Capacitors are used at the outputs of some pre and power amplifier designs to block DC offset. Some people have the opinion that these capacitors can be subjectively detected and that they ‘do not sound good’. I would like to ask for your opinion on the subject…
> Preamplifier:
Have you changed a circuit to replace an output capacitor with a DC servo, and if so what was the subjective difference?
Have you made objective distortion measurements with and without the capacitor… any difference?
> Power amplifier:
Large value electrolytic capacitors are placed at the output of single ended transistor amplifiers that use single supply voltages…
Have you heard amplifier designs with electrolytic output coupling capacitors (your own designs or commercial)... what was your opinion of the performance with these capacitors in the circuit?
> Preamplifier:
Have you changed a circuit to replace an output capacitor with a DC servo, and if so what was the subjective difference?
Have you made objective distortion measurements with and without the capacitor… any difference?
> Power amplifier:
Large value electrolytic capacitors are placed at the output of single ended transistor amplifiers that use single supply voltages…
Have you heard amplifier designs with electrolytic output coupling capacitors (your own designs or commercial)... what was your opinion of the performance with these capacitors in the circuit?
I don't believe there is any audible difference. That said, the main reason for using such a circuit is for economy. It only requires a single power supply, mainly.
The capacitor itself is important but most modern units are just fine. There needs to be feedback around the capacitor, not from its input.
The capacitor itself is important but most modern units are just fine. There needs to be feedback around the capacitor, not from its input.
I use a DC servo on my power amp... which is single ended, DC coupled running on split supplies.
http://focus.ti.com/lit/an/sboa092a/sboa092a.pdf
I have built several AC oupled amps over the years and never had cause to suspect the coupling cap was doing anything to the sonics.
Caps in preamps are transparent if the design is right. I have no issues at all with using them.
Built a Radford HD250, that was single ended.
http://focus.ti.com/lit/an/sboa092a/sboa092a.pdf
I have built several AC oupled amps over the years and never had cause to suspect the coupling cap was doing anything to the sonics.
Caps in preamps are transparent if the design is right. I have no issues at all with using them.
Built a Radford HD250, that was single ended.
Member
Joined 2009
Paid Member
my recent experiences are here (and quoted below)
http://www.diyaudio.com/forums/solid-state/145652-tgm2-amplifier-12.html#post2194885
Output Capacitor: just for interest, and although not needed, I thought I'd listen to the amplifier through an output capacitor. I found that the bass was noticeably worse off through a 220uF Nichicon MUSE. But through a 2,200uF Nichicon MUSE the bass was fine. I also tried a 10,000uF Nichicon HE cap. Overall I was hard pressed to hear a difference between the use of the two larger capacitors but I could hear a very subtle difference between cap and no cap. The output without a capacitor seemed to have just a tad more 'space' around the upper mids. That's about all I know how to explain what I heard. I concluded that I could be happy with an amplifier that used an output cap.
http://www.diyaudio.com/forums/solid-state/145652-tgm2-amplifier-12.html#post2194885
Output Capacitor: just for interest, and although not needed, I thought I'd listen to the amplifier through an output capacitor. I found that the bass was noticeably worse off through a 220uF Nichicon MUSE. But through a 2,200uF Nichicon MUSE the bass was fine. I also tried a 10,000uF Nichicon HE cap. Overall I was hard pressed to hear a difference between the use of the two larger capacitors but I could hear a very subtle difference between cap and no cap. The output without a capacitor seemed to have just a tad more 'space' around the upper mids. That's about all I know how to explain what I heard. I concluded that I could be happy with an amplifier that used an output cap.
Member
Joined 2009
Paid Member
I agree that even without a dc coupling cap on the output you still have capacitors in the signal path because of the PSU. However, doesn't mean that it's good for sound. There are many reports (look at the Class A web site dedicated to JLH) of people improving the sound by swapping out the PSU caps with higher quality units.
Member
Joined 2009
Paid Member
I need some help to understand this - as far as I know, which isn't very far, the PSRR performance of the amplifier is how well it prevents power rail contamination reaching the output but does not necessarily mean that the reverse is true - that it prevents the output from passing through the power rails. So I don't see the connection between PSRR and the fact that the signal does pass through the PSU caps.
psrr is the measure of how much the circuit responds to an independent source added in series with a ps rail (actually 2 #, either +/- ps rejection or diff/common mode ps rejection)
as long as the circuit is operating in the linear (or "weakly nonlinear") region then it doesn't "care" about the cause of the ps V change - line ripple, circuit current demand to drive the load or capacitor DA, ESR, etc..
so a large psrr means the response is immune to ps V changes from any cause - specifically including the load current loop completing through the ps cap's imperfections
high loop gain enables high psrr - but other means such as ccs bias, cascodes, ground referencing compensation/gain stages, bridged/"balanced" outputs can improve psrr
the "graduate level" psrr paper:
A General Relationship between Amplifier Parameters and its Application to PSRR Improvement (IEEE CAS v38#10 1991) Eduard Säckinger , Josef Goette , Walter Guggenbühl , Walter Guggenb�uhl Senior
[pdf can be found cached on CiteSeerX - but seems to be a dynamic link]
as long as the circuit is operating in the linear (or "weakly nonlinear") region then it doesn't "care" about the cause of the ps V change - line ripple, circuit current demand to drive the load or capacitor DA, ESR, etc..
so a large psrr means the response is immune to ps V changes from any cause - specifically including the load current loop completing through the ps cap's imperfections
high loop gain enables high psrr - but other means such as ccs bias, cascodes, ground referencing compensation/gain stages, bridged/"balanced" outputs can improve psrr
the "graduate level" psrr paper:
A General Relationship between Amplifier Parameters and its Application to PSRR Improvement (IEEE CAS v38#10 1991) Eduard Säckinger , Josef Goette , Walter Guggenbühl , Walter Guggenb�uhl Senior
[pdf can be found cached on CiteSeerX - but seems to be a dynamic link]
Last edited:
Member
Joined 2009
Paid Member
well if someone tells me they can "clearly hear" ps cap changes in a "no feedback" circuit with 0 to maybe 10-20 dB psrr then I think there is 100-1000 X less "error" to hear with a high loop gain, high psrr circuit with psrr ~60 dB over the audio frequency range when the same ps caps are compared under similar signal/load conditions
I'd like to think you can ignore ps Cap "audiophile qualities" when ( load current * ps impedance / psrr ) is samller than any of source, amp, room or human hearing noise floor
I'd like to think you can ignore ps Cap "audiophile qualities" when ( load current * ps impedance / psrr ) is samller than any of source, amp, room or human hearing noise floor
Last edited:
Right, i misswrote. I should have said: no more influence than psu caps.
I just want to kill a myth.
So I have been told. But there are other advantages of a dual supply that overcome this limitation.
Besides, the psu caps can be bypassed with very high quality ones.
This causes confusion.
Does the output signal pass through the PSU caps ?
Yes but not in the way most imagine.
The common misconception is that the PSU caps somehow define the low frequency performance... that they are "in series with the output".
They do not and are not.
If that were true then amps could not be made to work down to DC... how can DC pass through a cap ?
The output of the amp is determined by feedback and it's points of reference for that feedback and the input signal.
It doesn't matter what the current is in the PSU caps, it's irrelevant to what the amp outputs.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.