Power Supplies: Which is More Quiet?

Some people believe that modern switching power supplies and DC-DC converters are noisier, and that using a traditional larger AC power supply is going to be quieter. Others think that using a battery and isolating the pedalboard from wall power is going to avoid AC noise, ground loops etc, and so this will be quieter. So, which is correct?

To find out, I did some tests connecting the same pedalboard to a high-quality AC reference power supply, a low cost generic switching power supply I got on the internet, and a Mission 529 with both battery and wall power, and measured the noise to compare them. Let’s find out what happened.

Just before the turn of the 19th Century, engineers and industrialists were trying to figure out how to get a practical electrical supply into homes and businesses. Electricity was anticipated to be a cleaner, safer, and more reliable source of energy to replace candles and gas lighting in residences, and steam powered machinery in industry. In the US, a battle of technology and business took place between Thomas Edison, proponent of Direct Current, and Nikola Tesla, and George Westinghouse, pioneers of Alternating Current.

The principle challenge was that low voltage DC, such as from a battery is ideal for small devices and local power, but a significant amount of energy is lost to heat when transferring over distance in cables. The voltage needed to be raised to much higher levels to be efficiently sent over a long distance, but this is hard to do with Direct Current. Tesla and Westinghouse developed Alternating Current which is much easier to convert between different voltages using simple transformers, and this is the key reason this won out over DC.

This is the system we still use today. AC is generated in large power stations in industrial areas, stepped up to high voltages; sometimes hundreds of thousands of volts for transmission in power lines around the country. Then it’s stepped down again a few times, eventually to the hundred or two volts at the wall outlet. Then we often convert it to DC for use in our small devices such as guitar effects pedals.
For the tests I used Pedaltrain Nano with a mixture of small analog and digital pedals. I added to the board an iRig Pro to provide the USB audio interface between the pedalboard and a PC to do the noise analysis. My thanks to the folks at IK Multimedia for providing the iRig Pro to test.

For the AC power supply, I used the MXR MC403 power system. We use these in the Mission lab as our reference power supplies because of their good performance. This is a linear AC DC power supply. The wall power plugs directly into the side of the unit. For the low-cost power supply, I used the AGPtek CP-05. I purchased this on Amazon for around $30. This one uses a wall wart to convert the AC to 18V DC to power the unit. To test DC, I used a Mission 529 which uses a lithium ion rechargeable battery. The 529 can use any USB power source, so I tested this with a wall wart too, to see if there is any difference.

The MXR uses internal transformers to drop the voltage, and rectifiers to convert to DC. With AC supplies we are looking for issues with 60 cycle hum. Unlike DC where current flows continuously in one direction, AC oscillates back and forth. This is what allows AC to be easily transformed between voltages, thanks to the properties of electro-magnetism. In the US, wall power oscillates at 60Hz. In some other countries, it’s 50Hz. Unfortunately, those same properties that allow transformers to work, can also cause electro-magnetic interference. We sometimes hear it as a hum in audio systems.

Direct current does not oscillate, but we have another problem: Converting DC voltages. Battery voltage is determined by its chemistry; for example, 1.5V from an alkaline cell, 1.2 for Nickel, 3.7 for Lithium-Ion, etc. Transformers don’t work for DC, so to provide other voltages we use modern integrated circuit based DC-DC converters. A key mechanism behind how these work is the switching of current flow on and off quite quickly using transistors. By controlling the on and off times (called the duty cycle) the switching converters can easily convert between voltages.

The trade -off is that now we are no longer just providing a continuous current flow in one direction, but are switching current on and off through inductors. This can create a similar issue with noise from electro-magnetic interference as we had with our AC supply. The main difference is in the frequency. Switching noise is generally a higher frequency and sounds more like a whine or whistle compared to the lower frequency hum or buzz from AC. A good power supply design will filter these out as much as possible, so let’s go measure these, and see how they do.

Here’s the first result. The pink trace is the AC linear supply with large toroidal transformer, and the blue trace is the 529 switching supply with a USB battery. The response is almost identical except the AC supply has a small amount of extra noise at 60Hz, which is our expected 60 cycle hum. The switching supply has nothing at this frequency because no AC is present. So, right there, the suggestion that DC switching effects pedal power supplies are inherently noisier is totally busted. Both of these are very quiet. Even the 60 cycle peak on the AC supply is at -98dBu which is slightly less than the -95 dBu at the very top of the noise floor.

OK, batteries seem to work, but what about using the 529 with a wall power supply? Surely a wall wart is going to cause lots of switching noise?

We’ll that’s busted too. Here the green trace is the 529 with a decent quality USB wall mount power supply, and although we see a little 60 Hz noise starting to creep in now, it’s still as good or better than the linear supply.

So, what’s going on? Don’t some people get noise in their rigs when using switching power supplies? A conspiracy by makers of expensive transformers?

Well it’s really just a matter of getting the best thing for the job. Here the orange line is the low cost power supply, and we can definitely see some increase in noise. It’s not terrible, but it’s there. Some can be much worse but switching converters that are properly designed for audio use filter this out or move into frequency ranges outside the audible spectrum.

This last scan compares the 529 with the MXR, but this time with all the effects pedals switched on. Here we can see that any small difference in noise performance of the power supplies is wiped out by the increase in the noise floor once we enable a few effects. In guitar rigs, there is often so much noise from amps, effects, pickups etc that power supply noise is going to be the least of your problems.