trated - Tactual) / 10 ) eg 85 Celsius capacitor 35 Celsius working 2,000 hour rated Actual lifetime = 2000 x 2^(85-35)/10) = 64,000 hours = 8 years In practice it's not exactly double per 10 degrees C and manufacturers provide a graph of lifetime versus temperature. 2. Electrolytics operated at high temperature will die MUCH FASTER if the power is OFF !!!! The electrolyte dries out. Manufacturers provide a rating figure on this too. AFAIR an 85C cap will die in about 10% of its lifetime if run at 85C with power off !!!!!!!!!!!!! 3. Electrolytic capacitors are best run at voltages below but near their rated voltage (rather than well below their rated voltage) , all things being equal. It may be that other factors such as ESR, ripple current rating etc lead you to other conditions but if these are adequate then it is better to use a 10V cap for a 5V supply than a 50V cap for a 5 V supply. General: - Self heating due to eg ripple current must be taken into account. - Ripple current specs must be observed to achieve rated lifetimes. - Avoid polarity reversal - Most cheap ecaps are rated at 85C but remember that 105C rated caps exist. These are dearer but may make a vast difference to lifetimes in high ripple and/or high temperature applications. regards Russell McMahon _____________________________ >From other worlds - http://www.easttimor.com http://www.sudan.com What can one man* do? Help the hungry at no cost to yourself! at http://www.thehungersite.com/ (* - or woman, child or internet enabled intelligent entity :-)) {Original Message removed}

Interesting thread. I recall once reading that aluminum electrolytic caps have, to a certain extent, the ability to resurrect themselves after they have been hit with large "overvoltages", like ESD - eg, 5KV. Supposedly, the aluminum oxide layer can rejuvenate, even after some damage. Apparently doesn't work for tantalums, which don't have the oxide layer. Any additional experience/enlightenment along this line? Like how much damage can they take, etc? - Dan Michaels Oricom Technologies http://www.sni.net/~oricom ========================== > >{Original Message removed}

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 10:46 PM 3/8/00 -0700, Glenn Golden wrote: >In a nut shell > >1.) Heat kills. Current kills too, (because of heat) The equation by Jim Williams takes into account the part ratings, Temperature, ESR, current, (and a few more terms, but I forget) Look in the linear tech apps for it, he and Bob Pease have a special place in my bookshelf, next to AofE. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMdqroFlGDz1l6VWEQLmjwCfVzcNOtaf+dsFEkJcMO5RRZ/Q9KUAoPTj QW8/DI1yY7FD0UzwmHZbB5De =x3PQ -----END PGP SIGNATURE-----

In message <.....4.0.1.20000308131038.00fa6ca0KILLspam@spam@mail.cedar.net>, David VanHorn <dvanhornKILLspamCEDAR.NET> writes >Jim Williams has a nice formula for this, that is darned accurate, but in >general terms, every 10 degrees C you take off the cap, doubles it's life. >A typical garbage electrolytic is 2000 hours at 85C, so roughly 4000 hours >at 75, 8000 at 65, 16000 at 55, 32000 at 45, 64000 at 35 (where I generally >figure my designs will operate) > >This is correlated by my repair experience were most failures are traceable >to failed electrolytics, especially in anything that runs hot. Other parts >fail too, but ask anyone who repairs computer monitors or terminals, or >power supplies, or TVs (all use electrolytics at elevated temperatures, and >moderate to high currents) The biggest capacitor failure is in satellite receivers, most have now switched to 105 degree versions, and we only keep 105 degree ones as replacements for use in everything we repair!. -- Nigel. /--------------------------------------------------------------\ | Nigel Goodwin | Internet : .....nigelgKILLspam.....lpilsley.demon.co.uk | | Lower Pilsley | Web Page : http://www.lpilsley.demon.co.uk | | Chesterfield | Official site for Shin Ki and New Spirit | | England | Ju Jitsu | \--------------------------------------------------------------/

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 >The biggest capacitor failure is in satellite receivers, most have now >switched to 105 degree versions, and we only keep 105 degree ones as >replacements for use in everything we repair!. I use only 105 degree units in new designs too, the price difference isn't that large. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMd1A4FlGDz1l6VWEQKCewCgzpL8TPgClMdzEse5h3495xV9x9gAoLf/ jjFvMA1CRUTnZJHlgFu6AFSN =7Dq1 -----END PGP SIGNATURE-----

