|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
72.161.242.185
In Reply to: RE: Is The Last Cap In The PSU Still Important... posted by Lew on April 24, 2015 at 13:50:17
If you reference the Walt Jung article you can see that the impedance begins to fall off at 1kHz or so at 6dB/octave. To put this in perspective, W. Jung measured the rejection of the cascode DN2540 CCS at 1kHz to be -150dB. This represents an impedance of about 32M ohms. At 130kHz this drops to about -115dB. The impedance is now about 560k ohms, a fifty fold difference. Why be concerned about this out of audio band crap? Because devise nonlinearities can inter-modulate this crap with the audio signal. That's why I believe it's best to not generate this crud in the PS in the first place.
"It is better to remain silent and thought a fool, then speak and remove all doubt." A. Lincoln
Follow Ups:
Obviously I do need to read the article. I never would have guessed that the roll off starts so low in the bandwidth. I thought that the cascode topology ameliorated the effect of device capacitance on hf impedance, for all practical purposes. Meantime, I am using two of the K&K regulators where I once used Vacuum State Super-Regulators, and the K&K do seem to be an upgrade. Do you think it would be worthwhile for me to switch over to the lower voltage MOSFET that you mentioned above? (I'm way below 500V in and out.)By the way, where does the "crap" come from? I thought the issue is that the CCS simply loses impedance at very high frequencies because of device capacitance. This could make it more important to use good quality filtering before the CCS, but when you use the term "crap", I am thinking the issue is spurious "noise".
Edits: 04/26/15
Lew,
By crap I'm referring to any high frequency crud generated within the main PS. Do an Harmonic Analyses of any wave form with sharp corners and you will see high frequency harmonics being generated (cap input supplies for example). Filtering this crud is not trivial exercise do to devise parasitics (interwinding capacitance in transformers and chokes, ESR and lead inductance in capacitors just for example). It would be informative for you to read Morgan Jones treatment of power supplies in the 4th edition of his book "Valve Amplifiers" in which he discusses how these issues degrade the filtering ability of PSs, especially at high frequencies.
The voltage rating for the "lower devise" in the cascode does not matter as this voltage is constrained to the Vgs of the "upper devise" and is only about 3 or so volts.
What cascoding does is "shield" the lower devise from the signal voltage (the drain/source voltage of the lower devise is held constant). The capacitance of MOSFETs is NOT a constant, it varies with the voltage across the devise. If this capacitance is modulated by the audio signal this can result in distortion. Also, cascoding decreases the overall capacitance of the current source (capacitors in series decreases overall capacitance).
If you want to play with part substitution the 500v IXSYS parts are available fron DigiKey and Mouser. I would suggest trying them in the "lower" position and NOT mounted on the heat sink.
"It is better to remain silent and thought a fool, then speak and remove all doubt." A. Lincoln
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: