Hi all,
I have just finished building an amplifier based on the LM6171 with a Buf634 on the output.
Its all going well & sounds very nice, I have measured for DC offset & there is nothing much at all, however I am concerned about the possibility of something going faulty & destroying my headphones, for example the buffer sending the V+ (15v) through to the output, or anything else going faulty come to that.
I have fitted 470uf electrolytic's as output caps, I have read about the fact they can degrade the sound quality but so far I havent heard anything bad.
I also have a 15-0-15 supply & have read that because my output would essentially be swinging from + to -, that I would require a bi-polar cap.
But here is the thing, everywhere I read on the net says the same thing, that by using a polarised cap in the way I am should sound terrible, as half of the music cycle is causing the cap to work the wrong way, against the polarity, its too short to make it go bang I believe, but everyone says it will sound terrible!!
I have wired up a few caps to test in a way that I can go from, through cap, to no cap, & I have always had good ears but I honestly can not hear even the slightest difference with or without the output caps in place.
It kinda makes a mockery of what ive been reading online about all this subject, it should be sounding terrible, but there honestly is no difference, ive even hooked it up to a scope & then my PC (RMAA) again no visual difference to a square wave or anything else.
Can anyone shed any light onto why im able to use a polarised cap in this way without any sound degradation? ive even tried the done thing of bypassing the caps with 0.22uf polyesters, as I read people saying that it made a big difference to the highs, but again I havent found that do anything either, there is no loss of highs with the electrolytic in place or not.
One last thing, ive also tried several caps, ranging from a cheap 7p SMG brand, to a Nichicon FG, also from 0.22uf up to 1000uf, bass was effected with too lower values as expected (corner frequency I believe its called) but from 220uf onwards it was back to being flat from 10hz to 100khz.
So im puzzled!
Nick
I have just finished building an amplifier based on the LM6171 with a Buf634 on the output.
Its all going well & sounds very nice, I have measured for DC offset & there is nothing much at all, however I am concerned about the possibility of something going faulty & destroying my headphones, for example the buffer sending the V+ (15v) through to the output, or anything else going faulty come to that.
I have fitted 470uf electrolytic's as output caps, I have read about the fact they can degrade the sound quality but so far I havent heard anything bad.
I also have a 15-0-15 supply & have read that because my output would essentially be swinging from + to -, that I would require a bi-polar cap.
But here is the thing, everywhere I read on the net says the same thing, that by using a polarised cap in the way I am should sound terrible, as half of the music cycle is causing the cap to work the wrong way, against the polarity, its too short to make it go bang I believe, but everyone says it will sound terrible!!
I have wired up a few caps to test in a way that I can go from, through cap, to no cap, & I have always had good ears but I honestly can not hear even the slightest difference with or without the output caps in place.
It kinda makes a mockery of what ive been reading online about all this subject, it should be sounding terrible, but there honestly is no difference, ive even hooked it up to a scope & then my PC (RMAA) again no visual difference to a square wave or anything else.
Can anyone shed any light onto why im able to use a polarised cap in this way without any sound degradation? ive even tried the done thing of bypassing the caps with 0.22uf polyesters, as I read people saying that it made a big difference to the highs, but again I havent found that do anything either, there is no loss of highs with the electrolytic in place or not.
One last thing, ive also tried several caps, ranging from a cheap 7p SMG brand, to a Nichicon FG, also from 0.22uf up to 1000uf, bass was effected with too lower values as expected (corner frequency I believe its called) but from 220uf onwards it was back to being flat from 10hz to 100khz.
So im puzzled!
Nick
That is because the difference between caps is so small that you can't hear it. This might also mean that you are not filled with all the audiophile crap and you're still able to judge by yourself if it sounds good or not.
See these and have a laugh and/or feel sorry for them:
Matrix HiFi --> Blind testing high end full equipments
ABX Double Blind Comparator Data
BTW, nice option for the buffer.
So im puzzled!
Nick
You should be happy, rather than puzzled. Or you can continue experimenting. How about a diode in series with that cap? An LED? There must be something you can hear.
Unrelated question: do you listen long and loud on your cans?
