WHAT voltage across the capacitor?
It would be perhaps better to consider what is happening in reality before the nonsense conclusions that caps will blow up!
In your cicuit, ignoring very small amp offsets, thers is NO voltage across the capacitor. This in the audio passband. That is why a bipolar will work OK!
The actual voltage (not DC unless DC in or a fault) will be in the ratio of the cap's impedance at f to the 'phones RL (=32R?)
To blow up a cap requires considrable heat, ie Power! Where can this come from?
Bipolar elctros are just two caps back to back.
It would be perhaps better to consider what is happening in reality before the nonsense conclusions that caps will blow up!
In your cicuit, ignoring very small amp offsets, thers is NO voltage across the capacitor. This in the audio passband. That is why a bipolar will work OK!
The actual voltage (not DC unless DC in or a fault) will be in the ratio of the cap's impedance at f to the 'phones RL (=32R?)
To blow up a cap requires considrable heat, ie Power! Where can this come from?
Bipolar elctros are just two caps back to back.
Last edited:
After some more reading about this ive decided to go with my first idea, the nichicon Fine Gold capacitors on the output to block any DC incase of anything failing on the circuit & sending DC to my cans, they arnt the best caps you can buy but there was no audible or measurable difference between these & a cheap 7p SMG brand cap, so its a case of why spend more!.
I just read that in general electrolytics are tested to withstand a reverse polarity of 15% of its rated voltage for 125 hours with no damage to the capacitor at all, so thats 2.4v in my case which isnt anywhere near the voltages im seeing, im in millivolts & then only very short bursts into the reverse polarity, so if I was fresh out of college having just learnt all this then technically maybe its not quite right & a non-polar should be used but I should go by past experience & that it doesnt make any difference with these miniscule voltages, the caps remain stone cold even after 5+ hours of playing & there arnt any indications that its degrading the sound quality whatsoever, not by listening or by measuments, so im happy to leave them in place now, I think why I was questioning this was too much reading on the internet about how various capacitors sound etc etc, getting myself drawn into the whole debate
, chances are nothing would ever fail as all parts are rated higher than actually required so are all running cool & well within spec, but I dont want to take any risk with my cans, hence the need for output caps purely as a safeguard only.
I just read that in general electrolytics are tested to withstand a reverse polarity of 15% of its rated voltage for 125 hours with no damage to the capacitor at all, so thats 2.4v in my case which isnt anywhere near the voltages im seeing, im in millivolts & then only very short bursts into the reverse polarity, so if I was fresh out of college having just learnt all this then technically maybe its not quite right & a non-polar should be used but I should go by past experience & that it doesnt make any difference with these miniscule voltages, the caps remain stone cold even after 5+ hours of playing & there arnt any indications that its degrading the sound quality whatsoever, not by listening or by measuments, so im happy to leave them in place now, I think why I was questioning this was too much reading on the internet about how various capacitors sound etc etc, getting myself drawn into the whole debate
, chances are nothing would ever fail as all parts are rated higher than actually required so are all running cool & well within spec, but I dont want to take any risk with my cans, hence the need for output caps purely as a safeguard only.
Last edited:
I think the polarity refers to DC which is always present regardless of the AC waveform. In your case there might be some small negative DC at the output with regards to the speaker terminal but it will be mV. The AC waveform is not likely to be over a few volts peak to peak (2 volts maybe) and that would be so loud your ears will bleed.
Regarding explosions I have had plenty of caps exploding, tantalum ones go with a bang and hit the ceiling, electrolytics just swell up slowly, no bang. But then again I have not worked with valves
Regarding explosions I have had plenty of caps exploding, tantalum ones go with a bang and hit the ceiling, electrolytics just swell up slowly, no bang. But then again I have not worked with valves
I have seen distortion from polar Al electrolytics as output coupling caps
in a SACD player at the lowest frequencies on the test disk we could see a few harmonics just above the noise, this was on the slope of the HF curve - above fc there was no visible distortion
interestingly the visible low frequency harmonics were from the channel with the Nichicon Muse "audiophile" caps - the surround channels with "lesser" grade Nichicons showed no distortion above the noise floor
Cyril Bateman"s "Capacitor Sound" series shows extensive cap distortion measurements - his prescription for large values was nonpolar Al electros with high V rating - thicker Oxide, formed on both foils gave lower measurable distortion than any other electrolytic arrangement
in a SACD player at the lowest frequencies on the test disk we could see a few harmonics just above the noise, this was on the slope of the HF curve - above fc there was no visible distortion
interestingly the visible low frequency harmonics were from the channel with the Nichicon Muse "audiophile" caps - the surround channels with "lesser" grade Nichicons showed no distortion above the noise floor
Cyril Bateman"s "Capacitor Sound" series shows extensive cap distortion measurements - his prescription for large values was nonpolar Al electros with high V rating - thicker Oxide, formed on both foils gave lower measurable distortion than any other electrolytic arrangement
I'm surprised that no one has mentioned the fact that this headphone amplifier uses a bipolar power supply which is +/-15V and that a reverse biased electrolytic is most certainly going to vent if the full supply is impressed across it.
