[sdiy] programmable pots

[john mahoney]

> >The current versions are only availabe with (at most)
> >100 taps, so quantizing would be a problem...
>
> As a work-around, could you use more than one, to get more resolution?

Here they are, called "digitally programmable potentiometers." Add DPP to
your acronym list, folks.
    http://www.mouser.com/catalog/615/202.pdf
They are non-volatile, which is quite handy. You can use them as digitally
controllable trimpots, for example.

There are 256-tap units; all those are 28K Ohms.
The 32- and 100-tap models are available in 10K, 50K and 100K.
Precision is not specified. <assumption> Presumably it is is pretty tight. A
100-tap unit needs greater than 1% precision, otherwise why have 100 taps?
</ass.>

The math (or "maths" in British):
    32-tap 100K resolves to 3.125K Ohms per step
    100-tap 100K resolves to 1K Ohms
    100-tap 10K resolves to 100 Ohms
    256-tap 28K resolves to 109 Ohms

You could combine a 100K/32 and a 10K/100 to get the equivalent of 1000
steps of 100 Ohms each, approx. 10-bit resolution. This wastes some
resolution (only using 1K of the 10K unit) but requires no special tricks.

Maybe the PhDs can figure out a way to use 2 of them, one at full strength
and the other through a divider to lower its effective resistance. Is this
possible? Approx. 13- or 14-bit resolution would then be achievable, like
so:

Primary: 100-tap 100K provides 100 steps of 1K
Secondary: 100-tap 10K divided by 10 (to yield 1K) provides 100 steps of 10
Ohms each. Total of 10,000 increments.
    or 256-tap 28K divided by 28 (again, to yield 1K) provides 256 steps of
4 Ohms each. Total of 25,600 increments.

Again, I dunno about this divide by 10 or 28 stuff. :-)
--
john


----- Original Message ----- 
From: "Glen" <mclilith at charter.net>
To: "Cornutt, David K" <david.k.cornutt at boeing.com>;
<synth-diy at dropmix.xs4all.nl>
Sent: Friday, August 15, 2003 3:36 PM
Subject: RE: [sdiy] tubes, MOVs, etc.


> At 03:21 PM 8/15/03 , Cornutt, David K wrote:
>
> >The current versions are only availabe with (at most)
> >100 taps, so quantizing would be a problem...
>
> As a work-around, could you use more than one, to get more resolution?
>
> later,
> Glen Berry
>