Easiest way of comparing 1100H TVC to BeeQuiet attenuator?

[Deke609]

Because I am impressed with the SILK (SAC Thailand) grid chokes in my Kaiju, I picked up a pair of SILK's supermalloy TVCs (transformer volume controls).  The manufacturer's reported specs are -1 dB @ 10Hz and 35kHz referenced to 1KHz, and greater than 1,100H. Most folks used these things to build passive pre-amps (which I did this wknd - just to see). But I bought them to test whether they might sound different/better than a resistor-based attenuator. But before getting to that: a quick report on the results of making a simple passive preamp out of them - b/c it informs why I want an "easy" way of testing the TVCs with my BeePre 1.

The passive pre I put together is super simple: bascially a pair of rca inputs -> 23 step TVCs -> rca outputs. Nothing fancy. (I'll be testing it with XLR inputs and various switchable grounding configurations when I get the correct switches from Mouser this week). 

I tried the passive pre this evening in place of the BP1 feeding my Kaiju. It actually sounds ok - vocals are very nice.  But the bass lacks punch and doesn't go as deep, and the edges of everything sound a little too rounded over. It sounds "musical": pleasant, smooth, nothing to annoy. But nothing to excite either.  No jump, attack, etc.

That's not a problem - to the contrary I'm pleased by the results: they confirm my already tested belief that the BP1 is pretty extraordinary. Plus, I never intended to replace the BP1 with a TVC-based passive pre (or, indeed any other thing, except maybe the BP2 - but I consider that a rejigged BP1). However, having read people rave about TVC and AVC passive preamps, and having the makings for one on hand, I figured I should throw one together and listen for myself before trying the TVCs as attenuators in the BP1.

Which brings me to where I am at now. I think the TVCs *might* hold some promise of sounding good as replacements for the BeeQuiet attenuators -- but I don't have enough faith that they will to justify pulling parts out of, and rewiring, my rebuilt BP1.  So I'm looking for insight, conjecture or even wild guesses as to whether and how I can get away with testing the passive pre I built as an external replacement for the BeeQuiet attenuators. IOW: disconnect the BeeQuiet and have the BP rca input directly feed the grid of the 300B, with the TVC attenuator box sitting between my DAC and the BP.  Not ideal, b/c the attenuator will have to drive 2 ft of interconnect cable instead of only a few inches of wire - so there will be some additional capacitance to drive.

What I haven't been able to wrap my head around is whether I need/want an input grid resistor that would load the TVCs and provide a purely resistive path to ground for any tiny amounts of grid current that might arise. My first thought was "Yes, obviously", but given the +1,100H inductance, I'm wondering whether I need a reflected load. That's a ton of inductance - nearly 70K of impedance at 10Hz, and tons more at higher frequencies. So I'm thinking of simply connecting one leg of the TVC secondary to the grid (but leaving the grid stopper resistor in place) and the other leg to 0V/chassis. Since the 7000H SILK grid chokes seem to do a good job of replacing the 300B grid resistors in my Kaiju, maybe the 1100H is enough at the input of the BP?

What do people think?  Any risk of harm to the gear in doing this? 

I'd also try loading the TVCs via a grid resistor as well. And I'd appreciate suggestions as to values.  Since SILK reports that the TVCs have "very low output impedance", maybe I should try lower load values - e.g., maybe as low as 25K or 10K?

As always, all thoughts, speculations, etc. welcome.

Edit: And I should add that if the TVC box sounds good as a replacement for the BeeQuiet, I'd proceed to try them properly installed in the BP.

MTIA, Derek

[Paul Birkeland]

Just turn the BeeQuiet all the way up and use the TVCs between your source and BeePre.

[Doc B.]

FWIW a TVC was one of the things we tested when coming up with our own attenuator. We used Stevens and Billington. I think it came in about fourth place to my ear after our current design, a Sweet Whispers style attenuator and a most likely fake Alps Blue that somehow had some magic. The reasons were as you describe - nice midrange but lacking bass definition.

[Deke609]

Many thanks PB and Doc.

Just turn the BeeQuiet all the way up and use the TVCs between your source and BeePre.

I thought about that, but I want to test whether the TVC sounds better/different than the BQ resistors. My guess is that the BQ is built as two resistive voltage dividers in parallel (coarse and fine each comprising a divider) - so that at any attenuation setting there are always 4 resistors in the mix, two in parallel making R1 and two in parallel making R2 in series with R1 (I'm not a 100% about this b/c I haven't actually traced the operation of the BQ switches - but this is the only config I can think of that would work). So I need to remove the BQ from the circuit - either without replacement or replaced by a single resistor.

