AUDIO
OUT
AUDIO OUT
L
R
By Jake Rothman
Potentiometer scratching – Part 1
E
ver since electronic audio
started over a century ago,
there have always been volume
controls based on variable resistors –
potentiometers. And, right from the start
those pots suffered from an annoying
fault – when operated they cause the
audio output to go ‘scratch, crackle and
bang’. Pot scratching is the most common
analogue audio fault. Often, this is down
to potentiometers just wearing out, but
it’s also the result of poor circuit design
and parts choice. I’ve always believed the
rotary potentiometer or ‘pot’ is one of the
most effective human controls invented.
It would be a shame if the physical knob
became just another software icon in a
skeuomorphic world. This month and
next I’ll discuss how to overcome its main
weakness, rotational noise.
Potted histories
As a child in the 1970s, I was able to
earn steady pocket money fixing scratchy
volume controls in many radios and
other consumer audio products. Some
still scratched even when replaced due
to leaky capacitors and bad design, but
I didn’t know what was going on at the
time. There’s not much material on the
subject. The earliest article I’ve found
was by E Dexter in Practical Wireless
from August and September 1962. It
was mainly concerned with valve circuits. James Kerrick did an article called
Clearing Noisy Volume Controls in the
April 1974 issue of Radio and Electronics Constructor magazine. Another one
was by CH Banthorpe in Circuit Ideas,
Wireless World (April 1971).
I’ve been researching pots for 35 years and
since rotational noise is rarely mentioned
in device datasheets, real-world practical
experience is important when it comes to
parts choice and circuit design – and that
is exactly what I will write about here.
Basic construction
The construction of potentiometers has
been covered before by PE (November
2015), but it’s as well to be familiar with
the basic assembly. The best way to show
this is to pull a device apart. The Alpha
24mm pots supplied by Rapid are the easiest to disassemble – and they can easily
be put back together again. The technique
of opening up the metal tabs is shown
in Fig.1. The track, wiper and the centre
contact slip ring assembly are all visible.
Can’t stand the noise
The whole point of a potentiometer is
to provide an infinitely smooth change
in resistance with rotation. That is the
theoretical ideal, but in practice there are
always irregularities in the track resistance.
These modulate any voltage that is across
the pot and cause noise as it is rotated.
Pot technology
Manufacturers have developed quite
a few ways to create potentiometers.
The following is an overview of the
most-important varieties. Note that in
the following descriptions the ‘track’ is
the resistive material between the pot’s
two end terminals, and the ‘wiper’ is the
part that moves along the track and which
is connected to the pot’s third terminal.
Fig.1. The best way to understand the
component parts of a pot is to open one
up. Most are simply clamped together.
Just open up the tabs with side cutters.
Practical Electronics | June | 2023
Wirewound potentiometers
Wirewound pots (Fig.2.) give the best stability and accuracy, along with potentially
Fig.2. The earliest pots were wirewound.
This is an unusual dual-gang high power
tweeter level control.
massive power dissipation capabilities.
They were historically the first pots.
However, nowadays, their only audio
use is mid-range and treble level controls in loudspeaker passive crossovers,
where very low values of resistance (2Ω
to 100Ω) and high power ratings (2W to
5W) are needed. Wirewound pots have
stepped resistance changes which give a
perceptible ‘whooshing’ noise as the pot
is rotated, sometimes called zipper noise.
Carbon potentiometers
Early radio designers later moved on
from wirewound pots to using carbon
for the resistive track, employing a
similar material to carbon film resistors. Those made by Morganite (Fig.3),
who specialised in carbon brushes and
railway pantograph pickups were the
favourite, and were adopted by nascent
Hi-Fi industry pioneers such as Quad
in their 1950s valve pre-amps.
The tracks of carbon pots normally
consist of fine carbon-black mixed with a
phenolic resin and some kind of insulating
filler, such as clay. They still have irregularities, but they are of a smooth random
nature, as shown in Fig.4. Typically, they
have rapid transitions of around ±0.2 to 3%
of the resistance of the average track value
as the wiper is moved. If this parameter is
specified it is called dynamic noise. It is
always given as an initial value for a new
pot. As the track wears the surface gets
rougher and the noise increases.
63
Counter-clockwise
track end terminal
Fig.3. Very early Hi-Fi used these Morganite carbon-track pots. This
1964 germanium Leak Stereo 30 had them. They still work despite
oxidation and dirt.