Slightly more seriously: i doubt a reversely biased capacitor becomes a significant distortion generator at very low voltages. It would be a problem at speaker levels but not so much with cans. You mention nothing about your listening acuity or the quality of your system. Do you consider yourself an experienced listener? Per example can you hear absolute phase?
Yes I do get the feeling that there is a lot of exaggeration going on when I read about the so called differences.
I will have a look at those links next up..
Yes I am very happy with the circuit, as for a diode or an led, im not exactly trying to add components until I can hear an audible difference, im really only after protecting my cans against DC, the cap being a good blocker of DC.
Well I dont really buy into all that "Experienced Listener" title, I just know what I like when I hear it, but I am very good at hearing subtle differences, which is why im puzzled why I cant hear ANY changes to the sound by introducing an output cap, especially when I read online from others about how bad they are & that they seriously degrade the signal, but im wondering if its been blown out of all proportion, or maybe more the case it could be circuit dependant, maybe some circuits may put the cap under more stress & reveal any weaknesses, the output cap in my circuit isnt actually working all that hard at all, its basically a DC blocker.
Im more concerned about the actual sound of the whole idea of adding such caps, it would never actually explode at such miniscule AC voltages, but yes it could lower the life of the cap I guess.
I thought about what you said about not being filled with all the audiophile crap
especially when I saw just now a website selling (wait for it) oxygen free audiophile grade mains lead for an insane price, they actually said there was a noticeable difference to the sound by using this lead, what? how about when they plug it into the wall socket which has behind it metres & metres of standard cheap twin & earth mains cable throughout the house Its a world I am getting drawn into, but im trying to stay subjective about it all, I trust my ears more than what ive read, im slowly coming to the theory that it might be how you implement parts in a circuit whether or not they will sound good or bad or hear no difference, maybe in a tube amp where the output caps have 100v fed to them perhaps then they might start to have their own sound signature, being as they are actually working harder, I dont know, im clutching at straws here...
I can’t hear crap with headphones (maybe because I’m not used to them). With real room speakers though I can hear some fairly subtle changes. Not sure what that says about me, or the importance of the outer ear in hearing (headphones send the sound straight down the ear canal while speakers use the outer ear in a more normal way). You may want to try a good power amp with some decent speakers before making very many generalizations based on your finite experiences so far.
One thing I will say is that there is a big difference between a capacitor on the output of an amp compared to on the input. Distortions that capacitors will introduce will be small... if that distortion is on the output it will quite likely not be audible. If however it is on the input (before amplification) then depending on the gain of the amplifier, the distortions may be amplified many times over, and can become audible.
If for no other reason, you should change the caps to non-polars for safety reasons... Exploding electrolytics are not nice, and I would hazard a guess that if one exploded whilst you were wearing your headphones, that it would not be a pleasant experience.
Tony.
If for no other reason, you should change the caps to non-polars for safety reasons... Exploding electrolytics are not nice, and I would hazard a guess that if one exploded whilst you were wearing your headphones, that it would not be a pleasant experience.
Tony.
Welcome to the world of Audiophoolery,Snake Oil,Hearsay,and Myth.
Is a cap really gonna explode at the small voltages encountered in a headphone amp? I have my doubts. The signal level is unlikely to exceed a couple volts,at the most -and even then it would be unbearably/dangerously loud.
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.
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.
If it were on the output of an amplifier meant to drive a full-size speaker,that would be different.
Is a cap really gonna explode at the small voltages encountered in a headphone amp? I have my doubts. The signal level is unlikely to exceed a couple volts,at the most -and even then it would be unbearably/dangerously loud.
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.
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.
If it were on the output of an amplifier meant to drive a full-size speaker,that would be different.
Hi DigitalJunkie whilst you are probably correct, If one is building something then using the appropriate components is never a bad Idea. What if someone uses the wrong components for design A, and gets away with it, then for design B thinks well it worked for design A so I'll do it again, without realising the implications that because design B is different that it will most definitely not work.
IMO taking short cuts whilst it might work is never an acceptable solution, that is just me. I've seen too many times (in all different areas of life) where taking short cuts ends up being a costly mistake.
Tony.
IMO taking short cuts whilst it might work is never an acceptable solution, that is just me. I've seen too many times (in all different areas of life) where taking short cuts ends up being a costly mistake.
