RT-P48 - 48V Phantom Power Supply Generator

This is a design to generate standard 48V phantom power supply for microphone preamplifiers. A general purpose design suitable for retrofitting to older equipment or stand-alone use, with capability of using a wide range of input supplies, either AC or DC.

Design considerations

Is there a readymade solution? I can't find one. There are lots of things on ebay which could possibly be modified, but nothing I think that would be wanted at the front end of a serious recording studio. There is no way of cutting corners on this. Reliabilty is as important as quality of output, inevitably it will require relatively large and expensive components and very careful PCB design.

Circuit type: I think it has to be a simple boost converter for maximum flexibilty. I want it to accept a range of supply inputs, both AC and DC, so I will include a position for an optional bridge rectifier.

Switching device: Frequency of operation should be moderate, from experience I would want around 100kHz or slightly more. This is well above audio frequencies, high enough to make filtering very effective with simple circuits and standard components,but avoiding the special considerations of operating at very high frequencies. I have chosen the LT1172, an expensive device but with good performance and reliabilty. It operates at 100kHz using the standard PWM techniques. I will add a second inductive filter as suggested in the device datasheet, also a Sziklai compound transistor capacitance multiplier to the output. The input supply will also have an inductive filter. For convenience I will use the same inductor for all three positions, which is admittedly excessive.

Input supply: For a DC supply, I will specify between 5V absolute minimum and 30V absolute maximum. I know for sure the converter will run outside these limits, so that's a pretty safe specification. For an AC supply if the optional bridge rectifier is fitted, the same numbers on the rectified volts.

Output voltage: There is a widely-used loose spec for the volts, but many microphones derive their polarising charge directly from the phantom supply, so I will specify a minimum of 47.5V and a maximum of 50V.

Output current: Microphones need from about 1mA up to 7mA. 100mA is going to fit most needs. I will design for 100mA minimum capability with an input supply of 10V or more. Operation from as little as 5V will be possible but I will not specify the output current capability at that level. Output well in excess of 100mA will be possible with 15V or greater supply, but there are too many variables for me to want to specify a level. Suck it and see.

Output noise: Obviously needs to be as low as I can get it. With a switching converter, I need to keep the radiation low as well, meaning toroidal inductors.

Noise in electronic circuits is something that has been studied intensively. It is a mathematician's paradise. But it is not always easy for a designer to find a method of measuring or specifying that is appropriate for a given situation. Here we have some advantages arising from the application, the current is small and constant and is used by circuits that are expressly designed for good CMRR. Also there is a useful attenuation at higher frequencies arising from the combination of the 6k8 resistors which the current is supplied through with the microphone circuit. I will provisionally adopt a crude test method until I have given it more thought, that is whether I can see anything with my old 20MHz analog scope cranked up to maximum sensitivity.

With the fairly large components needed, it is reasonable to keep to a through-hole design. The LT1172 has built-in protection for many fault conditions, and I have not added anything of that sort. It is possible that a direct short on the output might damage the output transistor, so don't do that.

Click image for pdf.

The basic design result is very close to the examples on the device datasheet, some changes of component value were needed.
Dimensions 43 x 75 mm.

If you would like to buy a PCB, please email me.
This one is for a DC supply, so the optional bridge rectifier is not fitted.

Specifications were met. My rather inexact noise test showed a slight thickening of the trace on the 5mV/div scale, probably less than 1mV RMS.

Component selection

Under review, may contain errors. The three 100n plastic capacitors here will be replaced by ceramics. I use the plastic capacitors during development as they have somewhat higher ESL and will tend to show up any problems. However, they do seem to perform much better than would be expected here. The output noise is rather less than I have seen in many linear regulator circuits.


Farnell order code Notes
Qty, Refdes
AVX SR151A471JAR Multilayer Ceramic Capacitor, 470 pF, 100 V, SkyCap SR Series, ± 5%, Radial Leaded, C0G/NP0 1100376 1 C7
WIMA MKS2C031001A00KSSD Film Capacitor, 0.1 µF, PET (Polyester) 1006030 3 C6 C2 C1
VISHAY BFC237021104 Film Capacitor, 0.1 µF, PET (Polyester) 1215515
WIMA MKS2C041001F00KSS Film Capacitor, 1 µF, 63 V, PET (Polyester), ± 10%, MKS2 Series 1006040 1 C4
PANASONIC EEUFR1J471 63 V, FR Series, ± 20%, 12.5mm 2508149 3 C3 C8 C9
TE CONNECTIVITY LR1F100K 2329987 2 R7 R6
PULSE ENGINEERING PE53112NL 47µH 1209544 3 L1 L2 L3
Schottky diode
Multicomp SR1100 1861419 1 D1
Switching IC
Linear Technology LT1172CN8 9560181 1 U2
MULTICOMP MPSA92 Bipolar (BJT) Single Transistor, PNP, 300 V, 50 MHz, 625 mW, 500 mA 1574392 1 Q2
FAIRCHILD SEMICONDUCTOR MPSA42 Bipolar (BJT) Single Transistor, NPN, 300 V, 50 MHz, 625 mW, 500 mA 1017719 1 Q1
MULTICOMP W04G Bridge Rectifier Diode, Single, 400V, 1.5A 2675406 1 U1