Phantom power discharge time

[57sputnik]

I just saw this "similar threads" post below this one about phantom power switch failure that occurred when someone forgot the inrush-limiting resistor.

https://groupdiy.com/threads/phantom-power-toggle-switch-issues.74851/
I agree with winston that the circuit ramping phantom is widely-disseminated but upon close inspection it can be improved simply by moving the +48V bypass capacitor. I certainly won't argue about using an RC for +48V bypass but let the supply deal with the inrush on power up and keep phantom on/off switching currents minimal. It's not fair to say the approach shown in post #2 is "wrong" but it can be improved by simply moving the RC.

It's worthwhile to question the simple things we often take for granted which is why I investigated and co-wrote "Phantom Menace Returns."

EDIT: I remembered that there is a discussion of Phantom switching, which I also co-wrote, here: http://www.thatcorp.com/datashts/dn140.pdf

Types "B" and "C" deal with the situation in which a paralleled multed mic split is used and two devices galvanically-connected are capable of supplying phantom at the same time. I doubt this occurs too often under non-emergency situations with most splits transformer-isolated.

 

[Matt Syson]

As usual there is never a 'perfect' solution to a 'system' that was basically a kludge during the early days of microphone design (and a way to get power to it).
The scheme as drawn by Jensen, is basically sound although 220uF could be considered 'excessive' if the basic 48 Volt rail is actually clean as was usually the case before switchmode supplies became common. The capacitor should have a defined discharge path, either a permanent 100K or so resistor in parallel with the capacitor. or a minor rearrangement of the switch contacts.
Other considerations for a given application are where 'fault' currents might go when you have a faulty mic lead, or a patchfield is used whereby an amp or so of peak current can be stuffed into the input (or an output during those inevitable 'whoops' moments) which is covered in the application notes for the THAT (and previously SSM) input and output chips. Ensuring the phantom is applied (and removed) symmetrically is more important than ramping as such as transformer core magnetisation can occur. Although small switches that are rated at 100mA and 30 Volts may appear to be 'inadequate' having been 'party to' service reports covering many tens of thousands of inputs, I have never heard of a phantom power switch failure where obviously in many implementations the peak current is limited by the pair of 6K8 resistors to around 14milliamps or more depending on an R/C network as shown earlier.
While ramping a supply up and down is a 'nice thing to do' and might help sell some product when you factor in the most likely 'abuse will be through patching, unplug/plug while 'live it becomes a bit pointless.
The negative of a filter capacitor (as indicated) should of course be referred to pin1 of the input XLR (or equivalent position) so that common mode noise (presumably entering via a mic cable) is presented to the preamp circuit to be 'rejected' by it's CMMR.
Matt S