Chasing Hum

When hum develops in a piece of production equipment, it is a technician's task to trace it.
In that case, the hum may be due to a dried-out electrolytic, a defective tube or a leaky
power transformer which has to be traced. But when you have a new piece of equipment
or a prototype and it hums, the job often proves to be one of chasing.
It can be a chore. The senior engineer may not regard it as engineering or he may not want
to be bothered with it…let a junior engineer or technician fuss with it. He's already
engineered the thing and it works. It's just a little hum that needs to be
eliminated…anyone can get rid of hum, from his perspective.
All too often the junior engineer runs into difficulties. The books don't help him much but
he's tried everything he's learned about hum. Some of the things he does may vary the
hum somewhat but none of them stops it, except removing all the tubes or switching off
the power.
He's checked for ground loops, magnetic and electric or electrostatic induction, humming
tubes…etc., whatever he tries, the hum's still there.
Changing the value of one component inexplicably reduces the hum. There is no valid
engineering reason why it should but he accidentally found out it does. So he makes the
change. The hum is still unacceptable so, he continues to work on it; now, he finds the hum
is worse than ever and he cannot even repeat results as good as when he started working
on it.
Sound like a familiar story? The lesson is that hum chasing is an engineering task, not to
be pursued in an aimless or frantic manner. Approaching it from an engineering basis
does not eliminate the "chasing" aspect of the job but it can eliminate the hum.
Keep in mind that when you are working on new equipment that has never been "de-
hummed," you may be tracing several sources of hum; at one point in your endeavor, you
may try additional filtering in the B+ supply and discover that it makes no difference or it
may increase the hum slightly however, don't rule-out the B+ filtering as a possible
contributing factor to your problem.
All hum comes from the same ultimate source: the power supply. It has the same
frequency or harmonics, with many addition and cancellation occurring down the line.
When you provide additional filtering capacitance, the ground point you choose for the
extra connection may inject some ripple-current into the signal circuit. So, while you may
be reducing the ripple on the B+ supply, you may be inducing as much or more hum to a
signal circuit and the net effect is no change or an increase in hum.
It may be you are eliminating one source of hum however presently, it is not your major
source; cancelling some of the major source, makes little difference or even makes it
louder. Later, after you have eliminated what proved to be a greater source of hum, you
perform the same test and discover that now it produces a dramatic reduction in hum.
Until you have eliminated the hum, do not accept any conclusion as final.
A change that should reduce hum may just change its harmonic content or increase its
magnitude; you will have more than one type of hum to trace down; it’s doubtful that a
simple change will get rid of all of it.
It is helpful to view the hum on a high-gain scope, as well as listening to it. This will help
identify the source and often will help identify interacting effects of the type of hum.
Sometimes a circuit change may not reduce the magnitude of the hum but it may change
either the harmonic composition or its phase relation to the line frequency. Harmonic
composition can be identified much more readily by monitoring the scope than by
listening and the phase can only be recognized by checking your scope.

Fig 1: Some of the basic patterns encountered in hum chasing: a. 60Hz, b. 60Hz with 90°
phase-shift, c. 120Hz, d. 120Hz with 90° phase-shift, e. higher odd-order harmonics, f.
same with 90° phase-shift, relative to 60Hz triggering, g. spike due to half-wave
rectification, h. spikes due to full-wave rectification, all using 60Hz line-triggering.

If the equipment uses full-wave, B+ rectification with 60Hz heater supply, the
predominant frequency in the hum should provide a good lead. For an informative display,
use the scope's line-frequency time base, for triggering; the 60Hz components will
produce a slanting line or ellipse, as shown in Fig 1a or b, while 120Hz components
originating in the B+ supply, will produce a curve or figure 8, as shown in Fig 1c or d.
A third or higher order odd harmonic, as shown in Fig 1e or f, is usually due to induction
from a power transformer or motor. Higher order pulse-type hum that produce one or
more spikes on the trace as shown in Fig 1g or h, are usually due to rectifier circuit pulses
finding their way into the circuit.
This information is the source of the hum and most of these forms are introduced into the
signal circuits by a variety of routes.
A ground loop may not always produce hum, for it is possible that a ground system with
no loop, may be responsible for producing hum.
Fig 2: Basic principle of proper grounding system is to ensure that the ground "bus" is a
mirror image of the B+ distribution system.

