Page 1 of 4

Stepped attenuators

Posted: Fri Mar 09, 2018 4:19 pm
by karatestu
Been looking at the vast array of stepped attenuators on hificollective :shock:

I am going to go for a couple of mono 10K shunt models. Not decided on how many steps I should be ordering or what resistors. I think I have read that the audio note tantalums are the ones to go for, especially as they fit the first 6 values with shinkoh tantalums.

My questions;

1)How many steps -20 something or forty something
2)Some have circuit boards - avoid ?
3)Which resistors - shinkoh, audio note ? Oh and are 1W resistors better than 0.5W ?

Thanks, Stu

Re: Stepped attenuators

Posted: Fri Mar 09, 2018 6:31 pm
by keepitsimplestupid
If you want purely linear resistors then any decent metal film will do the trick. Anything else will add some sort of flavour. I've used Shinkohs in a stepped pot before and they sound great, not without their own sound but very minimally so. How many steps depends on how you listen and how fine you like your control of volume. I like a lot of small steps to start with and then am generally happy with 1-2db per step after mid volume levels. Whatever lights your candle basically.

IME a well designed circuit board isn't going to degrade the signal audibly. If what you want is simply the purest attenuator then go no board, but then its probably not Shinkohs you're after but something more like the Goldpoint stepped attenuator which to my ears had no sound of its own, having tried various audio note, noble, alps, DACT pots and a couple of relay switched stepped units.

Re: Stepped attenuators

Posted: Fri Mar 09, 2018 6:35 pm
by Dr Bunsen Honeydew
Speak to Nick at Collective, he is one of the few good guys, good prices good knowledge. He will talk you through the benefits, plus you can save money by buying then as a kit and doing the soldering.

Re: Stepped attenuators

Posted: Fri Mar 09, 2018 6:55 pm
by karatestu
Ok, thanks. I will give Nick at hfc a call. There is a bewildering array of them. More steps is probably better but then i think of all the resistors that i will never be using as i dont listen too loud and the room is smallish.

Re: Stepped attenuators

Posted: Fri Mar 09, 2018 7:06 pm
by Lurcher300b
If you want purely linear resistors then any decent metal film will do the trick.
Except its entirely possible to measure the non linear distortion created by "purely linear resistors", they are not the same, it can be measured and its not foo. As to the difference being audible I leave that up to the reader to decide.

Re: Stepped attenuators

Posted: Fri Mar 09, 2018 11:57 pm
by Rectified
Years ago when i was into Hi-Fi i built a passive Pre with the Sfernice 'P11VYN' Cermet Linear pot & Vishay Bulk foil Resistors LFR's. Was an improvement over a few off the shelf passives at the time..

Re: Stepped attenuators

Posted: Sat Mar 10, 2018 12:26 am
by Dr Bunsen Honeydew
I used them for a long time but they had severe problems for production use, in the end I replaced them with my standard pot I use now. Back then cermets were only available in linear law so they had to be blagged with a resistor to try and simulate log law, it made them not track very well and getting them both to start at the same time was next to impossible. They were also very noisy as cermet track is not as smooth as plastic film and people thought they were faulty, and the noise got worse over time, but not from wear as with a plastic film or carbon track.

I prefer the plastic film jobby I use now. The SA I use now is much better than both of them

Re: Stepped attenuators

Posted: Sat Mar 10, 2018 3:25 pm
by Thermionic Idler
This is probably academic and a bit overkill for what you're after, but just for interest, when I was planning my recent linestage build, I decided I wanted stepped attenuation AND full remote control, which somewhat narrowed my options!

In the end I built a linestage around this Arduino-based processor and it controls a set of relay boards, this does the I/O selection and this does the stepped attenuation - 256 steps at 0.5dB intervals if you go for the full 8 relays. Mine is an active linestage but it's entirely possible to build a passive stage with it too. I populated mine with Vishay RN60s, I would have gone for Takman but certain values I needed weren't available.

Re: Stepped attenuators

Posted: Sun Mar 11, 2018 10:16 am
by karatestu
Thermionic Idler wrote: Sat Mar 10, 2018 3:25 pm This is probably academic and a bit overkill for what you're after, but just for interest, when I was planning my recent linestage build, I decided I wanted stepped attenuation AND full remote control, which somewhat narrowed my options!

In the end I built a linestage around this Arduino-based processor and it controls a set of relay boards, this does the I/O selection and this does the stepped attenuation - 256 steps at 0.5dB intervals if you go for the full 8 relays. Mine is an active linestage but it's entirely possible to build a passive stage with it too. I populated mine with Vishay RN60s, I would have gone for Takman but certain values I needed weren't available.
Thanks for that, very interesting. It seems a bit hi tech for me though and goes against what i am trying to do - keep it simple.

Thanks for posting it though

Stu

Re: Stepped attenuators

Posted: Mon Mar 12, 2018 12:37 am
by Thermionic Idler
karatestu wrote: Sun Mar 11, 2018 10:16 am
Thermionic Idler wrote: Sat Mar 10, 2018 3:25 pm This is probably academic and a bit overkill for what you're after, but just for interest, when I was planning my recent linestage build, I decided I wanted stepped attenuation AND full remote control, which somewhat narrowed my options!

In the end I built a linestage around this Arduino-based processor and it controls a set of relay boards, this does the I/O selection and this does the stepped attenuation - 256 steps at 0.5dB intervals if you go for the full 8 relays. Mine is an active linestage but it's entirely possible to build a passive stage with it too. I populated mine with Vishay RN60s, I would have gone for Takman but certain values I needed weren't available.
Thanks for that, very interesting. It seems a bit hi tech for me though and goes against what i am trying to do - keep it simple.

Thanks for posting it though

Stu
No worries - I figured it probably wouldn't quite fit the KISS paradigm! There's a long thread on Audio Talk detailing its construction and my ongoing OTL project which you've probably seen already.