Volume Control in the Hi-Fi/High-End

Alex Torres, Netanya.

(part 1, 2004, revised in 2007)
Preamble

This article was written 3 years ago, but only Part I was published.
Unfortunately - the time is always not enough :(, so the second and
the third parts were not finished. Maybe Its even better - during these
3 years some designs were changed, some new ideas were implemented
Now I have decided to finish this work and to change the structure
of this article. The 1st part was totally revised, all information concerning
the nikitin volume control was moved to the 2nd part, than now totally
dedicated to this design. The Part III (multi channel application for the
Home Theater) was moved to the Part IV and the new Part III will describe
the advanced tube/solid amplifiers control (nikitin based).

The problems of the Volume control.
In spite of its simplicity, the volume control in the high level audio equipment still performs a
problem. The short pass oriented audiophiles sometimes even not want any volume control - the
music should sound as is. But in reality - life without volume control is not comfortable.
What the volume control device must to do? The sound pressure level must linearly depend of the
volume knob angle (in order to do this, the volume control element should be logarithmic). It is
not hardly to make, if you have only one channel, but in 2 channel stereo application (and multi
channel Home theater applications too) the big problem is the equality of the channels controllers.
Of course, volume controller is not allowed to produce any noise (as acoustic, as electric). The
good idea is to have the possibility of the remote control.
There are lot of solutions - audiophile ALPS potentiometers, simple or with the motor control
(by the way, why Alps? Bourns and Honywell make not worth potentiometers, but they are often
more expansive due to an aerospace orientation of these companies). Gold platted rotary switches,
variable input transformers (well known Sowter, transformers from Alex Vorobiev, etc.) and
even undeservedly contemned by audiophiles electronic volume controllers. The last ones
are very ambiguous - the wide range of the existing devices with the real high quality, to the low
end devices, that are suitable only for portable mp3 players, car audio, etc.
In this article I will try to observe some solutions - with electronic attenuators, with relays, with
micro controller control, from the simple up to the more advanced.
Short review of the volume control IC.
Electronic volume control usually based on the voltage control amplifier - VCA (or VGA -
Variable Gain Amplifier), for example at SSM2160, or switchable resistor network The most
popular IC are digital potentiometers (DS1802 from Dallas-Maxim, etc.), volume control chip
CS3310 from Cirrus Logic (Crystal Semiconductors) compatible with PGA2310 from Burr-
Brown (now - Texas Instruments), LM1972 from National Semiconductors. One thing - many
efforts to use 3-channel LM1973 fails. Im sure that LM1972 and LM1973 have the same internal
architecture, but the difference is very enormous.
Last time we have the new TI chips - PGA4311, than have 4 channels and better performance. (2
channel version - PGA2311). The old chip SSM2160 from Analog Devices provides 6 channel
control, but worse performance.
Various IC like LM1036 are out of out discussion due to their abominable sound. Certainly, this
short review does not cover all volume control IC that exists all over the world. We will talk about
some representatives.
Simple LM1972 based volume Control.
LM1972 produce separate volume control of the 2 channels in the range from 0 to --78dB. Step is
0.5dB in the 0 - -48 dB range and 1dB at -48 - --78dB. This chip use simple 3-wire protocol (SPI
bus):

If you need more then 2 channel control, you can use 2 solutions: it is possible to put some
IC in parallel and to control all channels simultaneously (for example - 2 parallel lm1972 are able
to control LF, RF, LS, RS, Center and Sub, but you cannot control all kinds of balance) Or - you
can use daisy-chain connection. The first way is simpler, because you dont need to make any
changes in the microcontrollers software. Both two methods are shown at the diagram (right -
parallel, left - daisy-chain) :.
SUBW_IN
DT
L_BACK_OUT
RIGHT_IN
SUBW_OUT
R_BACK_IN
CENTER_OUT
RIGHT_OUT RIGHT_IN
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
R_BACK_OUT
SUBW_IN
CLK
DT
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
CENTER_OUT
LEFT_IN
CENTER_IN CENTER_IN
L_BACK_IN
R_BACK_OUT
CLK
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
LEFT_OUT
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
4
20
2
17
12
11
9
10
IN1
IN2
OUT1
OUT2
DO
DT
CLK
CS
CS CS
LEFT_IN LEFT_OUT
R_BACK_IN
RIGHT_OUT
L_BACK_OUT
SUBW_OUT
L_BACK_IN

