Best way to provide 48v phantom supply for replacement audio mixer psu

[uslrs]

I'm repairing a Soundcraft EM6 mixer which had a fault that destroyed the switch mode PSU.  The original psu, apparently prone to catastrophic failure, fits beneath the main board in a relatively small space.  I'm happy to construct a linear psu, employing a toroidal transformer, which needs to provide 15v - 0 - 15v at approx. 1A , but I'd like advice on how to provide the phantom 48v for the six mike channels (I'm guessing 20mA might be adequate).  There are boost converters available for which I could use the 30v as input, but I'm concerned the converter might introduce unwanted noise.  Any recommendations would be very welcome.

[Audiorepair]

If this is your own mixer, for your own use, you could probably get away with less than 48v for phantom power, supplied from the same secondary that supplies the 15 volt rails.  Most mics don't need the full 48v.

Dynacord Powermates (powered mixers) actually use 24v as their phantom power.

[uslrs]

Many thanks Audiorepair, I'm expecting a rectified voltage in excess of 30v, using + & - regulators for the main board supply, so there may well be something approaching 40v available for the phantom supply.  I was assuming the phantom supply voltage had to be 48v, plus or minus something under 1v.  If I can get away with something over 35v I'm in business!
 

[Audiorepair]

The standard for phantom power is actually11v - 52v, though I would not suggest 11v should be enough to drive every mic or DI box out there!


Edit:  There are/were lots of different standards, so my quoted 11v - 52v may not actually be correct, but the gist is the actual voltage is not critical, and in your case you can suck it and see.

[Ian.M]

A capacitively coupled bridge could double that then drop it to 48V with a HV linear regulator.  As long as the input is over 60V Microchip's LR12 would be suitable.  If its  under 60V^*, a LM317HV would do the job. 

* A LM317HV can be used at higher input voltages if you put a TVS diode across it to protect it from over 60V input to output differential during power-up and fault conditions, and the output is protected against being shorted to ground.

[The Soulman]

Many thanks Audiorepair, I'm expecting a rectified voltage in excess of 30v, using + & - regulators for the main board supply, so there may well be something approaching 40v available for the phantom supply.  I was assuming the phantom supply voltage had to be 48v, plus or minus something under 1v.  If I can get away with something over 35v I'm in business!
 

No, your phantom power is referenced to gnd in this case the transformer center tap, so maybe 20V max unregulated depending
on the transformer and line voltage.
I'd use a small dc to dc converter perhaps followed with a lm317, depending on any rc filtering that maybe already in place.
Or a dedicated transformer with rectifier, cap, and voltage regulator.

[magic]

All sorts of diode capacitor networks can boost AC voltages at low load current.

https://www.allaboutcircuits.com/textbook/semiconductors/chpt-3/voltage-multipliers/

[Audiorepair]

Have a look at the Dynacord circuit:
https://elektrotanya.com/dynacord_powermate_600_powered_mixer_sm.pdf/download.html

They use a couple of caps to generate 50v from the same secondary the 15v rails use, then for some reason regulate that down to 24v for their phantom.

Note that you should really alter the values of the resistors going from phantom voltage to the 2 XLR pins if you are using much lower than 48v.
Usually 6.8k resistors are used with 48v, though the value is again not critical, Dynacord use 2k4 with their 24v.
These resistors should be closely matched, however, so 1% is probably a good idea.

[Messtechniker]

These resistors should be closely matched, however, so 1% is probably a good idea.

Neumann has specified here resistor matching <0.4% for decades.
Certainly no problem today.

[Audiorepair]

These resistors should be closely matched, however, so 1% is probably a good idea.

Neumann has specified here resistor matching <0.4% for decades.
Certainly no problem today.

Slightly off topic but this has got me thinking:  what is the typical variation between resistors on the same reel?
I guess there won't be any separate manufacturers specs on this, just the overall tolerance to the resistor value itself?

[uslrs]

Thanks guys, really appreciate your help.  I'm now in the process of repairing the main board where there has definitely been emission of the magic smoke!  I can test after repair using bench power supplies, so onwards and upwards!!

[Cerebus]

The standard for phantom power is actually11v - 52v, though I would not suggest 11v should be enough to drive every mic or DI box out there!


Edit:  There are/were lots of different standards, so my quoted 11v - 52v may not actually be correct, but the gist is the actual voltage is not critical, and in your case you can suck it and see.

IEC 61938 is the 'holy writ' on phantom power and there are 12V, 24V and 48V versions of it. The 48V version (what most people expect and mean by phantom power) requires a minimum 14V at 10mA to be available at the mic end of things (post the 6.81k feeder resistors). Hardly anything is as strict as that in practice at the mike end (I've got 48V mikes that will happily run off 9V) but to trust that always happens might be foolish. You don't want to be running around diagnosing a phantom power problem 30 minutes before the main act comes on stage because someone has plugged in something that relies on your having met the standard.

I'd go with the suggestion to use a classic voltage doubler/tripler to get the 48V rail. You need next to no real power available, so it ought to be possible to get all you need with only a moderate sized smoothing capacitor. It would be wise to put some over-voltage protection in - an amplified zener or even a beefy zener clamp with a current limiting resistor in the supply would do it.

[magic]

Slightly off topic but this has got me thinking:  what is the typical variation between resistors on the same reel?
I guess there won't be any separate manufacturers specs on this, just the overall tolerance to the resistor value itself?

My very limited experience:

I bought 20 identical metal film THT resistors from Vishay, 1% tolerance and 50 or 100 ppm/°C tempco - the most basic stuff.
My aim was to select a few with ±0.1% match over typical room temperatures. I compared them at 18°C, 23°C and 28°C.
IIRC they all satisfied my matching tolerance at all temperatures, but also all were off from their specified value by more than 0.1%.

[uslrs]

Thanks Cerebus - it's good to get some 'hands-on' feedback in relation to the phantom supply.  Unfortunately, this now seems to be of academic interest since having repaired the board and replaced the NE5532's that were incinerated, it appears the demise of the power supply also took out transistors at the input end of things.  Almost zero accessibility without desoldering multiple jack & SLR sockets, etc, so pretty much a write off by the looks of it 

[Audiorepair]

These pre-amps circuits often have 4 protection diodes running from the negative of the input electrolytics to the
+/- 15v rails.
These do get damaged and dc bias the pre-amp so it won't work. 
There could be a chance this is your problem, and you can find the dodgy one(s) with a meter as they do usually read bad or different in circuit.

You could just knock them off the board if you can get to them, they are not required to make the circuit work.

[uslrs]

Thanks Audiorepair, I was hoping someone with experience of this board might be able to point me in the right direction.  I have removed the XLR socket to get access as multimeter tests were a bit inclusive with the diodes in circuit.  I removed this pair of diodes (D703/D704 between input and negative rail) and they both test fine out of circuit, so I'm of the opinion the transistors in the input circuitry have given up. When I inject a signal into the line socket nothing seems to get past TR700/TR702.  I guess I'll have to bite the bullet and remove jack sockets to get to the offending articles so I confirm that diagnosis.   Thanks again for your help.

[uslrs]

This is the area of the board.

[Audiorepair]

Ah, the 4 diodes I have seen in the past are implemented slightly differently.
Two SMD single diodes on each cap going to both of the 15v supplies.

Looks like this time they wanted to save a fraction of a cent.