The ground, return system is a mirror image of the B+ supply circuit as indicated in Fig 2;
the area where the supply branches, the ground line should similarly branch so, the return
currents will follow paths corresponding to outgoing currents. A decoupling cap returned
to an improper location on a ground line or to a wrong branch, as shown in Fig 3, may be a
cause of hum, due to injection of ripple current in the wrong place.

Fig 3: Incorrectly returned cap can inject hum due to improper return of ripple currents.

Pay attention to electrolytic cap returns. If necessary, use insulating wafers for mounting,
to isolate electrically, from chassis-ground. It may be necessary to change the group type
cans because the common grounding of a single can, causes trouble.
You could try increasing the size of various smoothing or decoupling caps. If you're up
against smoothing versus DC voltage drop, 2×20µfd with a couple of 10KΩ resistors will
provide more ripple smoothing than one 40µfd cap with one 20KΩ resistor, for the same
DC drop, as shown in Fig 4.

Fig 4: Two-stage filter, using same voltage drop and total capacitance, is more effective in
hum reduction.

This brings up another aspect of supply-induced hum: phase. Each smoothing element, R
and C together, produces approximately a 90° phase-shift in the ripple-voltage. Check the
phases within the circuit for the injection of residual hum at different points, to see if they
add, subtract or appear in quadrature, as shown in Fig 5. This process can be helpful in
interpreting what you see on the scope; it may also help you arrive at another supply
arrangement that will allow the hum component to buck rather than add to one another.
Heater supply hum may be induced into the signal circuits in a variety of ways, internal or
external to the tube.

Fig 5: Using phase arrows, as shown, can help trace whether sources of hum are additive
or otherwise.

Internal, it may be due to capacitive transfer either to grid or cathode. To grid, the effect
will depend on the grid circuit impedance, as with other capacitive transfers; to cathode, it
will depend on the impedance between cathode and ground.
Electronic conduction between cathode and heater, due to the heater acting as a plate with
low applied AC potential, can cause hum, either at 60 or 120Hz.
A remedy for this is to bias the heater positive so, the conduction path is maintained at
saturation, as shown in Fig 6.

Fig 6: Use of positive bias to eliminate internally created heater hum.

In series-connected, transformerless heater chains, cathode injected heater hum may be a

difficult problem; a low-impedance return is necessary in the cathode circuit and often the
best solution is to connect the cathode directly to the ground bus or chassis and use grid-
leak biasing. External to the tube, heater wiring may induce hum due to poor lead dress, as
shown in Fig. 7.

Fig 7: Heater wiring lead dress is important; although twisted throughout rest of travel,
(a) leaves a loop under the socket which (b) eliminates.

Heater wiring should be twisted and kept well removed from signal circuits; the primary
power leads may create problems too. Pilot lights, fuse connections and power switches
on the back of pots, may position AC wiring into critical areas.
The remaining possible ways for hum to enter, are by direct induction in the signal
circuits, either magnetically or electrically.
When magnetic induction is a strong component, test-shorting of a grid or other point of
relatively high impedance, will likely increase the hum and if the induction is electric, it
will virtually eliminate the hum.
A method to eliminate magnetic induction is by shielding, twisting leads and use of proper
lead dress. Electrically induced hum may also be eliminated by shielding however, to be
effective, the shielding should be complete. Sometimes this may be impractical; a
reduction in circuit impedance may be a more effective approach.
In determining circuit impedance for this purpose, remember the active elements or
reactance. Plate resistance of a preceding stage may lower impedance just as effectively as
using a lower resistance value, as shown in Fig. 8 or feedback may help. Low plate
resistance is a good reason for using triodes in preference to pentodes. Two triodes can
replace one pentode, with less susceptibility to certain hum issues.
Feedback circuits may increase, as well as reduce, the output hum because of interaction
with other factors. To determine what happens when feedback is connected, create a
complete phase diagram, as shown in Fig 5, for the hum frequencies. Then decide on a
method that includes all possible points where residual hum may be bucked, rather than
added.
Fig 8: Hum pickup due to circuit operating impedance depends on total impedance to
ground. Don't forget that active impedances are part of this. A triode (b) allows lower
impedance to ground than a pentode (a).

The main idea for hum-chasing, is to be aware of all the possibilities; there's nothing
mystic about hum, although at times it does seem to come out of nowhere.