Solupeak RC-11 Review (Passive Volume Control)

[pseudoid]

...coupling impedance of 99.990k ohms (check my math) which is about 80 pF...

If you are talking about transmission line theory of 'coupling'; then, the frequencies involved have to be way up in the hundreds (=eyeball math) of MHz range for an 80pF cap...

I suspect the apparent flatness with frequency in the plots is because the capacitance is small enough relative to the resistance that circuit bandwidth is well beyond the audio band

Especially confusing that the crosstalk improves with frequency...

I see about 3dB reduction (=eyeball math) in crosstalk graph which had made me ask my original question.

 

[DonH56]

If you are talking about transmission line theory of 'coupling'; then, the frequencies involved have to be way up in the hundreds (=eyeball math) of MHz range for an 80pF cap...

No, t-lines are my day job, this is for "fun". Although the coupling effects are the same, capacitive and inductive, as calculated for the EM fields around the conductors, but the wave equations for t-line coupling are usually much more complex and of course you have to consider distributed elements at RF/mW frequencies. That is not what I did here -- it is a lumped model. Capacitive coupling matters here because (a) impedances are relatively high and (b) the crosstalk is very low (<-100 dB). In a 50-ohm transmission line system at 20 kHz it probably (almost certainly) would not matter, but we are talking a passive device with around 10k-ohms impedance. As I said, please check my math and see what you get for coupling, using a simple impedance divider with the pot on one side and capacitor across. I assumed an ideal source so the capacitance may be higher, but the magnitude was not a surprise to me in how it would be realized in the box nor the impact on crosstalk. IME achieving <-80 dB crosstalk in an audio system takes some good design and layout (especially layout!) practice.

The inexpensive units I have seen (and some expensive ones) twist signal and ground wires together from jacks to potentiometer, leading to relatively high capacitance to ground. I suspect the crosstalk reduction at higher frequency is due to the excess ground capacitance that begins to dominate over the coupling capacitance.

But you could be right and I am all wrong; best calculate it yourself and let us know so I can correct my errors.

 

[Tom C]

I’m no expert, just an enthusiast, yet not the least bit frightened to use one of these. Audible distortion, cross talk, or perceptible alteration of frequency response? No, not unless it’s broken, which is always a possibility. I think noise is the biggest risk, so it‘s good this tested well in that respect. If the form factor fits the use case, I wouldn’t think twice about using one, especially at the price.
I actually prefer this type of solution over a built-in volume dial, because of the control it gives when putting a system together from discrete components.

 

[Michael Fidler]

Don't disagree, but do you have any idea how much a factor that is for typical amplifiers? Curious, not arguing, I do not really know. Bipolar input stages have base current to deal with, but JFET/MOSFET inputs not enough to worry about, and for op-amps it depends so much on the op-amp and circuit implementation that I couldn't guess (not many multi-GHz op-amps so I have not paid attention to them in years except as a hobbyist with too much work for hobbies). I am wondering how many power amps have enough nonlinear input current to significantly impact the distortion (yes, that is a lot of weasel-words... ).

I have in mind a look at the frequency response but need some free time in short supply lately. I actually had a spreadsheet I made many years ago but cannot find it, probably a victim of a hard-drive crash a number of years ago (yes, I had a backup, and it got corrupted too, long painful story).

I have a few times mentioned explaining to friends why their super low noise bipolar input op-amps were not the best choice for a phono cartridge input stage...

In my experience for a typical BJT input stage doing this more than doubles the overall distortion of the amplifier. Most of this distortion is third harmonic however due to the nature of LTP inputs. The level is very low indeed comparatively speaking, we're talking about the best case scenario for a class B bipolar amplifier design.

I don't have any measurements to upload right now, all I can offer up is a quote from veteran designer Douglas Self protesting against this sort of thing having seen it featured in an issue of Electronics World which perennially publishes badly designed audio circuits as filler material... Screenshot from his site attached.

For JFET/MOSFET inputs you're already going to have much higher distortion than necessary (and a DC servo to cancel the offset unless they're very carefully matched) so I doubt there's any benefit to having a low source impedance in the first place.

RE bipolar inputs, I've consistently found the 5534 and NJM2068 to be the best op-amps for the job vs all the FET devices I've tested so far, although they do have very low current noise indeed. Unfortunately parasitic and more importantly nonlinear capacitances are much higher in FETs which will most likely heavily degrade the distortion at the high impedance, high level, high frequency characteristics that moving magnet cartridges offer up.

 

[Michael Fidler]

Also just spotted this one giving some real numbers as opposed to factors...

 

[solderdude]

So... highly amplifier dependent and only in some cases becoming potentially audible.

 

[Michael Fidler]

So... highly amplifier dependent and only in some cases becoming potentially audible.

If the amplifier already has rather high distortion generated by other means then input source impedance distortion will be swamped.

 

[Michael Fidler]

I had a hunt for the original source. It's from a 2003 article titled Power Amplifier Input Currents and Their Troubles

Here's the proof. Adding a source resistance of 2.2k ohms more than doubles the distortion in a well designed power amplifier, although we're talking the difference between 0.0007% and 0.002% here.

 

[Tom C]

A crushing blow…

 

[DonH56]

In my experience for a typical BJT input stage doing this more than doubles the overall distortion of the amplifier. Most of this distortion is third harmonic however due to the nature of LTP inputs. The level is very low indeed comparatively speaking, we're talking about the best case scenario for a class B bipolar amplifier design.

