TVC Preamps

[hifinutt]

Thanks for that. I wouldn't rush into selling Ref 3. Do you still have the 5SE?
I had heard of the Allegri until I saw this thread. It's about £800 cheaper than the MFA Classic V2, but doesn't have XLR inputs. If you asked me 6 months ago, I'd have said that the Ref 3 was staying for ever. But I find I'm sensitive to volume now (normally 60-70dB according to my phone app) and I didn't use the system for a few weeks because it is hard on my ears. My Yamaha piano is playing normally at 70-82 dB and I can play it all day.

Basically, if I can get the result of a Ref 3 from a tiny box that is 'ON' all the time, then a TVC may be a good thing for me.

sadly had to sell the 5se to pay for a house but using a modwright 36.5dm now which is exceedingly good and not far off the 5se methinks . it also costs a great deal less s/h !!

the good think about the ref3 is they really keep their price and you could almost do a straight swap for a baby ref which is an awesome pre amp . there are a number around s/h

 

[hifinutt]

oh and if you really want to push the boat out , jonathan billington will bring you a v2 baby ref to try in your system !! about 6k though !!

 

[Tony L]

Still not understanding the shunt thing. Here's a link to HifiCollective with some example units.

There is some implication there is a difference in the way shunted and ladder devices behave from an input impedance perspective. FWIW my Audio Synthesis seems to be a fairly constant 16.5k until the loudest few steps (which I never get close to) where it drops swiftly down to well under 10k. At all points I've ever used it it seems to be 16.xx k.

 

[GT]

Not in my experience. And that would make it an AVC NOT a TVC!

The Slagle designs definitely are which would account for the Townsend Allegri. So is the Tribute design. Some other designs were too back in the day, well 15-20 years ago when these were first being used to control the volume. I seem to recall S&B dabbled with an autoformer design before MFA was born.

 

[Julf]

Still not understanding the shunt thing.

The attenuator element is basically a switched, variable resistor. To make it attenuate, you make a voltage divider with the switched resistor as one element, and a fixed resistor as the other one. You can use the switched resistor as a series resistor, with a fixed resistor to ground (forming the output impedance), or use a fixed series resistor and have the switched variable element connected to ground (as a shunt).

There is a misguided belief that only the resistor that the signal passes on the way to the output has an effect on sound quality, not the shunt resistor to ground, so the shunt connection allows you to use a fixed, supposedly very high quality resistor as the series element, and the switched resistor "only" as the ground shunt element.

 

[Tony L]

So you are saying it is inferior to a conventional ladder like my AS with switchable dscreet resistors on both signal and return for each step?

PS For now please forget the 'single high quality resistor' aspect, assume I'd get top quality resistors everywhere. I just want to understand the core advantages/disadvantages for ladder vs. shunt. All things being equal which is the better approach?

 

[Julf]

So you are saying it is inferior to a conventional ladder like my AS with switchable dscreet resistors on both signal and return forveach step?

PS For now please forget the 'single high quality resistor' aspect, assume I'd get top quality resistors everywhere. I just want to understand the core advantages/disadvantages for ladder vs. shunt. All things being equal which is the better approach?

The shunt one produces larger impedance variations, so it is inferior in that respect.

The myth of shunt attenuators is based on not understanding basic circuit theory and assuming only elements that the signal "passes through" affect the sound.

 

[Tony L]

How wide an impedance variation could one expect, e.g. on a nominally 100k build where would it be low, middle-range, wide-open? Much of my interest in this is impedance related as I want to get a really source component-friendly solution for my Leak that adds no noise or distortion. That is the goal and really it is my only criticism with the AS aside from I'd like just a little more heft/weight to the sound, which again is likely impedance related.

I guess the benefit of a shunt is one less switch contact?

 

[Julf]

How wide an impedance variation could one expect, e.g. on a nominally 100k build where would it be low, middle-range, wide-open? Much of my interest in this is impedance related as I want to get a really source component-friendly solution for my Leak that adds no noise or distortion. That is the goal and really it is my only criticism with the AS aside from I'd like just a little more heft/weight to the sound, which again is likely impedance related.

Input impedance will always be fixed resistance + variable resistance (ignoring input impedance of subsequent stage for now). In the series circuit, output impedance will always be same as fixed resistor, in shunt circuit it will go from the maximum value of the variable resistor at full volume down to zero at full quiet.

I guess the benefit of a shunt is one less switch contact?

One less switch contact in series, but one more in parallel. So do you worry about voltage or current?

 

[awkwardbydesign]

The Slagle designs definitely are which would account for the Townsend Allegri. So is the Tribute design. Some other designs were too back in the day, well 15-20 years ago when these were first being used to control the volume. I seem to recall S&B dabbled with an autoformer design before MFA was born.

You mean they are AVCs, yes?

 

[awkwardbydesign]

This why sometimes it's better that a preamp has the attenuator on the output as it could be the preamp is adding its own noise in this case.

