Yes, it's just that, a feeling, beacuse substituning caps with a servo circuit is no guarantee of anything. You are substituting a single component or two with a whole circuit, with all its internal parameters. Once you hear the "breathing" or "pumping" effect of a not-so-well-done servo, you will know exactly what I mean.
Of course, you could now say that you meant a well done servo, but then I could say a beter choice of the capacitors would have removed many a problem.
Personally, I also prefer a DC servo to any capacitor, but that's another whole new hradache, how to make it so it does what it's supposed to do but with no negative side effects. The maths are simple enough, but the sound? Not necessarily so.
To be fair, I've been using servos on my preamp circuits since the '70s. David Berning was the instigator in that particular case.
I'm on month three of my "monster" H/K ... and boy - it is "on the money".
I'm glad I picked it's EF3 for the forum's projects. 100% thermal , no smoke ,
and really "on the money" !
OS
Fundamental is -10.7dB (you need to read fundamental amplitude), so it is -95dB in fact. Still good. Getting worse with lower frequency, as voltage across the cap is rising.
Attachments
Last edited:
I'm with John on this, I hate to admit
You can get no nonlinearity from linear models, but the model is only used to look at linear effects, like charging/discharging.
If you want to know whether DA can lead to non-linearities, you need a model that includes more accurately any nonlinear soakage-release effects - if these are there; I don't know that.
But saying that DA cannot lead to nonlinearities because the model has only linear elements is a prime example of circular reasonoing.
How does DA manifest other than during charging/discharging?
se
That's not decoupling. And since those caps are typically 10-100uF, good luck trying to get a ceramic cap of that value.
Yes, but they are pretty good at HF, no ? So why not use them in // with other brands as power shunts? If i'm looking for near perfect, I use a 1/10 value for each. Like 1000µ lityc,, 100µ lityc, 10µ film, 1µ film, 0.1µ ceramic.
PMA, thanks for all this very instructive measurements.
PMA, thanks for all this very instructive measurements.
Last edited:
How was it found to be linear if not by looking at this model? Isn't this the chicken versus the egg?
No I am not. I am perfectly aware that the use of a cap is not different in, say, space telescopes or audio. Except maybe the distance to the listener ;-)
Jan
Mouser's listing 2,198 ceramic caps between 10uF and 330uF. Looks like your luck's in
MLCCs | Mouser
Then if we can accurately model its charge/discharge behavior with linear elements, I'm just not seeing any room for non-linearity. What am I missing?
se
What I wonder is whether we can indeed accurately model charge/discharge behaviour with linear elements. Why are we so sure we can?
Jan
Steve it's pretty easy to demonstrate that there is non-linearity. My point in the first place was that the pulse bridge test showed the charge/discharge issues of DA NOT separated from any underlying additional distortion. IMHO this sensationalized the issues at hand. If you went back to what was written at the time you would find a lot of nonsense.
Progressive fitting to the R/C model until there is an irreducible residue, this is a mathematical exercise probably relatively trivial to a modeling/math expert. A few steps on the residual from the pulse test left very little residue, it was clear no one was interested so I dropped it.
But that residual would be the nonlinear component, at least if we are reasonably sure it is not the limitation of the modelling accuracy.
Jan
- Status
- Not open for further replies.