Gain,
with regards heating of capacitors you are correct provided you are referring to d.c. and the load is resistive. But as Andrew mentions, I also believe there has to be some heat generated when a.c. current flows through the capacitor, else there would be no reason to specify the ripple current limits.
with regards heating of capacitors you are correct provided you are referring to d.c. and the load is resistive. But as Andrew mentions, I also believe there has to be some heat generated when a.c. current flows through the capacitor, else there would be no reason to specify the ripple current limits.
AndrewT said:
ok i will concede the debate to Andrew. if he says caps can produce heat then i believe he is correct. i have read enough of his posts to know that he knows his stuff. REALLY knows his stuff.
i do not understand though. i have gone over my calcs and can't see where i went wrong. could someone who is good with math (unlike me) please help point out the error of my analysis. i'm just naive i guess and don't get how a voltage and current wave that are 90 degrees out of phase can produce power.
thanks, especially to AndrewT.
Nico Ras said:
imho you are backwards in your statement. heating occurs when DC is applied at first due to the resistance present in the legs of the cap. this will, however, decay to (almost) zero as time goes on and exponential function asymptotically approaches V. so cap will heat initially with DC but once charged will stay cool.
with AC there will be no heating due to the 90 degree phase diffrence between the voltage and current waves. except for the heating in the legs of the cap due to resistance (not reactance!) of the i squared r law of current AC or DC through a resistor (in this case not a resistor but the leg of the cap)
fwiw, in all my years of dealing with electronics (almost 8 now) i have never once felt a cap that was warm. ever. even filter caps in PSU's that were playing metallica cranked to ten.
Maybe the ripple currents produced by a normal amplifier
running off a PSU just don't produce enough heat to be
noticable ,still..., a little heat is produced.
BUT , try hooking up a cap reversed biased (I did this by
accident..reversed rails
) ,the cap got so hotI had to dip it in the sink!!
BTW ,the caps survived but giving it FULL AC at reverse
polarity could of boiled a cup of tea.
In cheap HT receivers where they use 63V/4700uf caps
(at +- 58V rails) for all FIVE channels I've felt warm caps
(with 6-10v ripple). The way we DIY'ERS use 40Kuf/100V
caps with our 70V rails one should enjoy 1-2V ripple
and 10K+ hours of listening pleasure (cold caps with overdriven
Metallica,as gain says
)
this thread has gotten off topic imo.
we do not care about whether or not caps make power when AC is passed through them. imho the answer to that question is obvious. i wait to be standed corrected.
i care about devices that OBVIOUSLY heat up. transistors, diodes, resistors, etc. whats you all's opinion?
will amp live longer if left on continuously or if turned off when not in use?
we do not care about whether or not caps make power when AC is passed through them. imho the answer to that question is obvious. i wait to be standed corrected.
i care about devices that OBVIOUSLY heat up. transistors, diodes, resistors, etc. whats you all's opinion?
will amp live longer if left on continuously or if turned off when not in use?
gain said:
I thought I had shown how power is dissipated in the effective series resistance of a cap.
Anyway to answer your question. I have done both. The stereo in my living room is an old Onkyo. It's been on pretty much continuously for 11 years. It was a good 5 years old when I got it. It has been as stable and resolute as any amp could be.
My special system get's turned on and off every time I change something in the system - sometimes as much as 10 times a day. I've been abusing it this way for 6 years now and the only fuse I blew was when I accidentally connected a faulty crossover.
So from my experience, it doesn't seem to make much difference - as long as the gear is designed right and appropriate specs/tolerance components are used.
gain said:
one quick comment on this one, then back to the original topic...
if you have a 1000uf cap that averages 1amp ripple current, if we assume no ESR, then no problem, there is no heat generated since the voltage and current are out of phase. if we assume .1 ohm ESR, that would give a 0.1 volt drop across the series resistance, giving an average wattage of 0.1W. not a lot, but it does still generate some heat. this heat is generated in the plates and electrolyte, and causes a miniscule amount of evaporation, which after time raises the ESR, and thereby the amount of internal heat. let's just say for argument that the ESR doubles for every 500 hours of use. if we started at an ESR 0f 0.01 ohm (good for a brand new cap), how long would it be before the ESR got to a critical point (where the cap is generating 2 or more watts of heat, enough for a catastrophic failure)? about 4,000 hours. of course the operating parameters as well as ambient temp have a large effect on the climb rate of the ESR. and in most circuits, the climbing ESR would cause other problems usually well before the internally generated heat has a chance of causing a catastrophic failure.
to connect this back to the original topic, the longer you operate electrolytic caps, the closer they are to the end of their limited life span. the problem with powering up and down all the time is thermal expansion causing 1) stress rings in solder connections and less frequently 2) failures of some semiconductors. in vacuum tube equipment constantly turning equipment on and off causes early failure of tuve filaments.
back in the 70's and 80's there were some transistors in TO-220 cases that had emitter lead failures regularly from thermal stresses and separation of dissimilar metals.
gain said:
You skipped the third option : stand-by mode.
Either by dropping the dissipation/temperature with a low bias level switch, or running the front end live all the time (separate powersupply) and only switching the output stage On/Off.
Added bonus for Class A amps is that it drastically reduces the warm-up period, you'll have a 90% performance level within 15 minutes.
90 degree V/I phase shift is theory book jabba, doesn't take ESR and leak current losses into acount.
You could consider taking the wrapper of lytic parts Off.
Keeping an amp clean inside is more effective than counting kwh, ime.
Hi Uncle Jed,
what you are saying from your calculations is that an electrolytic cap will not last for more than 166 days if power is left. Sorry, but I cannot agree with that statement. Surely caps are rated at 8000 hrs at maximum temperature and ripple current.
Our switch mode battery chargers are found in almost every high demanding environment such as in cellular radio networks, telecommunication repeaters etc. all over the country running 24/7 in active stand-by, thus the equipment is run 100% off the charger with spare capacity to bring the batteries up in four hours.
These stations are not air conditioned and ambient of 45 degrees centigrade is not uncommon.
We warranty the product under these conditions for five years, and have a failure rate of 0.2%. We would be bankrupt if they only lasted for 166 days.
what you are saying from your calculations is that an electrolytic cap will not last for more than 166 days if power is left. Sorry, but I cannot agree with that statement. Surely caps are rated at 8000 hrs at maximum temperature and ripple current.
Our switch mode battery chargers are found in almost every high demanding environment such as in cellular radio networks, telecommunication repeaters etc. all over the country running 24/7 in active stand-by, thus the equipment is run 100% off the charger with spare capacity to bring the batteries up in four hours.
These stations are not air conditioned and ambient of 45 degrees centigrade is not uncommon.
We warranty the product under these conditions for five years, and have a failure rate of 0.2%. We would be bankrupt if they only lasted for 166 days.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.