In search of a perfect attenuator

[Peter Daniel]

A year and a half ago, I posted some impressions on testing various potentiometers: So I did try all those different potentiometers

I came to conclusion that PEC pot appeared to be most musical. In reality it wasn't the most practical choice: first of all the tracking accuracy on some units could be off by 6dB between channels (although the pot was specified as 10%). Also, some listeneres didn't like the "mushy" quality of the sound, so in the end I was still using Nobles in my Integrated amps.

However, I spent some further time experimenting with attenuators and it seems that switching attenuators are much better than pots.

[Peter Daniel]

It seems like I've already found a perfect attenuator. It consists of a series, fixed resistor at amp's input and another, removeable resistor that sets the attenuation. In a picture, my current setup is presented. Please note that the socket is only temporary, it will be later replaced with much better grade, closer to amps' board.

All I need is a set of 6 resistors that I can plug in and depending on their values, I can adjust the attenuation in a 12dB range which is perfectly fine for most of my listening.

The quality of resistors is of great importance and in high resoulution system the differences can be easily observed. Specifically, for a GC amp, I'm choosing a nude Vishay in a series position and Caddock MK132 for shunt (Vishay does not work well here).

This setup is probably better than anything else presently available, although it's not particularly convenient

[Bas Horneman]

Hi Peter,

Where did you get the grayhill? I have a 7gang one..but know someone who is looking for grayhill switches like that.

Regards,
Bas

ps..I have not tried that many attenuators..and especially not in the same setup. But so far I like the DACT units and my Grayhill with intactaudio AVC. Next up....DiyHiFiSupply amorphous TVC's. On the same Grayhill (7gang) so comparison should be easy.

[Peter Daniel]

I think Brian got it as sample from Grayhill and he sent it to me. The one in a picture is 53M15-04-1-24S-F and you can probably order it from Grayhill directly.

[leadbelly]

Thanks for the info Peter!

In your listening evaluation of various pots and switched attenuators, did you find any correlation between things like poor tracking in a dual gang pot and poorer channel separation in a dual gang pot to overall listening preference?

[Bas Horneman]

The one in a picture is 53M15-04-1-24S-F and you can probably order it from Grayhill directly.

Thanks!

[Peter Daniel]

It's really hard to say, but when you get 6db channel imbalance you definitely loose imaging.

The Nobels that I'm using are pretty good with regards to tracking accuracy and I found only one that was obviously out of specs. However, none of the pots I tried have the level of resolution a quality fixed resistor offers.

[Peter Daniel]

It all started again, when I bought a used Placette RVC on the Audiogon. I've heard previously very positive comments on that attenuator, so finally I decided to give it a try.

Initially, inserted between NOS DAC and GC, I didn't really like it. It was a bit harsh, not liquid and kind of annoying. I still preferred the Levinson preamp as it was more laid back and tuneful.

Later, I completed the active buffer stage ( 6 ohm output imp.) and I plugged the buffer directly at the DAC output, and suddenly things changed. Placette became the best thing I ever had in my system, with a level of accuracy, resolution and openess I didn't expierienced before. it was not an issue of musicality anymore, things simply sounded right to the point that issues like tonality dynamics and musicality are not relevant any more.

I was kinda surprised.

[Peter Daniel]

Placette is based on a network of 14 resistors and 7 relays. It provides linear output voltage adjustement in 128 steps (if I'm not mistaken).

To see how good it is, I tried to "improve" on it with a circuit that consisted of a single series resistor inserted between two RCA jacks (input and output, series resistor being Vishay S102 or Caddock TF020) and the shunt element with DACT switch and nude Vishay across it. To my surprise, I couldn't improve on Placette with that circuit, contrary Placette was still better (in terms of resolution and opennes).

This leads me to conclusion that shunt element is equally important in a switching attenuator as the series counterpart.

[Peter Daniel]

Finally, I can fully agree that in a properly configured system an active preamp isn't really needed, even with a GC. I recommend the following link for more details: http://www.high-endaudio.com/RC-Linestages.html

Here are some further comments with regards to Placette: http://www.goodsoundclub.com/Forums/...postID=957#957