FWIW a TVC was one of the things we tested when coming up with our own attenuator. We used Stevens and Billington. I think it came in about fourth place to my ear after our current design, a Sweet Whispers style attenuator and a most likely fake Alps Blue that somehow had some magic. The reasons were as you describe - nice midrange but lacking bass definition.

I read your posts to that effect. Maybe it's a lost cause - but I'm going to give the Silk TVCs about a 100 hours before moving to the next testing step. It completely slipped my mind that these things are trafos and so may need some break-in. Late last night, the bass got a lot better and the thing has a bit of punch now - still far off from my BP1 but getting better.  Worst case scenario, I'll see if I have a buddy with SS system that's way too sharp around the edges and that doesn't need preamp gain - it might be perfect in such a setup.

My original idea was to experiment with fancy "silver tantalum" resistors in the BQ, but was scared off by the prices. But I think I'll try them in the Kaiju where I've installed a stereo stepped attenuator in place of the 100K linear pec pots with approx. 0dB, -6db, -9dB and -12db steps.  I can replace the -6dB step with another 0dB so that I have two side-by-side 0dB switch positions, one using a 100K Dale RN resistor and the other a 100K silver tantalum. Comparing a dB setting that uses two resistors (e.g., -6 dB) would be a better test of the sonic effect, if any, of the resistors, but I'd need identical resistor values in both RN and silver tantalum and the latter only come in a fairly small series of values -- but I will see if I can find matching values.

cheers and thanks, Derek

[Paul Birkeland]

so that at any attenuation setting there are always 4 resistors in the mix

All of the coarse resistors are always in the signal path and with the fine up all the way none of them are.  With the BeeQuiet turned all the way up, you'll have a nice high impedance load for your TVCs.

[Deke609]

All of the coarse resistors are always in the signal path and with the fine up all the way none of them are.

Thanks PB, but apologies, I am being dense here and can't make sense of this. I'll have to study a *Quiet attenuator. But I guess all I really want to know is how many BQ resistors does the input signal pass through when the BQ is all the way -- not just before the grid, but in total between rca input and return?

cheers and many thanks, Derek

[Deke609]

All of the coarse resistors are always in the signal path and with the fine up all the way none of them are.  With the BeeQuiet turned all the way up, you'll have a nice high impedance load for your TVCs.

Ah, I think I get it now.  So with Fine all the way up, the Fine attenuator is open circuit and drops out of the equation, leaving only the initial 35K input resistor in series with the 6 Coarse resistors that add up to 100K. So min. 7 resistors in the signal path. Tangent: so the BP with BQuiet has an input impedance that ranges from approx. 40K 50K [edit: couldn't even add 35 and 15 without making a mistake! - Jeez), with Fine all the way down, to 135K with Fine all the way up? Neat.  I didn't do the log calculations to figure out how the dB steps work, but I'll come back to that sometime soon b/c it's pretty cool and not how I expected the BQuiet to work.

Back to my experiment planning: 7 resistors compared to a TVC loaded by 1 resistor (or even none) should make for a good listening experiment. I think I can make a small rig for holding the new switch and TVCs that can sit on top on the BP1 chassis with connections running up through one of the many ventilation holes. So no need for dismantling the BP1 or for using 2 ft of interconnects to wire in the TVCs.  I should even be able to install a shorting switch for quick AXB comparisons. I'll work on that next wknd.

In the meantime, I followed your suggestion and put the TVC box between my DAC and the BP1. Bass and dynamics are back! But now it sounds like my usual setup with a bit of rounding off of the edges.  No surprise there b/c I'm stilling running the signal through 7/8 of the BQ resistors. But at least it sounds good again and I can enjoy listening while I "break in" the TVCs before next wknd's experiments.

Many thanks again, Derek

edited to correct my lousy grade 3 math skills

[Deke609]

After much head-scratching I figured out a switching rig to test the TVC against the BeeQuiet. I've been doing listening comparisons for a week. For detail/clarity, the supermalloy TVCs have the edge. But for spaciousness, dynamics and toe-tapping liveliness, I give the win to the BQ. My expectation biases were weighted the other way, so I trust my impressions.  I'm a bit shocked. I'm going to give it some more time b/c I've noticed that my ears are adjusting to the new level of detail, so maybe things will change a bit.  But the liveliness isn't changing. THe BQ just sounds more exciting -- and without having to listen closely. I'm a bit of a detail junky, but may have finally reached the "enough" point with my existing rig. I don't know that I want to listen as if my ear were right up against the strings of a cello or the drum skins, or to focus so intently on hearing all that detail. 

One thing seems clear: the sound signature of the BeePre (the amp) dominates over the signature of the attenuator -- which is what I would expect, but didn't know from experience.  Now I do. For my next experiment, I plan to add a BQ to the TVC box I slapped together to test the BQ against the TVC as a passive preamp. Just to hear the difference. I have a spare MourQuiet that can be put to this service.