Tony.
Could you give an example for "subtle differences"? I specifically asked about absolute phase as it is recognised as audible even by the cable/cap/amp-deaf and is easy and cheap to reproduce.
Well I meant anything in general really, perhaps a slight loss in highs, or even just something you cant explain what it is but you hear a difference, like when I swapped opamps on my soundcard (xfi), couldnt put my finger on what was different but it just sounded more forward & slightly better, really minimal difference as expected but audible (just about), but in this case there are no differences that can be heard or seen on a scope, so im trusting my ears with this one & going with the fact that if there are any differences, its not audible or measurable so thats plenty good enough.
As for source which I completely forgot to mention not refused to, its not important for this test, ive fed it via a computer sound card, my trusy old technics slp-1200 & slp-520 cd players & even a cheap mp3 player, they all gave varying levels of quality, the technics being the best source (naturally) but like I said, source is irrelevant in this case & was kept the same for all tests.
Last edited:
Your correct it cant explode with the levels we are talking about here, but im guessing it could shorten the lifespan a little, but being as they arnt that expensive im going to put 470uf non polar electrolytics in the output, I may bypass them with 0.22 polys which ive seen is the "done thing"
although I havent found them to make any difference, ive read they use the 0.22uf's to bypass the electrolytic to bring back the highs that have been lost, well I havent lost any highs that I can hear & so I wanted to get a visual of what was going on & the scope is reading up to 100khz with no loss which im sure is plenty high enough, bass was fine until around 5hz where it had a slight deviation, but again thats only air at that frequency, not real bass.
If you want to do some good measurements, try rightark RMAA I use it for testing distortion among other things... it is only as good as your soundcard, but it can reveal changes that come about after you modify a circuit.
here is a comparison before and after changing the input cap on a kit soundcard buffer preamp that I built. Original was a generic 10nf (probably polyester) cap replacement was a 68nf wima polyproylene. Note I was only changing to fix the low freq rolloff problem but got a pleasant surprise when the distortion went down.
Note that if what you are testing puts out more than a couple of volts you need to cut down the levels, or you will more than likely fry the inputs on your soundcard!
Tony.
edit: if your using firefox and the images in the link don't work try (shock horror) IE... (recent versions of firefox seem to have an issue with the png files...)
here is a comparison before and after changing the input cap on a kit soundcard buffer preamp that I built. Original was a generic 10nf (probably polyester) cap replacement was a 68nf wima polyproylene. Note I was only changing to fix the low freq rolloff problem but got a pleasant surprise when the distortion went down.
Note that if what you are testing puts out more than a couple of volts you need to cut down the levels, or you will more than likely fry the inputs on your soundcard!
Tony.
edit: if your using firefox and the images in the link don't work try (shock horror) IE... (recent versions of firefox seem to have an issue with the png files...)
Last edited:
Hi Tony,
Yes ive been using RMAA, its a great way to get a visual of whats going on with the amp, as you say its only as good as the soundcard used but I was happy to see that my headphone amp was giving identical readings to the soundcard, indicating that the soundcard was the limiting factor, as on a scope & a seperate signal generator I measured no deviation right up to 100khz, which the soundcard cant manage & drops off at 30khz, not that anyone on the planet will hear those frequencies but if it can do 100Khz then 20khz must be a breaze for it to manage ;-)
Overall im very happy with the circuit, I havent any caps in the signal path now except for these output caps im fitting so that could be helping with the great sound, simple is best, but im going to order up some non-polar just so its technically correct on the output.
Yes ive been using RMAA, its a great way to get a visual of whats going on with the amp, as you say its only as good as the soundcard used but I was happy to see that my headphone amp was giving identical readings to the soundcard, indicating that the soundcard was the limiting factor, as on a scope & a seperate signal generator I measured no deviation right up to 100khz, which the soundcard cant manage & drops off at 30khz, not that anyone on the planet will hear those frequencies but if it can do 100Khz then 20khz must be a breaze for it to manage ;-)
Overall im very happy with the circuit, I havent any caps in the signal path now except for these output caps im fitting so that could be helping with the great sound, simple is best, but im going to order up some non-polar just so its technically correct on the output.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.