[sdiy] Signals leaking into the PSU?

[Ben Bradley]

I'm not sure I read every word in these threads, but I want to add
something I didn't see mentioned.:

On Mon, 20 Feb 2023 at 06:19, Roman Sowa <modular at go2.pl> wrote:
...
> BTW, throw away that old LED requiring 330 ohms to light up, and buy new
> one which shines up brightly at 0.5mA. 30mA square curent jumps in power
> rails are troublemakers.

This is true, modules should not cause jumps in current pulled from
the common power supply. Likewise, modules should be immune from
voltage changes in the (allegedly regulated) common power supply.

A problem with the current multiplier (whether single BJT or op-amp)
is any current change in the output will still appear instantly on the
current pulled from the power supply. It may "regulate" the output,
but at the expense of what's powering it.

The best way (IMO) to keep constant current pulled from the power
supply is by using parallel regulators. I haven't used the LM4040 that
was just mentioned, but I've used quite a few LM431/TL431 for
regulation. When the circuit it's powering pulls more current, the
parallel regulator instantly pulls LESS current to keep the voltage
the same, and the current pulled through the resistor from the power
supply remains the same.

This also keeps the current returned to the power supply through the
ground (or the negative supply rail) the same as well. This leeps
small voltage changes between modules from happening, which would
otherwise change VCO pitch and such.

I of course agree with reducing current jumps by using a more
efficient LED. It doesn't seem efficient to have an LED powered from a
12V source. Eurorack has a 5V supply for such things, but there's
still a common ground (a non-zero resistance back to the power supply,
and in common with the other power sources) that can get polluted by
varying currents.