diyAudio (
-   Analog Line Level (
-   -   Digitally controlled pot distortion? (

Tenson 13th January 2013 01:58 PM

Digitally controlled pot distortion?

I need to make a volume control for a stereo balanced signal. Since 4-gang pots are difficult to get, I have been considering using a digital potentiometer IC.

There are lots about; Wolfson WM8816, Maxim DS1844, National LM1972 etc..

My question is whether these can really be thought of as passive resistive elements or do they introduce distortion of their own (say though transistor switching of the resistor ladder)? The WM1816 datasheet gives a THD+N figure of -100dB but maybe this is really down to the op-amp buffer?

nigelwright7557 13th January 2013 03:19 PM

I built a digital mixer using Microchip digital pots and it sounded fine.
So long as you keep the voltages between 0 volts and 5 volts then there wont be much distortion. I found I had to keep inputs and output biased around 2v5 or I got terrible distortion.

abraxalito 13th January 2013 03:29 PM

The switches are the most non-linear part in my understanding - CMOS switches have a non-linear voltage vs resistance curve. This should mean the distortion reduces as the signal level goes down. Whilst the output opamp won't show much THD, chances are the audio performance is limited by low level IMD.

counter culture 13th January 2013 04:37 PM

You can get passive-type digipots, but they tend to have quite high distortion.

You're better to go with the chips intended as volume controls such as the Wolfson you named or the TI PGA23xx series or Cirrus equivalent. I think the TI chips actually have the best paper performance, but these things are not set in stone.

The volume controls are not purely passive, they generally offer gain as well as cut, this is provided by internal opamps. This is not a good reason to reject them as the levels of distortion they achieve are almost certainly sub-audible. Some are only intended to work with maximum 5V levels, but some will work with +/-15V.

A device that shows low THD is very unlikely to have high IMD.

Chances are that the IMD is insignificant.

Tenson 14th January 2013 10:22 AM

Thanks for the replies chaps. Looks like the LM1972 is the best from the spec sheet. However I know nothing about programming micro-controllers! I think that is something to spend a bit of time reading up on :)

On the other hand, if I got a order of quad linear pots (Vishay P9) would people be interested in buying some off me? Obviously linear pots can be make log with the addition of a law faking resistor. I guess I'd go for either 1K or 5K.

qusp 16th January 2013 07:35 AM

the Ti PGA4311 A GRADE type. has better numbers 120db DNR and 0.0002% THD+N

its also 4 'deck'

abraxalito 16th January 2013 07:47 AM

That part has 10MHz CMOS opamps on-board so the THD+N figure is certainly not to be trusted as a FOM for audio SQ.

qusp 16th January 2013 08:43 AM

i'm not a fan either, but I dont like any of them since they basically become the dominant factor of your analogue stage whether you like it or not. however the spec has been tested, I suggested it as spec seemed to be the metric under discussion and he needed a 4 channel part

All times are GMT. The time now is 03:54 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2