[DZigas]

I thought this forum might find Paul McGowan's comments interesting, so I posted them from his forum to this one. I'm not technically qualified to evaluate what he is saying (and only minimally follow it). As I mention below, I have a DIY passive built around an autoformer made by Dave Slagle.

I'd be interested to hear what others in this forum think. -- David

Author
  Topic: Attenuation
 
Dzigas
 posted 10/15/2004 8:42:28 AM  

--------------------------------------------------------------------------------
Paul -- In your product lit for the Gain Cell, you state that "All volume controls, regardless of their design, attenuate or lower the signal level by redirecting the audio to ground."

I own a diy passive, built by Dave Slagle and Larry Moore. It uses an autoformer, rather than resistors, to attenuate inductively. My limited understanding is that the autoformer, while lowering the voltage as it attenuates, transforms the signal into current, rather than "throwing it away" as heat. Is this correct? Does it follow your description above?

I can certainly state that the sonic results are very pleasing. -- David  
 
Paul McGowan
 posted 10/15/2004 9:33:58 PM  

--------------------------------------------------------------------------------
Well, I guess I would have to stand corrected on that one. However, using an autotransformer for this purpose is probably the last thing I would do as you're not throwing away volume but you are reducing amplitude by basically choosing different taps along the transformer. It's an interesting idea, not one that I would choose.

I am glad it's working for you, I would never have guessed it'd sound good. I'd like to try one sometime just to see. I can only imagine the frequency response must be limited and there's probably a lot of phase shift at the extremes.  
 
Dzigas
 posted 10/16/2004 6:06:10 PM  

--------------------------------------------------------------------------------
>>>>I can only imagine the frequency response must be limited and there's probably a lot of phase shift at the extremes.

Paul -- What is it about an autoformer that leads to this observation? I'm curious. Thanks. -- David  
 
Paul McGowan
 posted 10/17/2004 3:11:38 AM  

--------------------------------------------------------------------------------
All transformers are magnetic coupling devices and only work on AC. So, for example, the Gain Cell is flat to below 0.1 Hz. That's so close to DC that it won't even register on a transformer of any kind.

A really good transformer, be it auto or otherwise, would be considered great if it's -3dB point was 1Hz - but that means it starts to roll off (phase shift and amplitude loss) well before that, maybe 20Hz or so. And, a transformer that is flat to 20kHz is really good, but they roll off at maybe 50kHz, which means phase shift again.

A really cool input transformer, like a Jensen - similar to what Jeff Rowland uses on the input of his amps, is pretty flat within the 20Hz to 20kHz range but not even close when compared to a Gain Cell or a good buffer or even a mechanical volume control.

Having never seen the device you're referring to it's not fair for me to suggest I know its specs, but I do know that any transformer is a magnetic coupling device and has to have AC to work - so as it appears DC it works less and less - and is certainly bandwidth limited on the other end of the spectrum as well.

Hope that's as clear as mud?  
 
Dzigas
 posted 10/17/2004 8:01:31 AM  

--------------------------------------------------------------------------------
Thanks for the info!

[John Chapman]

Hello!

This is kinda fun to read and thanks for posting.  The reason it is fun to read is because his reaction to the idea of using a transformer for a level control is EXACTLY the same as mine was. A local friend asked me about it and I thought is was a crazy idea - until I thought about it for a while.

For reference the tx102's are flat from below 10Hz (that is as low as my Audio Precision analyser goes so it is actually flat lower than that) to 100Khz. I have measured phase for someone a while back and posted the results either here or at the old HD forum - can't recall exactly but it was decent.  It took some doing to make the transformer behave this well and from what I know of Daves autoformer it also measures very well. Getting the core size and number of windings large enough to do a good job of low frequencies makes that upper frequency range a challenge.  

For a simple signal path and a passive design I think these transformer (or autoformer) based units do pretty well   .

Comparing the outer reaches of specs (like 0.1Hz or Mhz) does not teach much about how stuff sounds.  Kinda like 0.001% THD vs. 0.00001% THD. I can see the point that he makes however - it would likely not make sense to go to the trouble of building an ultra wide bandwidth actve circuit and then feed it with a device with less bandwidth.

If your philosophy is ultra wideband active circuits transformers would not appeal to you....I was comming from resistor based passives when I first discovered TVC based passive devices and the way the impedance translates was the thing that sold me from a technical point of view. Put a length of cable after a TVC and once the level is a few steps from the top it will measure a wider bandwidth than a pot.  I have run the test here for folks wanting to run long cables and it is easy to see. Unless you are carefull to keep the impedance of the cable (and circuits) following a resistive pot very low it may roll off sooner than a transformer or autoformer.  


Thanks again for posting this.

John Chapman
www.bentaudio.com

[DZigas]

Thanks, John. Paul is really one of my favorite people in the high end biz. Note that despite his skepticism, he was still interested in hearing one. The mark of an open mind. -- David

[giorgino1]

Hi

You might be interested in sending the specs to Paul:
http://www.stevens-billington.co.uk/page102.htm

I'm also going through my "research" phase before reaching for my cheque-book  

Regards

George