Originally appeared in the July'97 issue of EQ magazine

Tweaked on 5 April'98, 17 June'99 and 26 October 2000

copyright 1997, 1998, 1999 and 2000 by Eddie Ciletti

The terms "High-Voltage" and "Phantom Power" may sound scary, but don’t be afraid. A little caution, a dash of respect and this do-it-yourself project will come alive.  Even so, a few people have had trouble getting it right the first time 'round.  Please remember that this is a voltage doubler circuit and as such, is a bit unconventional when compared to full-wave and half-wave designs.  I have added a schematic drawing plus you can also refer to the Altec mic preamp power supply since the transformer and diode connections are identical.

Please do not connect the supply to a microphone until the voltages are confirmed ! ! !

The point of this project was to use as many off-the-shelf parts as possible.  The heart of the power supply is a standard 24 volt transformer available from radio shack.  You could start with a 48 volt transformer, or wire two 24 volt transformers in series, then rectify, filter and regulate, but a single off-the-shelf transformer is the approach I chose.  It's about a $5 part.  Remember, from a standard wall outlet, power transformers are processing an AC signal, that's an Alternating Current in the form of a sine wave -- 60 Hz in the US and Canada and 50 Hz in other places.  Hz is an abbreviations for Hertz, a gentleman who experimented with electricity.  Hz was formerly know as "Cycles per Second," or CPS.


The concept of voltage-doubling takes just a little extra thought.  In the photoschematic below, take a close look at Diodes D1 and D2 you'll see two mini-scope traces showing how the diodes are rectifying the AC sine wave.  D1 passes only the positive half of wave A and D2 passes only the negative half of  wave B.  The "trick" is that both diodes connect to the same "leg" of the transformer (the orange and blue wires).  The opposite, pink "leg" is common to the two capacitors, C1 and C2.  Measure the DC voltage from this "common" junction and you'll see +37.5 volts DC at point A and -37.5 volts DC at point B.  Add 'em up and you get 75 volts, much more than the 48 volts required.  That's what all the other "stuff" is about. 

The raw, rectified 75-volt DC is neither "clean" nor regulated power.  Zener diodes CLAMP the juice to 48 volts cleaning it up alot. PLEASE understand that these voltages are not exact.  The parts and the power in your area will never yield exact values with the exception of the zener diodes.  A ZENER DIODE is a diode operated in the reverse-bias mode.  When operated in a safe range, a Zener Diode provides the reference voltage and, if only one microphone was involved, that's all you need for regulation.  In this case, the supply is being used to power several microphones so the transistor "sees" the 48-volt reference but does not "tax" it.  The transistor gets the raw 75-volts and like a valve, passes the amount it sees as reference, less .6 volts.  So, in order to end up with 48 volts, the reference must be 48.6 volts.  This is not a sophisticated power supply, but  it does effectively nail the raw DC voltage down to a specific, constant and clean power source.

Also understand that most microphones can operate on 9volts to 48volts but you never want to feed more than 48 volts into any mic and you want to be very careful to make the correct connections.  Click here to learn how to correctly connect a phantom power supply to a microphone.


Below is a "photo-schematic" detailing the layout, built as it might be drawn from schematic.  Included are oscilloscope captures that show what half- and full-wave rectification look like.  The letters A, B and C in circles designate the test points for both the oscilloscope and voltmeter probes.  Some visitors have found the image a bit hard to follow, so a "real" schematic, sans transformer, is now included below.

The Photo-Schematic above and a "real" schematic below.  Please note the addition of C4.


All audio circuits require DC power before amplification can occur. Like an audio signal, the power that flows from a wall socket is also AC, specifically, a 120 volt RMS sine wave. (Line voltage varies with location and demand. RMS is the "effective" or equivalent power a sinusoid can deliver when compared to its DC counterpart. Actual peak-to-peak voltage is 343 volts.)


The power cable is attached to a switch and a fuse. The transformer converts 120 vac to 25 vac. Notice that there is no physical connection between "primary" and "secondary" windings. The "connection" is made by induction: the flow of current in a wire radiates a magnetic field into the iron core (the vertical lines). (Guitar pickups "gather" noise — hum from power transformers and buzz from light dimmers — also by induction.)  The ratio of turns between primary and secondary — in this case, 4.8 to 1 — will determine the voltage.


The conversion from AC to DC is called rectification. Logic would dictate that 48 volts AC makes 48 volts DC. This is almost correct. I chose a 25 volt transformer because you can buy one at Ray’s Shack. A little sleight-of-hand will squeeze out the necessary juice…

The oscilloscope insert to the right of the transformer displays a slightly clipped sine wave of about 75 Vp-p. Diode D1 passes only the positive portion of wave "A." Notice that diode D2 is reversed so that it passes only the negative portion of wave "B."  This circuit configuration is known as a "voltage doubler."

