capacitor "forming"

[Ograv]

I have heard so many "stories" about capacitors. One I heard recently is that you should never use a 100VDC cap in a 16VDC circuit because it will never "form" correctly or fully (whatever those terms mean). As I think about this one, I can see arguments on both sides.

True? Untrue? And why? Anybody know for sure?

Thanx.

[chrismh9999]

the 100vdc cap would have a higher esr value (series resistance) not so good with high frequency, and it would be a lot bigger. never herd those terms mentioned before

[woody]

I haven't read that recently but back 30 or 40 years ago that was ofton mentioned.
Now this is just a wild guess but caps seem to be a lot better than they used to be.
30 or 40 years ago the shelf life of an unused electrolitic cap was considered 3 years
I wonder if inprovements in materials and chemistry has extended this period before the oxide layer starts to brake down and need to be reformed.

[EBM_dude]

Well, I think it depends on the usage of the cap. A 100V cap is overkill for 16V, but apparently they seem to last longer when it comes to "dirty" PSUs (ripples and stuff). The temperature rating seems to be more important for that matter. I.e. a 105°C cap is less suceptible to failing when exposed to nasty ripples than a 85°C cap, even when they're never used at such elevated temps. The german magazine "c't" wrote something about that a while ago regarding blowing mainboard caps. Not only computer parts are affected, but even the trusty ECU in your car. I have noticed, that the caps in failing Mitsubishi and Hyundai ECUs are always rated at at least 50V and only 85°C. Must have had a reason they didn't use 16V caps. I replaced all those little guys with 105°C and 63V caps and never had any issues. I suggest you use the highest grade capacitors that are feasible.
If you want to "format" a 100V capacitor which will be used in a 16V circuit, then I suggest charging it to, say 60 or 90 Volts and slowly decharge it with a resistor. My father when he was still building stuff always told me I should "format" my caps, so I experimented a bit. The only difference I noticed was, that "formatted" caps seem to have a lower leak current for the first few power cycles. Then the unformatted caps have apparently the same leak current. This was a highly unscientific experiment and I don't want to be quoted on this! I think the whole "formatiing" issue is bollocks until someone shows me a doctor's thesis or scientific study.

[Conrad Hoffman]

IMO, a lot of things you might have heard about capacitors just don't apply to the caps being made over the last decade or two. My experience is that very low voltage caps have poorer performance and are to be avoided regardless of application. By low I mean under 16V rating. Go ahead and use a 50V cap in a 6V circuit- there's no life or performance downside. The only caps I've seen that may need to be re-formed (not formated) are high voltage caps that are a few years old or more, maybe lots more. The voltage on those should be raised slowly while watching the leakage current. I've never seen a cap rated under 100V that needed forming beyond just putting it in the circuit and turning it on, assuming it hasn't been sitting in the junk box for 20+ years. Caps are subject to contamination problems during manufacture and some of the off-brands can be junk. Get a few chlorine atoms in there and life will be short. If the seals are poor and some chlorinated pot cleaner gets in, life will be short. Mostly it's not an issue with brands like Panasonic, Nichicon, Rubycon etc. Caps aren't known for long life, but I have lots of electronics going on 30 or 40 years old (ok, some going on 60) with their original electrolytic capacitors. The values are ok, the dissipation factors are ok, and the leakage is ok. It would be silly to replace them and they'll probably last many years more. OTOH, I've worked on receivers and such with dozens of little low voltage caps, and they can all be junk in 5 or 6 years. Best practice is to measure!

I have an ancient Sprague leaflet that was given to new employees called, I think, "Birds, Bees & Capacitors" that I'd love to scan and post, but I don't know the copyright status. It covered all the things people building capacitors had to be careful about to avoid shortening the life of the products. Remember the old Sprague saying? "Don't be vague, ask for Sprague!"

CH

[Enzo]

Yes, I know guys who just HAVE to form every cap they use. I won;t say it is useful or useless, but I would point out that MILLIONS of consumer electronics products are produced every year, from little MP3 earbud players to screaming 400 watt tube bass guitar amplifiers, and not a one of those caps is formed before installation. Yet the products, not being aware of that, work just fine anyway. I have been soldering for over 55 years now, and maybe I am just lucky, but I have never seen a cap failure or performance issue that I could remotely attribute to a lack of forming.

[Damon Hill]

Yet the belief persists in 'break-in' for various brands of capacitors (and just about everything else) in audiophile circles. Without doing a thorough before and after test of ESR and dissipation factor, I can't account for any reasons why an electrolytic capacitor ought to sound different after extended in-circuit aging, nor for any other alleged passive component.

It might be helpful to have an objective analysis of electrolytic capacitors, but the required equipment isn't something the large majority of us would ever have access to. Short of passing a test signal through a capacitor and measuring residual noise and distortion, I'd be unable to suggest such a procedure myself. I think it would be worth the effort, just to eliminate one area of uncertainty.

[Ograv]

A lot of good thoughts.
I am beginning to think that there is no bad "forming affect" due to low application voltage on a higher voltage rated cap. Sometimes, I see good deals on caps that have much higher voltage ratings than the old ones I want to replace & I want to make sure I am upgrading & not down grading. Of course temperature rating & size also come in to play in this decision - Bob V

[Triodethom]

From the last 40 years of caps here what I have found. I agree with the statement if less the 100 volts don't worry. Over 100 volts or old parts forming of the caps is a safe way to go. Materials used 30 to 60 years ago are not as uniform as they are now . So the dielectic can fail if hit with a large inrush of voltage . High voltage caps of 300volts plus and materials of teflon etc may take 100 hours to break in from what the audiophiles say. That applies to caps in the signal path . Given that crystal oven freq reference must be at temp to work the high voltage caps may also need to be at operation temp and voltage to work best.

[Conrad Hoffman]

When you first put voltage on a new (electrolytic) cap, it will have some value of leakage current. This value will typically fall for hours, days and maybe weeks until it stabilizes. We're not talking much leakage here at all, and circuits should never be designed such that leakage of that magnitude has any effect on operation. IOW, I don't think anybody can hear the difference caused by the change in leakage current during so-called break-in. Everything else is usually stable from the start.

CH