Dropping phantom power with few losses

[fryingpan]

Say I wanted to bring phantom power down to 9 or 18V, considering you have very little current (~5mA, IIRC), what's the best course of action? A resistive voltage divider would be very wasteful.

[merlinb]

In a proper desk there should be at least 10mA available. Usually we do indeed drop it across a resistor when required, with a Zener to regulate. No special effort to make it 'efficient'. Why is that a problem in your case? You could use an emitter follower to buffer a fairly high-resistance divider, if you prefer.

[fryingpan]

Well, that's the problem. On the field, you can find all sorts of "desks" (sometimes it's just a cheap mixer). Say you want to power a bass preamp (a la Sansamp), you need all the juice you can get.

[antonis]

A more elegant way:

[amptramp]

Why not go with an LM317HV?

This is the high voltage version that can take 60 volts in as a supply source.  Just calculate the dissipation you are going to get but you can use a voltage dropping resistor in series with the input that will drop the regulator input voltage as the current drain increases, tending to keep the power in the regulator constant.

[merlinb]

Well, that's the problem. On the field, you can find all sorts of "desks" (sometimes it's just a cheap mixer). Say you want to power a bass preamp (a la Sansamp), you need all the juice you can get.

Then you need a DC-DC buck converter

[R.G.]

Say I wanted to bring phantom power down to 9 or 18V, considering you have very little current (~5mA, IIRC), what's the best course of action? A resistive voltage divider would be very wasteful.

Then you need a DC-DC buck converter

Any time you think "switching power supply" or "converter" your next thought should be "how do I deal with the switching noise". A poorly designed switching supply/converter could easily pollute the limited current from a phantom supply and might pollute the whole desk if it really has issues.

It's possible that the best answer to running 9V stuff from phantom power is "try to figure out some other way".
 

[fryingpan]

Any time you think "switching power supply" or "converter" your next thought should be "how do I deal with the switching noise". A poorly designed switching supply/converter could easily pollute the limited current from a phantom supply and might pollute the whole desk if it really has issues.

That's what I was thinking. I don't like relying on converters, charge pumps, etc. even though the good chips should produce little noise.

It's possible that the best answer to running 9V stuff from phantom power is "try to figure out some other way".

In fact I do think that it is a compromise, but sometimes a compromise is good enough. It appears to work decently for Tech 21 (and apparently the BDDI features three TL074s and a 4558, so not an insignificant current draw).

[Rob Strand]

In a proper desk there should be at least 10mA available. Usually we do indeed drop it across a resistor when required, with a Zener to regulate. No special effort to make it 'efficient'. Why is that a problem in your case? You could use an emitter follower to buffer a fairly high-resistance divider, if you prefer.

Well, that's the problem. On the field, you can find all sorts of "desks" (sometimes it's just a cheap mixer). Say you want to power a bass preamp (a la Sansamp), you need all the juice you can get.

A switcher is probably a bad choice because phantom power is also passing signal so any supply noise generated is going to backfeed into the signal lines.   Ideally it's common mode and it cancels but as the frequencies increase balancing gets worse and so does rejection.

If you have 48V with 2x6.8k feeds from the desk then at 10mA load the voltage at the remote end will be 48-10mA * 6.8k/2 = 48-34 = 14V.  However the open circuit voltage can be as low as 44V which leaves only 10V.

If your circuit wants 9V then with a source voltage of 10V the two supply resistors at the feed end cannot be more than 2*(10-9)/10mA = 200 ohm.    That's going to load the signal lines.

The thing is the desk can supply 10mA max but the consequence if trying to use all that current is extra loading on the signal lines.

The other way to spin it is once you set the signal loading that automatically sets the maximum current.  Lets say we want 2k loading, 2x1k resistors, then with a 44V supply the maximum current is only (44-9)/((1k+6.8k)/2) = 9mA.

One way out is to let the circuit run at lower voltages, say 7V, then with 44V and 2x1k's we can get (44-7)/(((1k+6.8k)/2) = 9.5mA.   We would need to drop the supply down to 5V to get the full 10mA.   (Clearly diminishing returns.  Also the assumption of 48V nominal vs 44V min makes a difference to the maximum current available.)

The thing about using the zener shunt regulator Merlin suggested is when the input supply is 44V the zener current can drop to zero but when the input supply is 48V the excess current will pass down the zener.   So it only wastes current if there is current to waste.   The zener is only there to cap the voltage.

[PRR]

~5mA, IIRC

Don't recall. Compute. Ideally you have 14mA at zero voltage, or 48V at zero current. Max available power is near 24V at 7mA. A useful condition is 12V at 11mA.

Assuming proper P48. P24 and P12 had their days, but cheap Asian microphones obsoleted most of the odd-voltage standards and hacks.

[Rob Strand]

I was going to post this circuit but I held back because it can cause some very weird clipping issues when you start pulling a lot of current from the phantom supply - especially when you are operating around the maximum current.

The idea is to pull current from the Phantom supply without much AC loading at all.   It does that by using current sources to feed the remote power rail.

See figure 4.3.   The circuit is Q1, Q2, R9, R10 and the circuit below it.  Rod's current is a little ill defined and risky as it depends on the hFE of Q1 and Q2.   The way I was thinking about it is to have the zener going to the base and having lower R9, R10 values.

https://sound-au.com/articles/p-48.htm

Anyway it's FYI and maybe not solving the problem you are trying to solve.