Volume Control At the Output

For over twenty years, I had an Aleph P as my pre-amp: first the original version, then the remote version. I finally moved on to an XP-20 about a year ago (yes, it's a lot better) but stumbled across the manual for the P (which I thought I'd lost and now need to send to the buyer) a couple days back.


In it, Nelson expounds upon the virtues of putting the volume control at the output: It allows the pre-amp always to operate upon a predictable, full-level signal, and attenuating the signal also attenuates the noise. Now, so far as I can tell, none of the Pass DIY pre-amps have adopted this strategy, and there's no mention of it in the manual for the XP-20 (though much of the other language is the same).



So, I'm curious: Why the change? What are the downsides to this kind of arrangement? I know that the Aleph P (at least the remote version) had a variable output impedence, though it never caused me any problems.



Of course, in the DIY world, we can try this out. What would it take (say) to adapt the Korg B1 to it? Should I put a 50K resistor to ground at the input and just put the pot after the output cap? If so, do I still need the 100R and 33.2K resistors or should those be changed?


FYI, I can follow directions build these things. I don't necessarily understand them....
 
What are the downsides to this kind of arrangement?

It is trading noise against headroom in the case of a low resistance attenuator at output. Have tried it and couldn't see/hear no particular advantage.

The only easy way to design a low impedance attenuator suitable for commercial equipment today is to use a relay attenuator as none of the common electronic pots are suitable.
 
If one considers an op-amp with a gain of 100, having an output of 5volts with a noise level of 5mV.
If there is no signal present there is 5mv of noise present.
At full signal the noise is not noticeable.
If then a potential divider, (pot) is place in the output and reduced to 50% or half, there will be 2.5volts of signal but only 2.5mv of noise so the signal to noise ratio is no better but there is less of it if your power amplifier only needs 2.5volts for maximum output.
That is exactly what the main volume control, on the power section of your amplifier, is for and why with PA systems, the engineer always sets a 3dB below maximum signal from the pre amp stages.

No exactly rocket science.
 
I think it is the same problem when using a passive preamp that the pot has to drive the interconnect cable? ......so maybe better to put the pot at the input of the power amp so pre-amp still drives the interconnect cable and pot has very short wires (few cm) to input of power amp?
 
The control on the output will cause an adjustable rise in the preamp's output resistance, which may interact with cable capacitance and amplifier input parameters. This could cause the sound quality to vary as volume is adjusted. I'm not a big fan of the idea.
 
What are the downsides to this kind of arrangement?

Any solid state source can drive a 10k load, so adding a 10k volume control at the preamp output is fine.
It would then take 1660pF combined cable and amplifier capacitance to cause just 1dB of loss at 20kHz
in the worst case -6dB, half resistance setting. Of course, many cables will be at least a factor of ten
better than that, so the -1dB point would then be over 200kHz. You would just not want so much gain
in the preceeding stages that they could overload. There have been line stages with ganged volume controls
both before and after the circuit.
 
Last edited:
Like I said earlier if they put a buffer at the output you can have your cake an eat it.

Pass labs generally run with a buffer output on their preamps anyway, so just make a 2 stage preamp with feedback then the volume pot then the buffer.
The buffer is left outside of the feedback loop.

Bob's your uncle

You can now use any pot load to match the preamp circuit, and still have any output impedance you want in the range of 50 Ohms to 100 Ohms which remains a fixed operating point.
 
Last edited:
If you don't have a gain knob in front, eventually you may meet a signal which will overwhelm the amplifier (distort). IAC you will want to super-size that amplifier to keep small THD down.

The gain pot goes where the cost of audio power starts to rise quickly. From 0.1V to 1V, no real difference. From 1V to 10V, different. From 10V to 100V, absurd rise of cost. Especially if you must turn-down from that near-100V to near-3V of a power amp input.

But in high end and in DIY, "cost" is a fishy thing. High power high voltage parts cost more than commodity audio can bear, but are not really expensive if you live in a fine home. Or if you live in a hovel to afford "fine" audio.
 
sole problem with attenuator on output, used as loading or not, is - it must be low in value to ensure low Rout (except if buffered)

then, to achieve minimally needed amount of symmetry and tracking between channels - one must use something better that regular pots - it must be (relays switched) matrix of resistors ........ or - worse scenario - rotary switch equipped with bunch of resistors

all road leads to Rome, some are just more paved
 
The performance advantage of locating the volume control at the output is that SNR is maintained as volume is reduced. However, In addition to loss of control over headroom that was mentioned above and the increase in stage output impedance, placing the volume control at the output also results in the highest relative level of distortion being generated by the gain stage. This is because the stage is always being fed the full input signal amplitude, which maximally excites the circuit's non-linear elements. If the volume control is instead located on the input side, the input signal fed the gain stage is never higher in level than is needed to produce the desired output volume.
 
Knowing the brainy people at Pass Labs, pretty sure that would be taken into account within the design.
In fact it's kind of common sense stuff, so not much intelligence is required, I'd say they'd have that covered.
It's a less efficient design obviously, but in comparison to a 100W Class A power amp, it's a drop in the ocean.
 
Yes, if by "taken in to account" you mean, decided to accept the objective degradation. Accepting it, however, doesn't mean that it's inconsequential. The fact that most (all?) of Pass Labs' own commercial designs utilize input side volume control should tell you they've concluded that, in net, the advantage of an an output side volume control is outweighed by it's disadvantages.