ESP Logo
 Elliott Sound Products Project 04 

Power Supply For Amplifiers

© 1999, Rod Elliott - ESP

I strongly suggest that the reader has a look at the article on power supply design for additional background and far more information than provided here.


In some countries it may be required that mains wiring be performed by a qualified electrician - Do not attempt the power supply unless suitably qualified.  Faulty or unsuitable mains wiring may result in death or serious injury.  All mains wiring must use mains rated cable, segregated from input and low voltage wiring as required by local regulations.


A power supply suitable for use with the 60W amplifier presented in the P3A article is perfectly simple, and no great skill is required to build (or design) one.  There are a few things one should be careful with, such as the routing of high current leads, but these are easily accomplished.  This article shows the general form of a cost-no-object version, but it can be simplified.

The first thing to choose is a suitable transformer.  I suggest toroidal transformers rather than the traditional 'EI' laminated types because they radiate less magnetic flux and are flatter, allowing them to be installed in slimmer cases.  They do have some problems, such as higher inrush current at switch on, which means that slow blow fuses must be used.

For the 60W amplifier, a nominal (full load) supply of ±35V is required, so a 25-0-25 secondary is generally ideal.  The circuit for the supply is shown below, and uses separate rectifiers and capacitors for each channel.  Only the transformer is shared, so channel interactions are minimised.  A single ±35V supply (i.e. using only a single bridge and set of filter capacitors) will work just as well in the majority of cases.

Figure 1
Figure 1 - ±35V Power Supply

  The 5A slow-blow fuse shown is suitable for a 300VA transformer, if a 120VA transformer is used, this should be reduced to 2.5A (or 3A if 2.5A proves too hard to get).  If you are even a little bit concerned about the fuse rating, contact the transformer manufacturer for the recommended value for the transformer you will use.  The correct fuse is critical to ensure safety from electrical failure, which could result in the equipment becoming unsafe or causing a fire.  The value also depends on the supply voltage where you live.  It may need to be a higher rating for 120V mains.

C2 (100nF X2 rated) is intended to minimise EMI (electromagnetic interference), and in particular conducted emissions.  It can be a higher value if you prefer, but more than 470nF isn't necessary.  Some people like to add low value caps in parallel with the diodes in the bridge, but this should not be needed.  They do no harm, but make sure the caps you use will handle the AC waveform without failure.

The capacitance used is not critical, and is somewhat dependent upon one's budget.  I suggest 10,000µF capacitors, but they are rather expensive so at a pinch 4,700µF caps should be fine - especially in the arrangement shown.  An alternative is to use (say) 5 × 2,200µF caps in parallel for each main filter cap.  This is often cheaper, and in many cases will actually have better performance.

When unloaded (or with only light load), the voltage will normally be somewhat higher than 35 Volts.  This is Ok, and should not cause distress to any amp.  The voltage will fall as more current is drawn, and may drop below 35V if a small transformer (or one with unusually poor regulation) is used.

Figure 2
Figure 2 - Dual ±35V Power Supply

Some constructors may prefer a 'dual mono' power supply, but using a common transformer.  This is shown above.  One thing that it vitally important is to ensure that the earth/ ground between the two sets of capacitors is as solid (electrically) as possible.  If there is any appreciable impedance between the ground points, this can lead to a ground loop, and hum/ buzz will be the result.  The ground connection between the filter capacitors is critically important!

Two parts of these circuits are critical:

When wiring the bridge rectifiers to the transformer, connect exactly as shown to ensure that ripple voltages (and currents) are in phase for each amp.  If not, mysterious hum signals may be injected into the amp's signal path from bypass capacitors and the like.  This is somewhat unlikely unless huge caps are used on the amp board(s) - not recommended, by the way - but why take the risk?

Bridge rectifiers should be the big bolt-down 35A types (or something similar) to ensure lowest possible losses (these will not require an additional heatsink - the chassis will normally be quite sufficient).  The transformer primary voltage will obviously be determined by the supply voltage in your area (i.e. 120, 220 or 230) and be suited to the local supply frequency.  Note that all 50Hz transformers will work just fine at 60Hz, but some 60Hz devices will overheat if used at 50Hz.

The transformer should be rated at a minimum of 120VA (Volt-Amps) for home use, but a 300VA transformer is recommended due to its superior regulation.  Going beyond 300VA will serve no useful purpose, other than to dim the lights as it is turned on.

Where it is possible, the signal and power ground should be the same (this prevents the possibility of an electric shock hazard should the transformer develop a short circuit between primary and secondary.  Where this will give rise to ground loops and hum in other equipment, use the method shown.

The resistor R1 (a 5W wirewound resistor is suggested) isolates the low-voltage high-current ground loop circuit, and the diodes D1 & D2 provide a protective circuit in the event of a major problem.  These diodes need only be low voltage, but a current rating of 5A or greater is required.  The 100nF capacitor (C1) acts as a short circuit to radio frequency signals, effectively grounding them.  This should be a device with very good high frequency response, and a 'monolithic' ceramic is recommended.

In some cases, the transformer secondary voltage may need to be higher than described above.  I tested some stock and custom transformers I have, and found that unless the transformer has extraordinarily good regulation, a nominal 28-0-28 secondary can be used.  This will provide supply rails of around ±40V, which is the highest recommended for P3A (for example).  Be careful when you test, since a relatively small (10%) variation in the mains voltage makes a big difference to measured output power - the secondary voltage also falls by 10%, so 60W becomes 48W if the mains is 10% low.

You also need to remember that the output voltage of transformers is typically quoted at full power with a resistive load.  This means two things:

1.   The no load voltage will be higher than expected
2.   The loaded voltage will be lower than expected

The first point is true because there is no loading, so the output voltage must rise.  The second is more complex, but happens because the conventional rectifier circuit uses a capacitor input filter (the rectifier feeds directly into the capacitor(s)).  Since the diodes only conduct at the peak of the waveform, the current is much higher, so the transformer and supply line impedance will cause the peak voltage to fall, and the DC voltage cannot exceed the peak output voltage (less two diode forward voltage drops).


HomeMain Index ProjectsProjects Index
Copyright Notice. This article, including but not limited to all text and diagrams, is the intellectual property of Rod Elliott, and is © 1999.  Reproduction or re-publication by any means whatsoever, whether electronic, mechanical or electro-mechanical, is strictly prohibited under International Copyright laws.  The author (Rod Elliott) grants the reader the right to use this information for personal use only, and further allows that one (1) copy may be made for reference while constructing the project.  Commercial use is prohibited without express written authorisation from Rod Elliott.