The 01A DHT pre-amp I just built has way too much gain (one of these - http://www.bartola.co.uk/valves/dht-pre-amplifier/01a-preamp-gen2/) and therefore noise, so using a TVC (Sowter) on the output overcomes both the gain and noise problems, as both are attenuated together. A pot on the input did neither.

 

[Paul R]

I think you misunderstand me, hiss, tube whistles etc are in the context of using an active tube preamp, the JC Verdier, with the Leak. It is just too high-gain a power amp to partner with an active pre IMO. The AS is obviously entirely silent.

OK. Still seems a bit unnecessary to tolerate with any output gain structure...

This why sometimes it's better that a preamp has the attenuator on the output as it could be the preamp is adding its own noise in this case.

Which takes us back to most of the passive issues.

Paul

 

[Paul R]

Input impedance will always be fixed resistance + variable resistance (ignoring input impedance of subsequent stage for now). In the series circuit, output impedance will always be same as fixed resistor, in shunt circuit it will go from the maximum value of the variable resistor at full volume down to zero at full quiet.

I think this is incorrect.

A conventional passive will have an output impedance at 0 attenuation of the pot in parallel with the output of the source. So in general it will be the output of the source. It will tend to 0 at maximum attenuation and be maximal when the pot is in its 6dB setting, where it will approximate to half the open resistance of the pot.

The input impedance of the conventional will be more or less constant, of the shunt it will vary depending on position of the pot. Because the source essentially sees the shunt in series with the adjustable part.

Paul

 

[Julf]

The input impedance of the conventional will be more or less constant, of the shunt it will vary depending on position of the pot. Because the source essentially sees the shunt in series with the adjustable part.

I guess we have a different understanding of how a shunt attenuator is constructed. My understanding is that in a shunt attenuator the adjustable part is the shunt (connected between the output and ground).

 

[GT]

More importantly to the user, shunt types do not have great control of the volume as the slope of the scale (the audio taper) is different to that of a good pot. Shunt types tend to give a great amount of level in a very small amount of control and this happens very quickly after zero volume, after that the taper is all but done. You can play around with the series resister to change this but only slightly. I find this a big disadvantage that out weighs any gains in sound quality you may experience over a good logarithmic scale pot. This would be a serious disadvantage with a high gain amplifier like the Stereo 20. Also depending on the pot used, its accuracy if poor in the region of most gain could give a big disparity in channel balance. However, channel balance shouldn't be a problem if a stepped attenuator used as the control.

 

[GT]

The 01A DHT pre-amp I just built has way too much gain (one of these - http://www.bartola.co.uk/valves/dht-pre-amplifier/01a-preamp-gen2/) and therefore noise, so using a TVC (Sowter) on the output overcomes both the gain and noise problems, as both are attenuated together. A pot on the input did neither.

More likely your power amp has too much gain. 15dB is about the smallest amount of active gain you can achieve with RC coupling and no feedback. The 18dB of your preamp is only 3dB up on that. In your circuit you really need a low impedance output to drive the TVC if you have it connected on the output.

 

[ChrisPa]

A conventional passive will have an output impedance at 0 attenuation of the pot in parallel with the output of the source. So in general it will be the output of the source. It will tend to 0 at maximum attenuation and be maximal when the pot is in its 6dB setting, where it will approximate to half the open resistance of the pot.

The input impedance of the conventional will be more or less constant, of the shunt it will vary depending on position of the pot. Because the source essentially sees the shunt in series with the adjustable part.

Paul

I guess we have a different understanding of how a shunt attenuator is constructed. My understanding is that in a shunt attenuator the adjustable part is the shunt (connected between the output and ground).

You may both have described the same thing. Paul's described it in terms of what is seen by the source, julf has described it in terms of how its output appears.

If either of you could search out a diagram of what you mean, that would help clarify

 

[Tony L]

More importantly to the user, shunt types do not have great control of the volume as the slope of the scale (the audio taper) is different to that of a good pot. Shunt types tend to give a great amount of level in a very small amount of control and this happens very quickly after zero volume, after that the taper is all but done.

Again could someone please try and explain this to me? I'd have thought the steps could be chosen just the same as on a ladder attenuator, i.e. they couod be made exactly the same steps in db. The argument that 20 odd steps is too few obviously still stands, but the better ones these days offer up to 47 steps which should be plenty to cover the whole range with sufficient precision I'd have thought.

 

[ChrisPa]

Again could someone please try and explain this to me? I'd have thought the steps could be chosen just the same as on a ladder attenuator, i.e. they couod be made exactly the same steps in db.

Yes, if it's a switched attenuator.
If it's a standard pot then you've got a choice of logarithmic or linear. When used in a shunt configuration as described, the shunt+log pot no longer gives a logarithmic resistance curve

 

[ChrisPa]

A shunt attenuator