resistive vs inductive attenuator

Magnet wire resistors, enough to make any engineer scream.
Post Reply
drdna801

resistive vs inductive attenuator

Post by drdna801 »

I am wondering if the difference heard between the autoformer attenuator and a stepped resistor attenuator is essentially the material used, rather than the difference between resistors and coils per se.

This is suggested by experience with materials changes in other areas of audio (e.g., teflon vs electrolytic capacitors, etc) as well as commentary on non-inductive copper wirewound resistors having a similar clarion-like character to the autoformer attenuators.

Possibly a stepped shunt attenuator using noninductive wirewound copper resistors might offer the best of both worlds: a clear copper path and fixed impedence load presented to the source.

What do you think????
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: resistive vs inductive attenuator

Post by dave slagle »

drdna801 wrote:I am wondering if the difference heard between the autoformer attenuator and a stepped resistor attenuator is essentially the material used, rather than the difference between resistors and coils per se.
i have also thought this, all you have to do is take the core out of an autoformer and you have a 50 ohm tapped pot with a small amount of inductance and some stray capacitance. I have tried to convince a few people to simply remove the core altogether and report back but have yet to get a taker yet.

I'll be clear, as resistors go, the magnet wire ones are really crappy. whatever you do to reduce capacitance increases inductance which opens a whole can of worms. Don't get me wrong, nothing beats a resistor as a tube load on paper, but chokes (with all of their flaws) still sound better. The only conclusion I can draw is a resistor by design needs to be a crappy conductor to keep the parasitics down, so maybe we need to accept some of the parasitics in order to use a better conductor!?!
This is suggested by experience with materials changes in other areas of audio (e.g., teflon vs electrolytic capacitors, etc) as well as commentary on non-inductive copper wirewound resistors having a similar clarion-like character to the autoformer attenuators.
I honestly dunno, when dowdy asked me to do it I said why not. I know his methodology and trust his ears and his results interest the hell out of me. I wish it were as simple as material change but I always fall back to experience and common sense. The thing that smacks me in the face is we strive for ultra-pure wire for a few inches of hookup wire then JAM the signal through a decidedly crappy conductor called a resistor. Am I the only one that sees this as ironic?
Possibly a stepped shunt attenuator using noninductive wirewound copper resistors might offer the best of both worlds: a clear copper path and fixed impedence load presented to the source.
in this situation given the parasitics, the clear path will meet up with lots of problems. I would suggest some work in spice to see how things work out.

dave
Guest

A multi-tapped resistive attenuator

Post by Guest »

dave slagle wrote:i have also thought this, all you have to do is take the core out of an autoformer and you have a 50 ohm tapped pot with a small amount of inductance and some stray capacitance. dave
Thanks for your response. I don't know what the disadvantage would be to using it this way? Perhaps it would even be possible to wire two linear tapped pots in an inverted configuration in a shunt type attenuator circuit to present a constant impedence to the source? I suppose it depends on what the accompanying capacitance and inductance will be. What do you think?

Adrian
drdna

Post by drdna »

Dave --

Wondered if you had any thoughts on using two multitapped unncored autoformers versus multiple resistors to wire up a ladder configuration as a attenuator to provide constant impedence to the source?

Alos, would it be advantageous to use lead wire instead of copper; wouldn't this be higher resistance and therefore require less length/size for resistors?

Adrian
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: A multi-tapped resistive attenuator

Post by dave slagle »

Anonymous wrote: I don't know what the disadvantage would be to using it this way? Perhaps it would even be possible to wire two linear tapped pots in an inverted configuration in a shunt type attenuator circuit to present a constant impedence to the source?
Adrian
I don't think so. Constant impedance (t-networks) attenuators only get "close enough" in reality with ideal devices (the resistor). using anything with additional L or C factors will only assure you of a math nightmare.
would it be advantageous to use lead wire instead of copper; wouldn't this be higher resistance and therefore require less length/size for resistors
i think the whole point of this thought line was using a "good" conductor and accepting the faults (L and C) that come with it. A resistor is just a crappy conductor, and opening the can of worms of the "best sounding" crappy conductor is too ironic for the scientist in me.

dave


dave
dirkwright
Posts: 56
Joined: Sat Jul 25, 2015 9:50 pm

Post by dirkwright »

The advantage of an AVC is low noise, as compared to a resistive attenuator. With an AVC, the noise is so low (apparently equivalent to less than a 50 ohm resistor), that you should be able to hear deeper into the music than with a 10k pot, provided the distortion of the AVC is very low as well. I doubt there are very many sources that can drive a 50 ohm resistive attenuator, so I don't think removing the core of an AVC is a good idea.
Daniel
Posts: 45
Joined: Fri Jan 10, 2014 4:50 pm

Post by Daniel »

I don't think Barkhausen noise is the crucial factor here.
dirkwright wrote:The advantage of an AVC is low noise, as compared to a resistive attenuator. With an AVC, the noise is so low (apparently equivalent to less than a 50 ohm resistor), that you should be able to hear deeper into the music than with a 10k pot, provided the distortion of the AVC is very low as well. I doubt there are very many sources that can drive a 50 ohm resistive attenuator, so I don't think removing the core of an AVC is a good idea.
dirkwright
Posts: 56
Joined: Sat Jul 25, 2015 9:50 pm

Post by dirkwright »

I thought I was talking about just the thermal noise from the difference in DCR. There's something like a 200:1 ratio between a 10k pot and Dave's AVC, so the thermal noise is going to be proportionally lower with an AVC.
Daniel
Posts: 45
Joined: Fri Jan 10, 2014 4:50 pm

Post by Daniel »

Sorry, I am a bit scatterbrained due to the heat here (thermal noise inside my brain??) and mixed up thermal and Barkhausen. Still I believe the thermal noise to be the least factor regarding the differences in resistive and inductive attenuation.
dirkwright wrote:I thought I was talking about just the thermal noise from the difference in DCR. There's something like a 200:1 ratio between a 10k pot and Dave's AVC, so the thermal noise is going to be proportionally lower with an AVC.
dirkwright
Posts: 56
Joined: Sat Jul 25, 2015 9:50 pm

Post by dirkwright »

Well, since I intend to use an AVC in my production headphone amp, with an active buffer driving the AVC, then most of the issues I read about here on the forum are not relevant to me. I'm not concerned very much with impedance, as long as the AVC has high enough input impedance that my buffer can drive at low distortion.

I chose an AVC over a TVC (like from Sowter), because the distortion for an AVC should be far lower (like 10 times lower) than for a TVC. This is why McIntosh power amps can still produce very low distortion power while using an autoformer. The autoformer is a low distortion device, but I would like to see some numbers on THD some day...

So, my main concerns are thermal noise and distortion. For those using these AVC's in a different situation, you will have different concerns.

It seems from my simulations that when the source impedance is nearly zero and the load impedance is extremely high, then distortion and bandwidth are never a problem for either a TVC or an AVC.
Post Reply