Dan, this is a good thread. I had to revisit the subject when my old Mag Innovision 17" monitor died a couple of years ago. It turned out that electrolytic caps spec'd at 85C died taking out several components from the H.O.T. to the flyback transformer... Talking with Mag confirmed the problem and I got the replacement parts with 105C caps... The "rehealing" feature with electrolytics relates to the working electrolyte in that any faults in the oxide layer will be repaired by further anodization. This is a factor is the design of the cap. "Self-Healing" refers to metallized caps where a partial discharge results in a localized failure of the dielectric which burns away the metallized electrode effectively isolating the fault. Someone mentioned "reforming" electrolytics. I had not considered this to be a problem but, after `blowing the dust off' ITT's "Reference Data for Radio Engineers", I ran across the following: "Conventional aluminum electrolytic capacitors which have gone 6 months or more without voltage applied may have to be reformed." A DC source at the rated voltage with a 1.5K internal resistance for caps with a rated voltage exceeding 100V or 150 ohms for <=100V must be applied for one hour after the cap reaches it's rated voltage +/- 3%. The cap is then discharged through a resistor of 1ohm/V. I would first check with the cap vendor to see what the shelf-life is in this `day and age', and what, if any, `reforming' procedure is required. Still, it's something to think about if you have a bunch of `ancient' electrolytics laying around and you are experimenting for home use. - Tom At 07:12 PM 3/8/00 -0700, Dan Michaels wrote: {Quote hidden}>Interesting thread. I recall once reading that aluminum >electrolytic caps have, to a certain extent, the ability >to resurrect themselves after they have been hit with >large "overvoltages", like ESD - eg, 5KV. Supposedly, >the aluminum oxide layer can rejuvenate, even after some >damage. Apparently doesn't work for tantalums, which don't >have the oxide layer. > >Any additional experience/enlightenment along this line? >Like how much damage can they take, etc? > >- Dan Michaels >Oricom Technologies >http://www.sni.net/~oricom ------------------------------------------------------------------------ Tom Handley New Age Communications Since '75 before "New Age" and no one around here is waiting for UFOs ;-)

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 08:23 AM 3/9/00 +0000, Graham North wrote: >I have a satellite receiver which does not work, I have tried but can not >find the problem. > >How can I tell if it's the caps? Your first clue would be if any have exploded, bulged, or ruptured their tops. After that, you'll likely need a schematic, and a meter at least, to track down the problem -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMf+OoFlGDz1l6VWEQIhdQCfcMaTAAv2MyAhrO6LpL0s1+d+iW0An3jo KcqrnFMD1CdQoB1mP3toyB/7 =ZQAg -----END PGP SIGNATURE-----

Tom wrote: > The "rehealing" feature with electrolytics relates to the working >electrolyte in that any faults in the oxide layer will be repaired by >further anodization. This is a factor is the design of the cap. > > "Self-Healing" refers to metallized caps where a partial discharge >results in a localized failure of the dielectric which burns away the >metallized electrode effectively isolating the fault. > So, it is reasonable to think that a cap *blasted by an ESD bolt* can possibly "reheal" by reforming its oxide layer, or may "self-heal" with the charge blasting the electrode away in a local area, but leaving intact the rest of the cap surface. (I'm not sure if your reference to "s-h" & metallized caps applies to aluminum electrolytics). I got onto this a few years ago, when consulting with a company that was using 6805 embedded controllers in industrial icemaker machines. A really bad ESD, EMI, and inductive switching transient environment. The controllers were constantly latching up (the orig design had NO protection on the sense lines at all - would you believe!). The contractor redesigned the controllers by adding pi-filters to the sense lines, consisting of 10uF/25v alum electrolytic right on the "external" pin, then 10K resistor, and finally .1 uF bypass at the uC pin. We did some inhouse ESD tests using a 5KV Zap gun, and I was amazed to discover that the little 25v electrolytics weren't completely blown off the boards. Leakage increased into the caps temporarily, but this seemed to diminish after a while. Thus, I got onto the healing bit. I also discovered that the small lead spacing on the caps acted as a spark gap, but still the caps must have been getting blasted pretty good. All in all, pretty amazing. Thanks, - Dan Michaels Oricom Technologies http://www.sni.net/~oricom

At 10:45 AM 3/9/00 -0700, Dan Michaels wrote: >So, it is reasonable to think that a cap *blasted by an ESD bolt* >can possibly "reheal" by reforming its oxide layer, or may >"self-heal" with the charge blasting the electrode away in a local >area, but leaving intact the rest of the cap surface. (I'm not sure >if your reference to "s-h" & metallized caps applies to aluminum >electrolytics). Dan, "Self-Healing" only applies to metallized caps (with metallized electrodes), not electrolytics. I brought that up to prevent confusion with the "rehealing" property of electrolytics. - Tom >I wrote: >> The "rehealing" feature with electrolytics relates to the working >>electrolyte in that any faults in the oxide layer will be repaired by >>further anodization. This is a factor is the design of the cap. >> >> "Self-Healing" refers to metallized caps where a partial discharge >>results in a localized failure of the dielectric which burns away the >>metallized electrode effectively isolating the fault. ------------------------------------------------------------------------ Tom Handley New Age Communications Since '75 before "New Age" and no one around here is waiting for UFOs ;-)

