Tom has already brought up an important point:
You may have to make a plan of the grounding layout (external connections included!) in order to identify potential culprits. A classic, for example, would be the Quad 405 problem - two RCA input jacks spaced far apart but feeding an amplifier with a common power supply ground. This scenario easily results in a massive internal ground loop once a stereo source with the usual common ground between L and R channels is connected, unless you deliberately route your input wiring to minimize it as far as possible.
BTW, due to the dual nature of E and H fields / capacitance and inductance, there has to be a magnetic equivalent of an electrostatic shield. Where the latter works by keeping the voltage between the shield and circuit ground constant, it would have to have a constant current delta. Methinks this is another inductor of good conductivity placed right next to the problematic loop.
You can actually find these as shorting rings in speaker drivers (used to knock down voice coil inductance along with its associated nonlinearity), and I think toroidal transformer "belly bands" serve the same purpose.
Magnetic coupling is only ever a problem if your circuit features loops to couple into. Should be obvious, right? So you want to hunt down and eliminate these, or at least greatly reduce their loop area, which is the #1 driver of inductance.
You may have to make a plan of the grounding layout (external connections included!) in order to identify potential culprits. A classic, for example, would be the Quad 405 problem - two RCA input jacks spaced far apart but feeding an amplifier with a common power supply ground. This scenario easily results in a massive internal ground loop once a stereo source with the usual common ground between L and R channels is connected, unless you deliberately route your input wiring to minimize it as far as possible.
BTW, due to the dual nature of E and H fields / capacitance and inductance, there has to be a magnetic equivalent of an electrostatic shield. Where the latter works by keeping the voltage between the shield and circuit ground constant, it would have to have a constant current delta. Methinks this is another inductor of good conductivity placed right next to the problematic loop.
You can actually find these as shorting rings in speaker drivers (used to knock down voice coil inductance along with its associated nonlinearity), and I think toroidal transformer "belly bands" serve the same purpose.
Magnetic coupling is only ever a problem if your circuit features loops to couple into. Should be obvious, right?
Yep. Should be equally obvious that you will always have loops. Current flows in loops, after all. So you're left with managing current flows and reducing loop areas.
Tom
Yep. Should be equally obvious that you will always have loops. Current flows in loops, after all. So you're left with managing current flows and reducing loop areas.
Tom
Tom,
I have done that in my current crossover unit. It does move the problem to e-field and line filtering which is easy. Though I will likely go back to linear as I want to get rid of the virtual ground as it is causing turn on and off problems. I have not found reasonably priced +/- small tracking switchers. I am thinking about several other boxes where they have multiple supplies for audio and digital mixed.
One question I have in SMPS use is turn on current spikes. Transformers have no problem with this and pretty easy to get diodes large enough, but unknown what is inside my various old laptop supplies.
I have done that in my current crossover unit. It does move the problem to e-field and line filtering which is easy. Though I will likely go back to linear as I want to get rid of the virtual ground as it is causing turn on and off problems. I have not found reasonably priced +/- small tracking switchers. I am thinking about several other boxes where they have multiple supplies for audio and digital mixed.
One question I have in SMPS use is turn on current spikes. Transformers have no problem with this and pretty easy to get diodes large enough, but unknown what is inside my various old laptop supplies.