Track coating recipes are generally closely guarded trade
secrets, although Omeg told me their tracks used carbon black,
phenol-aralkyl thermosetting resin, a toughing agent such as Butar
resin and a dry lubricant, such as PTFE and solvents. They call
it ‘conductive polymer’. The cheapest way to make a pot track
is to just spray or screen print a carbon-ink film onto a circle of
Paxolin or phenolic-resin-bonded paper. The best carbon-track
pots are the Alps and Alpha (Fig.5) pots that adorn most consumer
amplifiers. Carbon tracks have the shortest lifetimes typically
10,000 to 20,000 rotations. Morganite found the average radio
volume control in the 1950s was moved about 2000 times a year.
In busy professional audio studios, the main volume control or
monitor pot may only last a year or two.
Moulded track potentiometers
A longer-lasting technology is a thicker track of hot-moulded
carbon, typified by the old Plessey E series (Fig.6) which were
used in Neve modules and 1960s Bush radios, such as the
famous TR82 (shown in Fig.6). These radios are still made as
Chinese re-issues today. Resistance tolerance is the worst of all
pot types with moulded track, similar to carbon composition
resistors. However, in potential divider applications, such as
volume controls driven by low impedance and loaded by a
high impedance, variations in the track resistance value have
very little effect.
Conductive plastic/polymer potentiometers
For lowest noise, use conductive plastic tracks made from
extra-fine carbon bonded together with epoxy resin and given
a highly polished surface. This was first popularised in the
famous Penny and Giles faders used in expensive mixing
desks derived from aeronautical position sensors. These pricey
technologies still suffer from irregularities due to the finite size
and clumping of the carbon particles. Minimising the effect
of this is where the circuit design comes in. One day we may
get truly smooth carbon pots using nanoparticles.
Wiper
(moving contact)
Wiper
movement
Centre
terminal
Track material with
random carbon particles
and irregular surface
Clockwise track
end terminal
Track substrate: paxolin or ceramic
Centre terminal
Varying track
depth resistance
Wiper contact
resistance
Path of least resistance
Fig.4. Diagram showing how random resistance variations/
irregularities occur with rotation.
Cermet potentiometers
Another common track material is cermet, a ceramic and
metal glaze fired onto a ceramic insulating base. This is used
in instrumentation applications where high stability and high
dissipation are regarded as more important than rotational
noise. They a have a shorter rotational life, not because the
track wears out, but because the wiper grinds away on the
hard track. Cermet is the best trimmer technology to use for
adjusting quiescent current and offsets. Cermet pots do not
make good volume controls because the granular tracks are
scratchy and only come in linear taper. They are great for
setting values – trimming – and then simply left at that value.
Wipers
Most wipers are made from a springy metal, such as brass. For
lower rotational noise, multi-fingered wipers are employed to
average out track irregularities. If you want a quiet pot, four fingers
is a sensible number. Silver plating also helps. The Bourns pots
have a unique wiper with at least 10 fingers which accounts for
their superiority.
Moulded track pots use a single carbon brush wiper which
reduces wear and has a large contact area. Most importantly,
they do not oxidise, although the slip ring they run on may
do so. An important audio aspect of carbon brush wipers is
that they have a small but significant resistance of around 20Ω
called the ‘end-stop resistance’. Its presence means the volume
control may not go fully off. A cheap way
of getting extra wipers and lower noise is
to put both gangs of a dual Alpha pot in
parallel. Using this technique, the rotational
noise is reduced by 3dB.
Pot taper
Fig.5. Alpha pots give very cost effective
performance. Here the 16mm versions
are being used in a Colorsound Silver
Tonebender guitar pedal.
64
Fig.6. The stylish Bush TR82 radio used
Plessey E series pots. They rarely fail
apart from needing cleaning and their
leaky electrolytic capacitors changing.
Volume controls have always traditionally
had ‘logarithmic’ rather than linear tracks
to compensate for the logarithmic nature
of human hearing. I’ve put logarithmic in
inverted commas because the pot track
resistance actually rises exponentially as
the pot is rotated in a clockwise direction.
This gives a corresponding exponential rise
in signal voltage, which results in a smooth
subjective control of volume as the pot is
turned. The Austrian company Ruwido
mark their ‘log’ pots as ‘+exp’. The most
common log signifier is ‘A’ after the value.
Practical Electronics | June | 2023
Fig.7. Transition area of an Alps log fader.
Linear pots are marked ‘B’ and anti-log
devices with ‘C’. Just to confound audio
engineers, there are loads of contradictory
proprietary marking systems. European
manufacturers and the British company
Omeg swap the A and the B, while Bourns
uses D and Allen-Bradly uses R in their
long part numbers for log pots. Preh and
other German companies mark them +log.