The most simple control device is shown at the diagram - is has no display, control provide by
two potentiometers - not logarithmic, not double way, not audiophile quality. Any most
cheapest as you can find, even the value not need to be accurate - use what you have, from 5k to
50k. This controller has also mute button and two LEDs to display mute status ant the balance
control center position.
Maybe you have a question: Why we need all this, to use potentiometer?? The answer is
simple, and I give is above -this potentiometers has no influence to the audio signal. Both volume
and balance potentiometers change the voltage at microcontrollers ADC inputs. Micro measures
the voltage and produce the necessary signals to the volume control chip. The ICs control line
used also for LEDs, as these lines can be in any position in inactive time.
V
C
C
1
GP3/MCLR
4
GP0/AN0
7
G
N
D
8
GP1/AN1
6
GP2
5
GP4
3
GP5
2
U4
PIC12C671
IN1
4
IN2
20
OUT1
2
OUT2
17
V
D
D
1
3
V
D
D
1
5
GNDAC
1
GNDAC
5
GNDAC
6
GNDAC
14
GNDAC
16
DGND
8
G
N
D
3
G
N
D
1
9
V
S
S
7
V
S
S
1
8
DO
12
DT
11
CLK
9
CS
10
U3 LM1972
C2
100n
C6
100n
+
C3
100.0
+
C5
100.0
Remark: PIC12C671/672, PIC12F675
R6
10k
-9..-15v
R2
2k
C1
100n
D
1
M
U
T
E
D
2
C
E
N
T
E
r
R5
2k
R4
2k
R1
25k
R3
25k
R_OUT
L_OUT L_IN
R_IN
GND
SW1
MUTE
OUT
1
G
N
D
2
IN
3
U2 L78L06
IN
2
OUT
3
G
N
D
1
U5 L79L06
OUT
1
G
N
D
2
IN
3
U1 L78L05
V
O
L
B
A
L
-5VA
+5VA
C4
100n
C7
100n
+9..+15v

For the best performance, it should be connected to big input resistance. If you amplifier have low
- you should use the OP buffer
("$2132, "$2134 & '.(.). Take care about ground wiring - IC has the special digital and
analog ground pins... Input signal can be up to 3V RMS. level. Input resistance - 40kOhm.
Remark: I received many complains than PIC12C671 is not widely available, and because using
of the One Time Programming micro is not a good idea for the D.I.Y, I make the software also for
the flash micro - PIC12F675.
PGA2310 based Volume Control
Not all the people are able to make the changes in the software for LM1972, to work with
PGA2310. Both chips have the same SPI interface, but different internal structure. This device is
for those peoples.
At 1st, PGA2310 can not only attenuate the signal, but also to amplify! The control range is from
-95.5 dB up to +31.5dB wit the 0.5 dB step. The powerful output can be loaded with 600 Ohm
directly, without any buffers! But using of the full PGAs range looks very plenty and I decided to
make 3 software versions:
1. Full original control range, from-95.5 dB to +31.5 dB, but with 1dB step.
2. NoGain without amplification, i.e. from -63 dB to 0 dB, with 0.5 dB step
3. Gain3 with limited amplification, from -53 dB to +10dB with 0.5 dB step.
Balance control in all versions is +/-8 dB with 0.5 dB step.
With the full version the hardware limits are also possible (as will be described below).
+15v
R_OUT
AGND
U2
PIC12C671
1
4
7
8
6 5
3
2
V
C
C
GP3/MCLR
GP0/AN0
G
N
D
GP1/AN1 GP2
GP4
GP5
C2
100n
L_OUT
SW1
MUTE
R3
2k
D
1
M
U
T
E
R1
25k
R4
2k
GND
L_IN
D
2
C
E
N
T
E
R
+
C5
100.0
-15v
+5V
R_IN
R8
10k
B
A
L
+ C3
100.0
AGND
V
O
L
+ C8
100.0
C4
100n
C1
100n
Remark: PIC12C671/672, PIC12F675
AGND
R2
25k
U1
PGA2310
16
9
14
11
1
2
4 1 8
5
15 10
1
3
7
3
6
2
IN1
IN2
OUT1
OUT2
V
C
C
V
D
D
Z
C
E
N
-
M
U
T
E
D
G
N
D
AGND AGND
V
S
S
DO
DT
CLK
CS

As you can see - the schematic was not changed dramatically (power supplies regulators are not
shown). The major differences are in the software. Like LM1972 it is possible to use some PGA
in parallel to obtain the multi channel volume control.
The possibility to operate with the low impedance (600 Ohm, 1000pF) load (output current of
the PGA2310 35)$, and 2311 50)$) give us the elegant opportunity to use PGA as
preamplifier, or to get other benefits as the buffer. Below you can see some examples of these
solutions.
LM4780 (LM3886) inverting
Inverting schematic of the LM3886 (double version LM4780) have better sound then widely
used non-inverting. In order to solve the stability problem, gain must be more than 10 (some
people said even 20-30), and feedback resistor should be not more then 50 kOhm. This determine
the input resistor (input impedance) not more then 3-5 kOhm. With its low output impedance,
PGA2310/2311 can be buffers for inverting power amplifier. You must remember than input
impedance of the PGA is also not very big - 10 kOhm, so sometimes PGA need the buffer at the
input.

