When a transformer winding is driven from an amplifier whose DC voltage level must be blocked from the driven winding and high isolation requirements are an issue, it is better from a cross-coupling suppression standpoint to put that capacitor in the low side of the transformer winding instead of in the high side.
The capacitor in the high side on the left can act like a transmitting antenna, compromising the effect(s) of signal guard foils.
The capacitor placed in the low side as on the right half of this sketch will have far less signal radiation.
Interesting. What is the mechanism which causes this to be the case?
Posted by: Frank Walker | October 21, 2011 at 01:38 PM
There is electrostatic coupling from a capacitor that gets placed on the "hot" end of the winding, especially if that capacitor is physically large such as a CS13-style part. In the case that led to this, there were over a dozen different audio signal paths, each of which (to simplify just a little bit) needed to be isolated from all of the others by 100 dB.
Posted by: John Dunn | October 21, 2011 at 06:34 PM
Sorry John but I don't understand why this electrostatic coupling happens only with the capacitor in high side position. Maybe because in low side one terminal is connected to ground?
Posted by: Marco Ramondetti | October 22, 2011 at 07:37 AM
Hello, Marcos. When the capacitor is put on the low side of the winding, one end of that capacitor is indeed at ground. Since the capacitor is a low impedance device, the ungrounded end is at ***signal*** ground as well. The high side of the transformer stil carries the full signal of course, but that lead can be given shielding. There was no way to have shielded the capacitor itself.
Posted by: John Dunn | October 22, 2011 at 07:58 AM