Always check a datasheet or test a device
with a DMM set to Ω (ohms).
Many log pots make the required curve
by overlapping segments of different
resistance material. This is effectively a
combination of straight lines of different
slopes, sometimes called a piece-wise
linear approach. The noise tends to get
worse over the main transition area (Fig.7)
because the irregularities are greater. This
occurs in a band around at 70% of the
rotation in most log pots.
Pot position
+
It has to be accepted that pots are inherently noisy devices, they are, after all, a
moving contact. So it is sensible to place
them in the higher signal level parts of
the audio system. A pot placed just after
a microphone followed by +60dB of
gain would make almost
Output
Blocking
as much noise as the miDC offset
capacitors
crophone. Similarly, many
Input
Output
early radios were compromised in this respect
CW
because the audio level
W
Bias
was only around 20-100mV
current
Volume
CCW
followed by lots of gain. A
0V
pot placed after a preamplifier to bring the signal up
to around 1Vrms just before Fig.9. Block DC from a pot using capacitors.
the power amp is fine.
loaded by FET inputs, but in my experience such systems often drift, developing
Pot value
leakage currents and noise over time.
It is well known that fixed resistors develop
The basic idea of the pot scratching
Johnson noise that is proportional to their
articles of the 1970s mentioned earlier
resistance. However, higher resistance
was to increase the amplifier input imvalue pots suffer additional rotational
pedance loading so that a plastic-film
noise due to the lower proportion of carbon
coupling capacitor can be used instead
in the track material, making it rougher.
of a leaky electrolytic capacitor with
The higher impedance generates more
a pot’s wiper. Typically, this involved
noise voltage for a given leakage current
placing a buffer amplifier after the wiper,
and pickup of extraneous noise is also
as shown in Fig.10.
emphasised. It’s best to keep volume conThe main signature of DC leakage is
trols in the region of 1kΩ to 22kΩ. Some
that the crackling occurs when there is
old valve circuits will use values up to
no signal when the pot is moved. If it
2.2MΩ, but noise is expected with these
only occurs with a signal present, that
systems. Modern audio op amps will drive
is a sign that the problem is track wear
a load of 600Ω without suffering gross
or wiper contact problems. The most
distortion and many designers are now
revealing signal for this is a low distorusing low-value pots to minimise noise.
tion sinewave of around 300Hz. This is
a useful check for mixer faders. Using
Dump the DC
music usually hides the minor crackles.
It’s essential when controlling audio
Next month, we will dive deeper ito
with pots that there is no superimposed
the scratchy pot universe!
DC voltage across the track or wiper.
Always check for DC
voltage on the wiper
Post-fader amplifier
with a high impedance
+9V
(>10MΩ) multimeter.
Low impedance
The pot’s track irregularities do an excellent
Old or leaky
capacitor
job of converting DC to
C1
Output
Input
a random noise signal
BC549
as the wiper moves.
Leakage
The AC component is
current
+
5kΩ
passed on by the calog
+2V
Volume
bias
pacitor coupling the
wiper – see Fig.8. The
noise gets louder, the
faster the pot wiper
Post-fader buffer
Amplifier as before
is rotated.
+9V
The volume pot must
Small (10–220nF)
low-leakage film
be isolated with capac2.2MΩ
capacitor
itors on its input and
output as shown in
Fig.9. If the capacitors
C1
Output
Input
+2V
JFET
are electrically leaky,
2N5457
BC549
then noise will devel+4.5V
+
op. A scratchy pot is
Volume
10kΩ
often the sign of a leaky
2.2MΩ
coupling capacitor, typically an electrolytic
one. It is possible to use
pots without capacitors Fig.10. Avoid leaky electrolytic capacitors on a wiper by using
where the driving stage post fader buffers (PFBs). Driving a volume control from a
DC offsets are nulled- low impedance source makes the track resistance tolerance
out and the wipers are insignificant. Note reversal of interstage coupling capacitor (C1).
+
Static
value
Random irregularities band (dynamic resistance)
produces a rotational noise signal
Maximum resistance equating
to maximum output voltage
Theoretical straightline response for
potentiometer
+
Resistance
+
Real curve
AC-coupled
noise signal
Actual electrical travel
~1-3%
0%
Actual mechanical travel
~98% Rotation
100%
Fig.8. DC on an irregular potentiometer
track translates into a noise signal.
Practical Electronics | June | 2023
65