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 08:48 AM 3/10/00 +1100, John Orhan wrote: >Hi all, >I am searching for a good hi fi power output stage circuit around the 50 - >100watt mark having foolishely volenteered to help a good friend of mine >with his audiophilia problem . Can anyone suggest a good site for such >things or a known good circuit?? This will save months of angst. There is no known cure for audiophilia. Unfortunately, they tend to ignore sound engineering, and go with the "it's more expensive, so it must be better" approach.. :-P I lost the link, but I had a site selling a 3 Meter speaker cable for $7000.00 For my money, an LM-12 is pretty hard to beat, and you can connect it as a voltage to current converter, and drive the speakers that way, eliminating the effects of the speaker wire. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMhJDIFlGDz1l6VWEQKIBQCgxDSH85W67qsB8DTNApNZmq9n33kAnigB P5sEFmlAh9j9gfNVvbVvSNTD =0oMq -----END PGP SIGNATURE-----

-----Original Message----- From: David VanHorn [SMTP:EraseMEdvanhornspam_OUTTakeThisOuTCEDAR.NET] Sent: Friday, 10 March 2000 6:41 To: PICLISTspam_OUTMITVMA.MIT.EDU Subject: Re: [OT] Short life on electrolytic caps -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 At 08:23 AM 3/9/00 +0000, Graham North wrote: >I have a satellite receiver which does not work, I have tried but can not >find the problem. > >How can I tell if it's the caps? Your first clue would be if any have exploded, bulged, or ruptured their tops. After that, you'll likely need a schematic, and a meter at least, to track down the problem You forgot colour change - quiet or failed Sanyo TV sound was always easy to fix, you just looked for the grey cap in amoungest the pink ones. The dried out caps heated up and changed the wrapper colour as they failed. Bye.

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 >You forgot colour change - quiet or failed Sanyo TV sound was always easy to >fix, you just looked for the grey cap in amoungest the pink ones. The dried >out caps heated up and changed the wrapper colour as they failed. Nice of them to do that, but it's mostly peculiar to that brand. The ones I've seen don't change color until they are showing many other symptoms. (like fragmentation) -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com> iQA/AwUBOMhaSYFlGDz1l6VWEQJRrwCglY+AHjeaTnhUiwmliKWeXtd+xRUAnRkc 3jQQBu3Fym+y9y0dKTKd7Fli =ODPW -----END PGP SIGNATURE-----

Cap Meter or ESR meter will help find failed/ marginal caps. There is even a PIC project to build a cap meter, just don't have the address handy. Kev HKC Inc > > > >How can I tell if it's the caps? > > > Your first clue would be if any have exploded, bulged, or ruptured > their > tops. > After that, you'll likely need a schematic, and a meter at least, to > track > down the problem > > You forgot colour change - quiet or failed Sanyo TV sound was always easy to > fix, you just looked for the grey cap in amoungest the pink ones. The dried > out caps heated up and changed the wrapper colour as they failed. > > Bye. >

http://redrival.com/mcgahee/ That addy will get you to teh projects page Joe ----- Original Message ----- From: "Kev" <@spam@KKloppKILLspamEROLS.COM> To: <KILLspamPICLISTKILLspamMITVMA.MIT.EDU> Sent: Thursday, March 09, 2000 4:12 PM Subject: Re: [OT] Short life on electrolytic caps > Cap Meter or ESR meter will help find failed/ marginal caps. There is even > a PIC project to build a cap meter, just don't have the address handy. > > Kev > HKC Inc > > > > > > >How can I tell if it's the caps? > > > > > > Your first clue would be if any have exploded, bulged, or ruptured > > their > > tops. > > After that, you'll likely need a schematic, and a meter at least, > to > > track > > down the problem > > > > You forgot colour change - quiet or failed Sanyo TV sound was always easy > to > > fix, you just looked for the grey cap in amoungest the pink ones. The > dried > > out caps heated up and changed the wrapper colour as they failed. > > > > Bye. > >

In message <6621F2E12035D311B02208002BC33C593B1BB3@LANDUK1>, Graham North <RemoveMEgraham.northTakeThisOuTLANDINST.COM> writes >I have a satellite receiver which does not work, I have tried but can not >find the problem. > >How can I tell if it's the caps? Stick a scope on the PSU outputs, and measure the ripple - it should be fairly low on all the outputs. Obviously 1 volt ripple on a 30 volt rail isn't too bad, but 1 volt ripple on a 5 volt rail is much too high, it's normally pretty obvious which ones are duff. You also get problems with coupling capacitors drying out as well, you really need a circuit to follow the signal path to find those. -- Nigel. /--------------------------------------------------------------\ | Nigel Goodwin | Internet : spamBeGonenigelgspamBeGonelpilsley.demon.co.uk | | Lower Pilsley | Web Page : http://www.lpilsley.demon.co.uk | | Chesterfield | Official site for Shin Ki and New Spirit | | England | Ju Jitsu | \--------------------------------------------------------------/