Filter capacitors C1 and C2 smooth out the "ripple" that results from rectification. (See the ‘scope insert to the lower right of the xfmr.) Traces "A" and "B" were measured by first placing the ground lead at test point "B," measuring "A."  Then, leads were reversed so that ground is at "A," measuring B."  Notice that the frequency of the "A" and "B" traces is double that of the "C" trace.  This is full wave rectification.  Unfiltered, it yields a 120 Hz hum.


With the scope’s ground lead at test point "C," the wave form observed at test point "A" will yield trace "C."  (The polarity will be reversed when observing test point "B.")  This is what half-wave rectification looks like.  If the circuit did not continue beyond capacitors C1 and C2, the power supply would yield plus and minus 35 volts referenced to "C," or, 70 volts when measured from "B" to "A."

The relationship of 35 volts DC to 25 volts AC has to do with the fact that a sine wave is involved.  Dividing 25 volts RMS by ".707" (35.36) and then multiplying by 2 (70.72) yields the peak to peak value of the sine wave. (The square root of 2 — 1.414 — divided by 2 yields .707, hence R - M - S or root - mean - squared.)


Z1 and Z2 are 24 volt, zener diodes. When reversed biased (as shown) a zener diode acts as a DC voltage limiter.  Placing two in series creates the 48 volt reference.  R1 (2.7kW) sets the "knee" so that the circuit will maintain regulation even if the line voltage drops to 92 volts.  Increasing R1 will raise the knee and decrease regulation.  Decreasing R1 will overheat the zeners.  A side-effect of regulation is that it further reduces the ripple to almost a straight line.


Q1 is an NPN transistor in the "common base" configuration. It is a general replacement type, either NTE or ECG 210.  Its American counterpart is a 2N6551.  Raw 75 volts is fed to the collector (c) which serves as the input.  The reference voltage is connected to the base (b) and the output appears at the emitter (e) six-tenths of a volt lower than the reference. This drop is typical of a silicon junction device.


When the power line delivered 122 volts, 28.2 Vac appeared at the secondary. This became 75 Vdc, rectified. The reference, 49.5 volts, was a bit higher than expected and yielded 48.9 volts (at the emitter) with no load.  R3 serves as a current limiter (protection), but without it, the circuit can easily drive a 1kW load without budging.

.5 amp
25 volt 1 amp
D1, D2
C1, C2
220mF63 volt 
.1mF, 100
across both zeners
2.7kohm, ½ watt 
1.5kW, 1 watt 
current limiting
470 ohm
this serves as a "fuse"
NPN transistor 
2N6551, 52 or 53
ECG or NTE 210
Z1, Z2
zener diode
1N5252, ½ watt 
ECG or NTE 5031A
Fuse Holder
Terminal board
wire clamp
wood screws 


If you encounter problems, start with only the transformer, diodes and capacitor connected.  The voltage should be about 75 volts from diode to diode, that is, across both caps.  Don't confuse the common connection with the ground connection.  Measured from the common connection, the supply -- without the zeners or transistor -- will generate plus AND minus 37 volts.  The MINUS becomes ground so that makes 75 volts in the positive direction only.  The common connection is only for the two capacitors and one leg of the transformer.

The next step is to connect the voltage reference created by the two zeners diodes, each should have 24 or 25 volts across each, totaling 48-50 volts.  Also, the resistor feeding the zeners will be warm but not burning hot.  All this BEFORE the transistor is connected. The cap (C4) is across both diodes to ground.

The 2N6551 should cross refence to an ECG / NTE 210.  There is nothing special about this transistor except that it must handle at least 90 volts.

Once the correct transistor is connected, it should feed the 470 ohm resistor (R3) as well as the resistor feeding the LED.  Make sure the transistor you use is properly connected, noting that the pinout may be different from the unconventional type I used.  Most modern transistors are, from left to right, B-C-E and not C-B-E as shown.

Since I built this supply from parts laying around my shop, even the LED I chose required a bit more current than more "modern" versions.  You can use a larger value of resistor for the LED.   Since most LEDs need 10mA to 20mA to fire, measure the LED current.  If too high, the LED will explode.  Use a Volt-Ohm-Milliamp meter (VOM) or use Ohm's Law.  The LED requires 1.2 volts, so that's 48 - 1.2 = 46.8.  I = E / R so 10mA = 46.8 / R (10mA=.01A)  R = E/I = 46.8/.01 = 4,680 ohms OR 2,340 ohms if the LED requires 20mA.

Please, please, please be careful, ok?  I am not responsible for ANY damages!





E-mail Eddie (He's a lonely guy after staying up late uploading all these articles!)