The best volume control I have used is a shunt control. It has only one element (resistor) which can be a high quality Vishay, for example, in series with the input of the gain stage. The other part consists of a string of resistors which progressively shorts the input to ground.
The above can be realised with a four gang rotary switch or two, two gang switches.
Relays could be used to acomplish this but are noisy, these were used in early Pass Labs preamps. Today Pass uses a string of transistor switches to acomplish the same thing.
The shunt volume control is also used by Balanced Audio Technology in all their preamps (they acomplish this with cmos switches).
Actually, I suspect Nelson got the idea from one of the DIY'ers on his website gallary page ....
"... and an interesting electronic volume control. The volume control uses precision resistors and bipolar transistors to do the attenuation of the signal. Check Maik's web site for the complete description of the project.", viz:
Yes I've come across this before and I agree it will probably be the way I will go. I was, however, leaning towards a ladder type control, although I confess that in this application it probably won't make a huge difference.
If you look at a shunt control there is only one resistor in series with the next stage and in theory the shunt element should not affect the sound as it shunts the signal to directly ground. So that the signal that needs to be amplified only passes thru one high quality resistor and no switches.
This actually works but you have to be careful in the selection of resistors as noise and loading effects can come inyo play.
"...in theory the shunt element should not affect the sound
as it shunts the signal to directly ground"
I'm not aware of this theory. The only theory I know of is
the expression for an attenuator made of two resistors,
which says that the attenuation is R1/(R1+R2), where R1 is
the shunt resistance. The whole output signal is present
across this component, and the output is DIRECTLY
proportional to its value, so it might be expected to have
the most effect on the sound quality. I'd love someone to
convince me otherwise, but until then I suspect it's a
clever excuse by manufacturers to spend less on parts.
It is lower R2 which is the shunt resistance. Thus ,- the signal transfer will only pass through R1. The lower resistor, R2 will not be in the signal path, and the various sums or array forming the lower reistor can then be switched even by fairly unlinear transistors without affecting the signal path. 8 correctly calculated lower resistors, with correspondig transistor switches, controlled by an 8 bit counter will give an excellent 256 step volume control.
I'm not keen on the varying load presented to the source with a shunt control, but on the other hand the volume controls I've seen suggested as being best for CMRR in balanced circuits do look like shunt controls, as they can avoid the difficulty (impossibility?) of matching resistors closely enough in separate controls for each side of the circuit.
Is there an obvious (or not so obvious) way around the CMRR problem using ladder attenuation?
OK, I see the confusion. In my mind R1 was the shunt
impedance, but I've just looked at the "Fig 1" referred to
above, and it's the other way round. Anyway, I maintain that
it's the shunt component that is crucial.
I was skeptical about shunt controls till I built one. A shunt control will only have one resistor in series with the the input of the next stage, not a series of resistors, switch contacts or wipers as in conventional controls. ( Less contanimation of the signal ). Several resistors in series do not sound as good as one and switches and wipers affect the signal too. It might be an idea to build one and report your results.
Jake is correct about the varying load presented to the source and sometimes a buffer can help. The idea is to scale the resistors involved so as not present to difficult a load to the source and not too high as noise might be an
Yes, I accept that the absence of switches inserted between
input and output is an advantage, but I still don't see how
relegating these to the shunt leg should improve matters,
given my argument above.
Anyway, I'm close to finishing my preamp, and this by design
will allow me to compare different types of volume control.
I look forward to experimenting!
Most people have their pot wired as a series element.
To rewire as a shunt, choose a fixed series resistance, say 10-20KOhm (get a relatively decent one). Hook up pot so that it "shunts" the output (after resistor) to ground. Thus you need 0 Ohms at zero volume setting, probably 50-100K (is this not what your pot is?) when turned up all the way.
I run like this, works great. I use a motorized shunting pot of reasonable quality direcly powered by the volume settings in the CD player. Now you have remote controlled volume ...
After finishing two Aleph30 dual mono Amps, it's time to upgrade my Preamp again!
My current preamp is a Bal. Zen Line Stage with CCS on the bottom. I was looking for a better volume control to facilitate Balanced Input/Output. After going through various options, it appears to me that the one proposed by Maik is an attractive alternatives. I believed that the Pass XO also uses this method of Vol. Attenuation albeit a modified one.
After going through the notes and doing some simulations, I have made the following observations:
1) The "switches" are operated only one at a time.
2) You will need an extra input buffer stage. Six switches are placed before the buffer stage in -10db steps and another nine in -1db steps are placed between the buffer and amp stage. You will therefore get only 70 steps with a total of fifteen switches instead of more.
3) My simulation shows that with an input of 2.4V and Zero attenuation, the signal will go into clipping before entering the buffer stage. By changing R41¡¦R4n to 5K I was able to eliminate the clipping.
4) With a 10K series resistor for the first stage, the shun resistors works out to be 10,32,101,326,1111 and 4625 ohm (value rounded off).
5) With a 10K series resistor for the second stage, the value will range from 5499 to 81955 ohm
Here come my questions:
1) What is causing the signal to clip when all the switches are in off position? (All transistor in off state)
2) Any harm in changing R4 to 5K?
3) Can I use ZTX450 In place of BC547C
4) Is my calculation correct?
5) What is the best series resistor for the second stage?
6) Does any one out there who has done this have any advice before I actually start on the circuit?
7) Finally, is there a better way for implementing this circuit?
Isnt a transistor a lot more noisy then a good mechancal relay. I seldom change the volyme after setting it initially when listening.
The relays only sound then being changed but transistors are not totaly linear so they add something to the signal all the time.