Another thing that I hope to be able to test is the sound signature of the supermalloy cores of the TVCs. To my ears, supermalloy sounds a lot like the amorphous core OPTs I tested in my Kaiju - more detail, but a little too "polite" -- not enough attack/slam for my tastes. So it might be a core material issue. I'm trying to find someone to wind me Finemet TVCs, as I'm very pleased with the Finemet plate chokes and OPTs that I use in my Kaiju.  We'll see.

And I should add that the supermalloy TVCs make an excellent passive preamp. After about 200 hours, the low end came into its own and the whole thing was sounding quite nice. So I don't think the quality of the TVC is an issue. THE BQ just works really well.

So this is one of the rare tweaks that I've tested that I wouldn't recommend, unless you only get enjoyment from detail and don't care about the rest.

[Edit: and it just occurred to me that I've only tested the TVC unloaded (secondary open circuit). So for completeness I'll try it loaded with a 100K resistor.]

cheers, Derek

[johnsonad]

I tried them at one point too and they were quickly returned.  The sound wasn't for me. The BQ is an excellent attenuator. Nice write up.

[Paul Birkeland]

I plan to add a BQ to the TVC box I slapped together to test the BQ against the TVC as a passive preamp. Just to hear the difference.

The BQ is a pretty high impedance attenuator to just toss into a box and call a passive preamp.  When we offered this as a product, the resistor values were recalculated accordingly.

[Deke609]

Many thanks PB. That's helpful. Even though I still don't understand how/why it makes any difference where one puts purely resistive (non reactive) elements along a wire -- e.g., in the middle of a 3 ft interconnect, or at end (e.g., in the Kaiju). So there's something fundamentally wrong with my mental model of how electricity behaves in a conductor (no surprise there). I hope to make better sense of it one day.

Hmm. Now that I think more about it, perhaps the best way of testing the BQ as a passive would be to install it in the Kaiju in place of the 100K input resistors (that I use in place of the 100K linear PEC pots). Not sure I want to go to all that effort, since I'd also want to rig things so that the 100K resistors were in circuit when using the TVC.  Doable but would require some work.

Instead, I think I'll experiment with copper tantalum resistors in the BQ.  And if I also manage to find Finemet TVCs (not autoformers), I'll call it day on my experiments with attenuators. I can use the supermalloy TVCs for a remote controlled 2-channel TV/movies setup I have planned for the summer. I think the added clarity they seem to give to vocals might work well for movies. We'll see.

cheers, Derek

[Paul Birkeland]

Many thanks PB. That's helpful. Even though I still don't understand how/why it makes any difference where one puts purely resistive (non reactive) elements along a wire -- e.g., in the middle of a 3 ft interconnect, or at end (e.g., in the Kaiju). So there's something fundamentally wrong with my mental model of how electricity behaves in a conductor (no surprise there). I hope to make better sense of it one day.

Cables have capacitance.  Amplifier inputs have capacitance.  A passive control needs to be able to adequately drive these capacitances without incurring a bunch of HF rolloff.

[Deke609]

Thanks again PB.

... A passive control needs to be able to adequately drive these capacitances without incurring a bunch of HF rolloff.

I think get the capacitance thing. What I don't get is why the location of resistors along a wire (e.g., beginning, middle or end)  makes a difference.  In my (obviously faulty) mental model, a purely resistive element doesn't "drive" anything after it, the preceding stage (voltage/current source) does. Whereas, I can wrap my head around a reactive device like a transformer "driving" things. Maybe this is where the problem in my thinking is.

cheers, Derek

[Paul Birkeland]

What I don't get is why the location of resistors along a wire (e.g., beginning, middle or end)  makes a difference.

Doing some analysis may help.  Pretend you have two 200pF interconnect cables and an amplifier with 50pF input impedance.  If you have the two cables connected in series to make one long cable and all that going into the amplifier, that's 450pF.  If you have a 50K source impedance driving that load, where is the HF corner?

If you move the 50K source impedance to between the two 200pF cables and the 50pF amp load, what is the new HF corner? (for this we assume that there is a very low impedance source driving this whole arrangement, like the BeePre)

Now you can insert the 50K source between the two 200pF cables, so you have 200pF that you aren't driving and 250pF that you are driving.

[Deke609]

Excellent idea. Many thanks again PB. I will do that.  A quick mental sketch already revealed a stupid mistake: I had been thinking of the BQ as a resistor is series with the power amp load, not in series/parallel with it. But that only clears up part of the problem in my thinking. I will work through your example.

cheers and thanks, Derek