Premise: This article is not exhaustive, but focuses on the issue of radios that have an output transformer with a hum-cancelling tap. It assumes that the radio has already been fully repaired and is free of circuit problems, yet still exhibits an annoying hum in the speaker.

Hum in the speaker: what not to do…

Absolutely do not install oversized electrolytic capacitors in the radio! I often find 100, 220 or 330 µF capacitors installed inside vintage radios. This is something you should NEVER do, because the radio does not use silicon diodes as rectifiers, but a rectifier tube, which cannot withstand the charging peaks of such large capacitors. This could lead to premature tube wear or, worse, an internal discharge that might damage the power transformer!

What is the hum-cancelling winding?

Let’s look at this diagram: it shows a rather common situation, especially in radios from the 1950s and 1960s. The smoothed anode voltage coming from the rectifier tube enters the “I” terminal of the output transformer and flows through the winding up to the “F” terminal, reaching the plate of the output tube. The section of winding between “I” and “L” is called the “hum-cancelling winding”; in reality, it consists of a variable number of turns, usually between about 5 and twenty. What is it for?

At the time, producing large-value capacitors was not economical. Besides being expensive, they were also bulky, and most importantly, rectifier tubes could not deliver such high currents without failing—especially in inexpensive receivers, which for commercial reasons did not include filter chokes. The hum-cancelling winding is a short section wound in opposition to the primary, just like in a push-pull circuit, but much shorter and reversed. At the output of this winding there is always a resistor (470 ohm in the example schematic), which then connects to the second filter capacitor (50 µF in the image), from which all other tubes in the radio are powered. Since the supply voltage coming from the rectifier contains a certain amount of ripple audible in the speaker, the current flowing through the hum-cancelling winding induces in the core a ripple opposite to that generated by the primary, effectively silencing—or more precisely—“neutralizing” the hum. This neutralization is a matter of very fine balance, because the amplitude of the signal induced by the hum-cancelling winding must be exactly equal to that induced by the primary. If it is higher or lower, the hum will still be audible in the speaker, not from the primary but from the hum-cancelling winding itself!

At the time the radio was built, the ratio between the number of turns in the hum-cancelling winding and the series resistor was carefully calculated and tested. However, after so many years, the change in the value of the resistors in the radio, combined with variations in tube efficiency, has altered this balance. We must also consider cases where the output transformer has failed and been replaced with a non-identical part. For example, an output transformer for an EL84 might be installed which, while electrically compatible, does not have the same number of turns dedicated to hum cancellation, and therefore may not properly neutralize the hum, leaving it audible.

The same can happen even when using my universal output transformer…

This transformer is equipped with a hum-cancelling tap, but since it is universal, it has been made with a predefined number of turns. So what can be done when the hum-cancelling system does not work properly? To adjust the hum neutralization, you need to act on the resistor connected in series with the winding. In the example image, a value of 470 ohm is shown, but in reality, you can find very different values: 600 ohm, 1 k, 1.2 k, 1.3 k, and so on.

The simplest way to determine the correct value of this resistor is to temporarily disconnect it and connect a rheostat in its place (emphasis: a rheostat, which must be able to handle some current and power dissipation—so not a carbon potentiometer). Turn on the radio with the volume at minimum, so you only hear the 50 Hz hum in the speaker, and adjust the rheostat until you find the point where the hum almost completely disappears. Then turn off the radio, measure the resistance value of the rheostat with a meter, and replace the fixed resistor in the radio with one of a value very close to the measured one.

Here’s how I made this very handy little tool…

The materials I used are: a 1 k LESA rheostat, a 1 k 5 W fixed resistor, a switch, and a small box 3D-printed to fit. If you have a 2 k rheostat, the resistor and switch are unnecessary. I only had a 1 k rheostat and, to use what I already had at home without buying extra parts, I built it this way. The box is essential, considering that the rheostat carries the anode voltage, and handling it uncovered could be dangerous.

Box closed, a piece of wire, two alligator clips, the ugliest knob I could find, and the gadget is ready to use. It can also be useful for other purposes.