Do you know the probability that the output will go to the rail that will cause a unipolar electrolytic to fail? I don't..
Since the stated purpose of this capacitor is for protection against large dc offsets on the output due to a failure the whole argument about the small amounts of ac across the output coupling cap is probably specious, what counts is what happens when an arbitrary DC polarity is impressed across the capacitor if a failure occurs.
A bi-polar really is needed for the stated purpose, that said even that cap is not going to protect the wearer's ears if such a failure occurs whilst in use.
Were it me I would probably would not bother with the cap and if concerned just use a couple of comparators to sniff for excessive dc levels on the output and if present disconnect/not connect the load via a very good quality relay.
Do you know the probability that the output will go to the rail that will cause a unipolar electrolytic to fail? I don't..
Since the stated purpose of this capacitor is for protection against large dc offsets on the output due to a failure the whole argument about the small amounts of ac across the output coupling cap is probably specious, what counts is what happens when an arbitrary DC polarity is impressed across the capacitor if a failure occurs.
A bi-polar really is needed for the stated purpose, that said even that cap is not going to protect the wearer's ears if such a failure occurs whilst in use.
Were it me I would probably would not bother with the cap and if concerned just use a couple of comparators to sniff for excessive dc levels on the output and if present disconnect/not connect the load via a very good quality relay.
Last edited:
[/QUOTE]What about all the caps used in line-level circuits in every piece of consumer gear out there? I've never seen one that has exploded. The voltage and current is pretty low.[/QUOTE]
Because the the circuits are properly designed. Have you measured the voltages?
[/QUOTE]
I recall reading an app note/datasheet that stated that electrolytic caps don't mind about 1.5V of reverse bias,but after that they get cranky.That's about 70mw on 32ohm 'phones.[/QUOTE]
There are lots of 200 ohm headphones and some can reach 600ohms. Thats sqrt 20 * 1.5v= about 7 volts
Its not the heat that explodes caps (Dc voltages cause no heat on caps(neglecting the miniscule leakage R) its the force from the electric field between the plates, which is proportional to the voltage across them.
And as already stated, if there's a fault you can end up with the full + or -15v across the cap. Chances are slim but theres still a chance.
And caps WILL actually blow up, with a bang and confetti everywhere. Sure gets the heart going!
Because the the circuits are properly designed. Have you measured the voltages?
[/QUOTE]
I recall reading an app note/datasheet that stated that electrolytic caps don't mind about 1.5V of reverse bias,but after that they get cranky.That's about 70mw on 32ohm 'phones.[/QUOTE]
There are lots of 200 ohm headphones and some can reach 600ohms. Thats sqrt 20 * 1.5v= about 7 volts
Its not the heat that explodes caps (Dc voltages cause no heat on caps(neglecting the miniscule leakage R) its the force from the electric field between the plates, which is proportional to the voltage across them.
And as already stated, if there's a fault you can end up with the full + or -15v across the cap. Chances are slim but theres still a chance.
And caps WILL actually blow up, with a bang and confetti everywhere. Sure gets the heart going!
Last edited:
It sure does,not fun at all!
The point about the output latching to the rail and possibly reverse-biasing the output caps in a fault condition is a valid point. But,in normal operation there's no real reason to worry about it since the signal/voltage/current levels are small.
In this case they aren't there for "normal operation" though,use some bipolars rated for at least the supply voltage,and you're set.
Edit: Ohh,and bypassing them with a smaller film cap isn't a bad idea either.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.