R
6
2
5
k
V
O
L
SW1
MUTE
R2 4k7
U2
PGA2310
16
9
14
11
1
2
4 1 8
5
15 10
1
3
7
3
6
2
IN1
IN2
OUT1
OUT2
V
C
C
V
D
D
Z
C
E
N
-
M
U
T
E
D
G
N
D
AGND AGND
V
S
S
DO
DT
CLK
CS
+
C4
100.0
GNDA
Remark: PIC12C671/672, PIC12F675
-15v
R7 4k7
R10
?
R3
10k
R11
2k
-U
C5
100n
D
1
M
U
T
E
U3 PIC12C671
1
4
7
8
6 5
3
2
V
C
C
GP3/MCLR
GP0/AN0
G
N
D
GP1/AN1 GP2
GP4
GP5
+ C6
100.0
AGND
R4 4k3
+15v
R1
47k
AGND
R12
2k
L_IN
+5V
GND
+ C1
100.0
D
2
C
E
N
T
E
R
R_IN
U1 LM4780
241
0
1
1
1
8
5891
7
2
7
16
22
15
21
7
25
14
20
1
2
1
9
-
V
C
C
-
V
C
C
-
V
C
C
-
V
C
C
-
V
C
C
+
V
C
C
+
V
C
C
+
V
C
C
+
V
C
C
+
V
C
C
INA
INB
-INA
-INB
OUTA
OUTB
MUTEA
MUTEB
G
N
D
G
N
D
OUTPUT
C3
100n
R9
47k
+U
C2
100n
R5 4k3
B
A
l
R
8
2
5
k


There are no capacitor between the volume control and the power amplifier, and PGAs output
offset voltage will be amplified by the LM. Thats why the output stage gain should be as small as
possible. The same reason - to use PGA2311 instead of PGA2310 due to it's half offset. We not
discuss LM3886/4780 in this article, you can pass to the data sheets and to the various FAQs in
the Internet. Bypass capacitors and Bushero output circuit re not shown at the diagram above.
SE Headphone Amplifier.
In opposite to the previous schematic, in this design more preferable is PGA2310, because the +/-
5v power supply limitation of the PGA2311. PGA2310 have up to +/-15v working range, with
shown at the diagram below +/-12v power supply, PGA2310 with the emitter follower is able to
provide output voltage 3.2v @ 32 Ohm load and 6.5v @ 300 Ohm, that is quite enough!
This design enjoys the Full control version with the hardware limit adjustment. With the help of
two internal trimmers you can limit the maximum and minimum volume, and to get the most
comfortable control according to you headphones.
Remarks: p-n-p transistors MJE170 can be changed to BD140, etc., or mirroring the schematic
and use n-p-n transistors like MJE180, BD139, etc. The sound quality depends of the transistors -
MJE170 from Motorola sound very nice, also BD140 from Philips. But MJE170 from ST
sounds worse... Transistors idle current is about 220 mA and can be adjusted by R1R2.

+5V
AGND
R1
51R
R
5
1
0
k
J1
CON3
1
2
3
C3
100n
R6
1k
R
3
1
0
k
+12V
-12V
+
C2
2200.0
R9
1k
R2
51R
D
1
C
E
N
T
E
R
R
7
1
0
k
Q1
MJE170
+
C1
2200.0
R
4
2
k
2
5V_GND
H
E
A
D

O
U
T
P
U
T
R_IN
+/-12V_GND
LINE INPUT
U1
PGA2310
16
9
14
11
1
2
41 8
5
15
10
1
3
7
3
6
2
IN1
IN2
OUT1
OUT2
V
C
C
V
D
D
Z
C
E
N
-
M
U
T
E
D
G
N
D
AGND
AGND
V
S
S
DO
DT
CLK
CS
R
1
0
2
k
2
R8 2k
V
O
L
L_IN
U2
PIC12F675
1
4
7
8
6
5
3
2
V
C
C
GP3/MCLR
GP0/AN0 G
N
D
GP1/AN1
GP2
GP4
GP5
Q2
MJE170
B
A
L


Remember - the efficiency of the Class-A amplifiers is like the efficiency of the steam train. All
class-A amplifiers are extremely HOT - be careful, use the heatsink. With 220 mA idle current,
about 11W going to the heat on the two output transistors and their resistors. If you will use only
low impedance headphones (32 Ohm) - its better to drop the power supply voltages from +/-12v
down to +/-9v or even to +/-5v (using 7809/7909 and 7805/7905 regulators). If in opposite - you
will use only high impedance headphones (300, 600 Ohm) it is practical to reduce idle current
by increasing the emitter resistor.

-- end of part 1 --