Hi there,
I wonder if real world experiences show that the "big-input-caps-leads-to-end-of-the-world" is true?
Here and there some states that large input caps will lead to oscillations/antennas wich means frieds amps and bad sound. Tiny film caps like PEPT are then the best.
But great amps around, in real world then and with moderate power, features big caps and long leads; Virtue Audio, Red Wine...
If this to be a question of casing (like sub-case in the amp case and/or cap's metal case soldered to a ground wire) more than cap size it seems ok to go for good caps, were a cheap MKT deserves to be replaced with a top xxxx caps ( PIOs, Silver-Oil... ) on a nice tweeter... but yes, they are high-voltage parts for XO, not needed for the input.
What to believe?
I wonder if real world experiences show that the "big-input-caps-leads-to-end-of-the-world" is true?
Here and there some states that large input caps will lead to oscillations/antennas wich means frieds amps and bad sound. Tiny film caps like PEPT are then the best.
But great amps around, in real world then and with moderate power, features big caps and long leads; Virtue Audio, Red Wine...
If this to be a question of casing (like sub-case in the amp case and/or cap's metal case soldered to a ground wire) more than cap size it seems ok to go for good caps, were a cheap MKT deserves to be replaced with a top xxxx caps ( PIOs, Silver-Oil... ) on a nice tweeter... but yes, they are high-voltage parts for XO, not needed for the input.
What to believe?
Well now lets see, in a crossover the cap is sized to act as a filter (and thus there is by definition significant signal voltage developed across it), yep that can excite all sorts of weird behaviour as is well documented (Self, Jung, all the rest).
In an input coupling cap we want the value to be large enough that no meaningful signal voltage is developed across the part (Of course circuit impedances are much higher which helps a lot), as long as the value is sufficiently large more or less anything non microphonic should do with film of some sort being an obvious choice, or even if you can arrange a couple of volts of DC bias a (Sacrilage, burn him!) a good quality electrolyic.
Physical size is not a virtue in this position, but as long as you dont get stupid (motor run caps, that sort of nonsense) it is only a consideration if you have mechanical envelope constraints or such poor layout that feedback from the switching node is a real concern (In which case you have worse problems).
Regards, Dan.
In an input coupling cap we want the value to be large enough that no meaningful signal voltage is developed across the part (Of course circuit impedances are much higher which helps a lot), as long as the value is sufficiently large more or less anything non microphonic should do with film of some sort being an obvious choice, or even if you can arrange a couple of volts of DC bias a (Sacrilage, burn him!) a good quality electrolyic.
Physical size is not a virtue in this position, but as long as you dont get stupid (motor run caps, that sort of nonsense) it is only a consideration if you have mechanical envelope constraints or such poor layout that feedback from the switching node is a real concern (In which case you have worse problems).
Regards, Dan.
Self and Cordel both give you some very good guidelines on cap size. Note it depends on cap technology, as to keep distortion down, electros need to be much bigger than films. The input to a class D amp has the same DC blocking and RF filtering requirements as a class B amp, so just go read what they have to say on your input network.
- Status
- Not open for further replies.