I don't have any measurements to upload right now, all I can offer up is a quote from veteran designer Douglas Self protesting against this sort of thing having seen it featured in an issue of Electronics World which perennially publishes badly designed audio circuits as filler material... Screenshot from his site attached.View attachment 210922

For JFET/MOSFET inputs you're already going to have much higher distortion than necessary (and a DC servo to cancel the offset unless they're very carefully matched) so I doubt there's any benefit to having a low source impedance in the first place.

RE bipolar inputs, I've consistently found the 5534 and NJM2068 to be the best op-amps for the job vs all the FET devices I've tested so far, although they do have very low current noise indeed. Unfortunately parasitic and more importantly nonlinear capacitances are much higher in FETs which will most likely heavily degrade the distortion at the high impedance, high level, high frequency characteristics that moving magnet cartridges offer up.

Got it, thank you. That makes sense (as an old bipolar designer albeit not audio circuits but dabbled now and then) and is in line with my earlier thoughts; a potential problem with a bipolar input stage, not so much with other devices. And even for bipolar inputs it depends upon the design; base current cancellation circuits are common in my old designs, but that adds noise (always a trade). JFETs are common because they do not have the low-frequency noise that MOSFETs have (the latter often incorporate a special circuit to suppress the LF noise, but that circuit can itself add correlated noise).

 

[Michael Fidler]

Got it, thank you. That makes sense (as an old bipolar designer albeit not audio circuits but dabbled now and then) and is in line with my earlier thoughts; a potential problem with a bipolar input stage, not so much with other devices. And even for bipolar inputs it depends upon the design; base current cancellation circuits are common in my old designs, but that adds noise (always a trade). JFETs are common because they do not have the low-frequency noise that MOSFETs have (the latter often incorporate a special circuit to suppress the LF noise, but that circuit can itself add correlated noise).

This can of course be completely mitigated by using an op-amp input buffer for the power amplifier, such as a balanced line receiver, but worth keeping in mind. For power amplifiers with single ended input connections the line input almost always goes straight to the input stage.

As a staunch supporter of bipolar devices I would posit that FETs on the input would give you the required deterioration without the inconvenience of a passive attenuator .

 

[Michael Fidler]

Also I would imagine that the input current noise of a bipolar power amplifier input biased for good linearity and then resistor degenerated for further benefit would be high enough to generate a very significant amount of noise across a 2.2k ohm source.

 

[KSTR]

RE bipolar inputs, I've consistently found the 5534 and NJM2068 to be the best op-amps for the job vs all the FET devices I've tested so far, although they do have very low current noise indeed. Unfortunately parasitic and more importantly nonlinear capacitances are much higher in FETs which will most likely heavily degrade the distortion at the high impedance, high level, high frequency characteristics that moving magnet cartridges offer up.

That's certainly true for the generic/legacy types of JFET OpAmps, but you might want to take a look at the "new breed" of JFET (audio) OpAmps from TI like OPA1642, OPA827 and some others which have fully bootstrapped/cascoded input stages. Input capacitance is low and essentially constant over the CM input range (IIRC, this is also true for the already exceptional CMOS OPA1656).
And pretty low noise for the class, though not exactly close to the best bipolars.

 

[Michael Fidler]

That's certainly true for the generic/legacy types of JFET OpAmps, but you might want to take a look at the "new breed" of JFET (audio) OpAmps from TI like OPA1642, OPA827 and some others which have fully bootstrapped/cascoded input stages. Input capacitance is low and essentially constant over the CM input range (IIRC, this is also true for the already exceptional CMOS OPA1656).
And pretty low noise for the class, though not exactly close to the best bipolars.

Both the OPA1642 and 827 are noisier in terms of voltage noise than even the super-cheap NJM2068. Although the datasheet gives confusing measures of noise (RIAA, A-weighted etc.) that don't mean much I measured 3.5nV/sqrtHz and 0.3pA/sqrtHz, putting it firmly ahead of the OPAs 164x and 827.

Regarding the OPA1652, it has lower voltage noise than the 5534 and 2068 but is twice the price of an LM4562 which beats it again (and has the advantage of a DIP package).

 

[DonH56]

I added a couple of posts to my thread on how passive attenuators affect the sound, including a sweep of a passive volume control such as this. See https://www.audiosciencereview.com/forum/index.php?threads/attenuator-bandwidth.34268/ -- for the example values chosen it should still not have a big impact, but with some other choices (components) it certainly might.

FWIWFM - Don

Edit: Added a crosstalk sim for @pseudoid since I am sure others are also wondering...

 

[dananski]

Looks quite decent for the money, but how can you tell if it'll be any better than the pre in your integrated amp or DAC?

 

[pseudoid]

Looks quite decent for the money, ...?

If I follow you correctly, then it would mean that it'd be as "quite decent" if the price as well as the power was half as much as this one.
Please don't bother to extrapolate what would happen if the price (and power) were doubled...

 

[dananski]

If I follow you correctly, then it would mean that it'd be as "quite decent" if the price as well as the power was half as much as this one.
Please don't bother to extrapolate what would happen if the price (and power) were doubled...

Not sure if you do follow me correctly, because I can't follow you at all! It costs $17 and can comfortably go to -10dB without audible crosstalk, frequency response colouration, noise or distortion. The Schiit Sys preamp costs about three times as much for similar if not slightly worse performance, so I say this is good value based on an apples to apples comparison, not some made up extrapolation. I have no idea what you're on about with regards to power given this is a passive device, and I've no idea why you latched onto a trivial, uncontroversial statement made by others already, rather than the actual question I posed!