That might depend on the capacitor. I suspect there's more than one "root cause."
There are known non-linearities of dielectrics:
dC/dT (temperature)
dC/dV (voltage)
dC/df (frequency)
dC/humidity
DA does not seem to be the cause of any above.
That's why I said 'postulated'.
AFAIK aso there are no formal 'proofs' of DA causing throughput distortion.
After digesting your measurements over I expect a wide range of capacitor types, do you have reasons for capacitor throughout distortion ?.
That's why i mentioned level/operating point shifting...this can cause dynamic long period distortion spectrum changes.
Dan.
Dick Marsh claims that he does, but he forgot what they were. So see if you can jog his memory.
Originally Posted by PMA View Post
There are known non-linearities of dielectrics:
dC/dT (temperature)
dC/dV (voltage)
dC/df (frequency)
dC/humidity
DA does not seem to be the cause of any above.
What is meant by dC/df (frequency) non-linearity ?.
Is DA not a VLF non-linearity ?.
Dan.
There are known non-linearities of dielectrics:
dC/dT (temperature)
dC/dV (voltage)
dC/df (frequency)
dC/humidity
DA does not seem to be the cause of any above.
What is meant by dC/df (frequency) non-linearity ?.
Is DA not a VLF non-linearity ?.
Dan.
Last edited:
So what really is DA ?.
A cap can be shorted for a relatively long period and then return a voltage ramp.
Why ?.
Dan.
DA in those C elcaps?
Would not you rather take care about diode switching transients?
You raise a good issue. Standard practice is to put capacitors across each diode. I have seen 10 nF suggested as the value.
JC of course uses a diode that has low switching noise.
I forgot to make one thing clear, this discussion is about low power supplies as would be used in preamps!
There are so many places to start the discussion.
AC mains noise?
Avoiding unbalancing the transformer?
Diode noise?
Transformer noise?
Surge currents?
Ripple?
Filtering vs frequency?
Circuit theory components not shown in the schematic?
...
ES
Ok, simple description.
A capacitor is essentially two conductive plates separated by an insulator (dielectric). When you charge a capacitor by putting a voltage across it, you're basically sucking electrons off one of the plates and shoving them onto the other plate. So now one plate has a net positive charge from removing some of its electrons, and the other plate has a net negative charge from putting an excess of electrons on it.
So now you have a force between the plates as the positive plate is pulling at all those electrons on the negative plate to try and equalize the charge.
Dielectrics are insulators, but they're not perfect insulators. And some of those electrons on the negative plate will slowly move into the dielectric. Sort of like a sponge, which is why DA is also referred to as "soakage."
However the movement of electrons into the dielectric is very slow. So if you short the capacitor and discharge it for a shorter period of time than you charged it, not all of those electrons in the dielectric will have had a chance to make it back to the conductive plate.
This gradually happens once you open the circuit again, and now the negative plate, which had equalized its charge with the positive plate during the short, begins to become more negative again as the electrons that were absorbed into the dielectric previously return to the plate, and so you begin to see a rising voltage across the cap.
I'm pretty sure this is a pretty accurate description of the phenomenon. I'm sure SY will correct me if it's not.
se
Did someone measured the energy ? The one restored during the first short circuit, the second one after the cap recharge itself etc. ?
Can we model a high DA cap by a normal one paralleled with an other one of high value and high ESR ?
Ahhh. A couple thinkers in the crowd..... First off, you have to realize that energy stored in the Cda portion(s) is not readily available and is subtracted, if you will, from the total energy applied. Can the output then be exactly the same as the input in real time?
Later, the soaked-up energy is released into the external circuit at the Time Constant of the RC rep DA and the external Z.
For the thinkers.... how much later? Suppose a uni-polar impulse waveform with a period within the audio range. Will the DA/TC release modify the original waveform? How so? This one form of distortion.
The primary comparison was bipolar vs film..... that was the issue of the day. So, in your model, use a bipolar cap.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.