 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
72.211.141.237
Ok so looking more into lclc power supply for my little gm70 amp I realized that the last cap could be in series with the output transformer meaning its in series with the music??
If this is so, IYO what would be better For the last C? more capacitance or less? I am thinking most people would say LESS but... but will this affect Low Freq response?? hmmm yeah
Lawrence
Follow Ups:
One of the goals for optimum power supply performance is to allow no AC voltage to appear at the cold end of the output transformer. If one could accomplish this - literally zero AC voltage at the output of the supply - it could be safely assumed that the power supply injects no audible characteristics into the amplifier. While this isn't possible in practice due to the nature of the physical world, it is more nearly so with large capacitors than with small.
--------------------------
Buy Chinese. Bury freedom.
This is a slippery slope of simplification. While the lack of AC suggests that the supply will add nothing to the signal the achilles heel of that argument is batteries must be ideal.
The thing that is neglected in this line of thought is that the PS must also complete (or is in parallel to) the audio current loop at which point a whole new can of sonic worms is opened.
dave
I'm not neglecting anything. A zero ohm conductor is transparent, regardless of the audio current loop. Does that make batteries ideal? I don't know. That's a different discussion.
--------------------------
Buy Chinese. Bury freedom.
People have been experimenting with large and/or huge capacitors for decades, in the hopes that at some point they would have enough to function like regulators. I have not seen enough repeatability to say there is any consensus. In such experiments with regulators as I have done or heard, more capacitance is usually better but never good enough. And as noted by others, electrolytics especially have HF limitations requiring bypassing, one of the more arcane arts of audiophilia.
I speculate that frequency response is not the right criterion. My tentative candidate is envelope modulation - that is, the fluctuations in musical loudness. This is clearly dependent on the music as well as the amp time constants, which does muddy the waters. And if you are worrying about THAT, then the time constant of the interstage cap or choke or transformer should also be considered.
If you have access to simulations, it would be an advantage to look at the PSU/output stage together, allowing for the OPT inductance (and plate choke if parallel feed) to play its part. It is genuinely more complex than we think initially!
"I speculate that frequency response is not the right criterion. My tentative candidate is envelope modulation - that is, the fluctuations in musical loudness. This is clearly dependent on the music as well as the amp time constants, which does muddy the waters."
One needs to distinguish between class A and class AB amplifiers here. In class A, the average current draw from the power supply is essentially independent of the loudness of the music. The fluctuations in the current draw are on the timescale set by the audio frequencies in the musical signal. Essentially, the job of the final smoothing capacitor is to hold the supply voltage nearly constant during the period of the lowest frequencies that one wants the amplifier to reproduce without colouration. (Bypassed by smaller capacitors too, if the main capacitor has higher-frequency deficiencies.)
By contrast, in a class AB amplifier the average current draw from the power supply can increase very substantially when the music gets louder. In this case, sagging in the power supply voltage during extended loud passages could certainly be a problem in the case of a poorly designed power supply. It seems to me that it is in these class AB cases that the envelope modulation you are alluding to would be a problem.
Of course, if one actually wants a colouration in the sound of a class A amplifier, such as a low-frequency fall off, then a less than "ideal" power supply might give a more pleasing sound. But this, or any other colouration of the sound, might be better achieved in a more controllable fashion, using purpose built audio filters.
Chris
SETs in general are borderline AB - that is, while the output triode is never really cut off, its average current does increase with power. It's not as severe as typical PP amps, but it's not insignificant either.
For example, in the 1950 data sheet for the 300B (which is the one I have at hand this morning) there are two suggested operating points. One has 62mA no signal and 74mA at full power; the other is 60mA to 77mA.
"SETs in general are borderline AB - that is, while the output triode is never really cut off, its average current does increase with power."
Isn't that another way of saying that SETs generate a large amount of even-harmonic distortion? To the extent that the average current increases with the power output, this means that the current on the positive halves of the cycle increases disproportionately more then current decreases on the negative halves of the cycle.
That is, it seems to me, the essence of even-harmonic distortion.
If the distortion is so great that the average current draw is increasing significantly with loudness, then one is probably talking not so much about high fidelity reproduction as about sound effects.
Chris
You can exaggerate the effect by reducing the quiescent current without changing the plate voltage or transformer impedance. There are some on this forum who like that effect.
It's a legitimate question whether the brief against even-harmonic distortion is due to the distortion itself, or the effect on envelope modulation. It will be a long time, if ever, before measurements catch up with psycho-acoustics.
"Ok so looking more into lclc power supply for my little gm70 amp I realized that the last cap could be in series with the output transformer meaning its in series with the music??
If this is so, IYO what would be better For the last C? more capacitance or less? I am thinking most people would say LESS but... but will this affect Low Freq response?? hmmm yeah"
Yes, the last capacitor in the power supply is in the audio signal path. Cold logic would say that the ideal power supply is one that presents essentially zero impedance at all relevant frequencies. This ideal would, other things being equal, be approached by using a large value for the final power supply capacitor. (Maybe bypassed by smaller ones, if there were reason to think that the impedance of the large capacitor had some sort of frequency-dependent anomalies at higher frequencies.) In any case, any shortcomings that represented a deviation from the ideal would be demonstrable by looking at any residual audio signal appearing across the final power supply capacitor(s).
However, there is absolutely no reason to think that the sound quality that the listener will find most desirable is the one achieved by the "ideal" power supply. For example, if the listener prefers a bit of bass cut, then a lower value might lead to a more satisfying sound.
Personally, my view would be that if one wants to experiment with boosts or cuts in certain ranges of the audio spectrum, it would be better to do this by means of carefully designed tone controls or equalisers. It is a bit of a hit and miss approach, if one does it by swapping power supply capacitors. But I understand that my viewpoint is considered to be too clinical and engineering-based by some audiophiles.
Chris
There are ways to mitigate the effects of the last cap in the power supply. See the link below that explains why the signal path is actually a loop.
I read the article thank you :)
but this didn't actually answer my question.
So will having a smaller last cap in the PS cause a low freq response dip?
If no then i would assume that the smaller the last cap the better and possibly bypassed with many smaller caps too.
Lawrence
Yes, you are correct: that link reveals only part of the story. See the link below for the rest.
> So will having a smaller last cap in the PS cause a low freq response dip?
It is not possible to answer this question because it is too vague. What value would you assign to a "smaller" cap? By "low freq response dip" you mean an attenuation of low frequencies relative to the overall frequency response?
> If no then i would assume that the smaller the last cap the better and possibly bypassed with many smaller caps too.
Assumptions are just that: a thing that is accepted as true, without proof.
I suggest that you experiment. Size the PS capacitor that feeds the output tube commensurate with its function. Its function is to ensure that there is sufficient filtering in the PS to eliminate power supply noise and to supply instantaneous current to the output tube for transients (some argue that on transients a low impedance PS is more important than large reservoir caps; another reason to experiment for yourself).
Notice that for the "Elliano Ultrapath" the example shows 100uF for the PS cap and about 40uF for the "ultrapath" cap. I would start there. Just as important as the amount of capacitance is the type of capacitor; they should be very high quality such as an Wima MKP 4.
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
