advertisement


High-quality potentiometer or attenuator for DIY passive pre-amp?

I've got both TKDs. The bigger gives a very slight improvement over the 601, maybe that's due to better matching?

The soundstage is a bit more real - like when you upgrade the clock in your CDP, but not THAT huge a difference.
 
Last edited:
Just to add more confusion, I preferred the Noble pot to the TKD. And prior to that I preferred the Panasonic precision pot to the Noble. I now wish I'd tried to fix the panny, rather than bin it :(
 
If/when I ever get round to it I'm planning on a 256 step reed relay based attenuator. I've designed the electronics and gathered together the relays and the IC's... Fook knows if I'll ever do it though... so many potential projects and new products and work for customers and this is hardly high priority.
The concept does have an issue in that it has constant ish input and output resistance and the two are the same figure... which pretty much demands a buffer at either input or output. A compromise of around 10K and no buffer could work OK I guess but would be a bit limiting in compatibility with other equipment. A mega quality buffer at the input leaving the output to be a nice purely resistive 1K ish (varies from about 900R to 1.1K ish depending on setting of attenuator and in a "pseudo random" looking [it's not] manner) is front runner. I guess this glitch is why some of the commercial relay attenuators have loads of relays and only say 64 steps! This works on the R/2R principle and could be termed a crude DAC. It gets you 256 steps with 8 relays though.
 
I have a leg endery Panasonic pot. It's no longer in use as I have digital attenuation (but it's not for sale), I seem to remember it's a bit better than an Alps blue.
There are bigger fish to fry elsewhere.
 
Arkless Electronics said:
If/when I ever get round to it I'm planning on a 256 step reed relay based attenuator. I've designed the electronics and gathered together the relays and the IC's... Fook knows if I'll ever do it though... so many potential projects and new products and work for customers and this is hardly high priority.
The concept does have an issue in that it has constant ish input and output resistance and the two are the same figure... which pretty much demands a buffer at either input or output. A compromise of around 10K and no buffer could work OK I guess but would be a bit limiting in compatibility with other equipment. A mega quality buffer at the input leaving the output to be a nice purely resistive 1K ish (varies from about 900R to 1.1K ish depending on setting of attenuator and in a "pseudo random" looking [it's not] manner) is front runner. I guess this glitch is why some of the commercial relay attenuators have loads of relays and only say 64 steps! This works on the R/2R principle and could be termed a crude DAC. It gets you 256 steps with 8 relays though.
Click to expand...

Unless I'm missing something, the 256 steps would be linear? So in any particular context many of the steps are subjectively redundant. A log attenuator might inevitably have fewer steps, but I guess each step is more useful. Not sure how many steps needed in order not to miss the continuum provided by a log pot?
 
gustavm said:
Unless I'm missing something, the 256 steps would be linear? So in any particular context many of the steps are subjectively redundant. A log attenuator might inevitably have fewer steps, but I guess each step is more useful. Not sure how many steps needed in order not to miss the continuum provided by a log pot?
Click to expand...

It would be linear yes but I fail to see any connection between this and anything else in your post I'm afraid...
It's 256 steps because it's so easy to do so using this technique... 6 relays = 64 steps, 7 = 128, 8 = 256 etc
 
Arkless Electronics said:
It would be linear yes but I fail to see any connection between this and anything else in your post I'm afraid...
It's 256 steps because it's so easy to do so using this technique... 6 relays = 64 steps, 7 = 128, 8 = 256 etc
Click to expand...

As you well know, normally volume pots are logarithmic because the ear's perception of loudness is logarithmic. A linear volume pot tends to have all its useful range up at the top end. So I would expect the same to be true of a linear step attenuator - in other words, many of the steps would never be used in practice.

I was then just speculating how many steps in a log attenuator design to be sufficient not to miss the full flexibility of a log pot.
 
gustavm said:
As you well know, normally volume pots are logarithmic because the ear's perception of loudness is logarithmic. A linear volume pot tends to have all its useful range up at the top end. So I would expect the same to be true of a linear step attenuator - in other words, many of the steps would never be used in practice.

I was then just speculating how many steps in a log attenuator design to be sufficient not to miss the full flexibility of a log pot.
Click to expand...

OK I see your jist on that. It's not ideal I guess but the number of steps available should mean decent control of perceived loudness.
In a line level pre I've designed and built recently I added log faking resistors to an already log pot to stretch out the useful control range as I didn't want to reduce the gain of the pre and I'm using with a rather sensitive power amp. This worked well in practice.
 
64 db total attenuation is enough, ill settle for 128 half db steps. I listen quietly at night sometimes, so with a hot xlr out dac that I like to run digitally at 0db the step size is important.
 
Tumeni Notes said:
Thanks for all the input, everyone.

I haven't pulled the trigger yet, was half thinking of spoiling myself for New Year with the Tisbury, off the shelf, rather than building, but I see that one response above is unenthusiastic about their design ....

https://www.tisburyaudio.co.uk/mini-passive-preamplifier
Click to expand...

I don't know if how relevant it is but I recently tried one of those as a stopgap for a few weeks. I got it partly because I have an LDR hybrid pre in another system which sounds excellent and liked the idea of a passive pre which was basically just a volume control. But when I then replaced the Tisbury with a traditional line preamp, the difference (to the Tisbury's detriment) was huge. (I couldn't judge before because the RIIA was also new). So I don't know how much better you could do for how much money or by what means, but I wouldn't get that Tisbury if I were you.
 
The Valab ones sound better than the pot in a 62, and a Goldpoint sounds even better.

Pete
 
Surely you have to consider if the output impedance of the source is low enough/the input impedance of the power amp high enough to make a passive pre a good choice in the first place. In less than ideal circumstances with a impedance 'mismatch' you'd expect an active pre to superior. Conversely, there's not going to be anything more transparent than a passive if the situation is well suited. Of course some pre's act as subtle tone controls/flavour the sound in a way that people like but that's another matter.

A passive works spectacularly well with my Leak ST20 which has a whoppingly high 1 mega ohm input impedance.
 
Tumeni Notes said:
Thanks for all the input, everyone.

I haven't pulled the trigger yet, was half thinking of spoiling myself for New Year with the Tisbury, off the shelf, rather than building, but I see that one response above is unenthusiastic about their design ....

https://www.tisburyaudio.co.uk/mini-passive-preamplifier
Click to expand...
I had a Tisbury and liked it. I found the increments a little coarse, though (about a dozen?) so eventually I replaced it with a Khozmo which has 64. But as a bonus the latter is much bigger and more expensive, which of course means it is also much better.
 
Unfortunately much of what is repeated almost daily about passive attenuators and their use is is simply wrong...
The stuff our own Mike P said above is kind of "the standard mantra" on them but unfortunately just repeated the main errors. Sorry but explained it all in detail too many times before so won't repeat it yet again...
 
You must log in or register to reply here.


advertisement


Back
Top