[sdiy] Log/exp conversion

patchell patchell at silcom.com
Sun Dec 23 22:15:58 CET 2001



Jonathan Lippard wrote:

> At 02:28 PM 12/22/01 -0500, media at mail1.nai.net wrote:
>
> >While I'm sure there are much more VCO-specific designs on the web, you
> >might want to go to National's website and download application notes AN-30
> >"Log Converters", and AN-311 "Theory and Applications of Logarithmic
> >Amplifiers".
>
> While we're on the subject, has anyone ever used the exp converter from the
> app note for the LM13600? (Sorry, I don't remember the number.)  It's the
> one that uses one of the linearizing diodes for temperature compensation,
> and it's used in PAIA's 9700 modules...or has anyone used the 9700 and can
> comment on it?  I like it for the obvious reason that it doesn't need a
> tempco resistor.  I can't build one right now though, or likely for some
> time to come...if anyone can shed light on its quality, I'd appreciate it...
>
> -Jonathan

    Acutally, that circuit does not use one of the linearizing diodes as a
temperature sensor.  What it uses in the transconductance of the OTA as the temp
sensor, but, still, basically the same thing (the Gm of an OTA has the same gain
dependency on temperature that an exponential converter does).

    The circuit does work.  And, it should work just as good as a tempco type
compensation.  But, here are the things you need to be aware of:

    The OTA needs to be physically connected to the exponential pair (just like
when you use a tempco).

    The OTA will introduce it's own noise into the mix, although, this will give
your pitch a little more randomness, so that might actually be desireable to
some degree.

    The OTA will introduce a little more nonlinearity into the exponential
function.

    Now, you can also so a similar thing more directly using the exponential
transistor itself as the temp sensor.  If you go to my web page below, you
should eventually find the circuit diagrams...
    By using the expo transistor itself for the temp sensor, there is no temp
lag, plus, no messing around with glueing things together.

    There was also link posted to an article in EDN that should a third way to
do this.  I don't have the link anylonger, but you might be able to search
around at EDN and locate it.  A very interesting article.

    And there are a few other members on this list that have their own way of
doing the same thing...can't remember the URL's off the top of my head for those
either...I am sure they will stand up an make themselves counted.

--
 -Jim
------------------------------------------------
* Visit:http://www.silcom.com/~patchell/
*-----------------------------------------------
*I'm sure glad Merry Christmas comes just once a year
* -Yogi Yorgensen
------------------------------------------------





More information about the Synth-diy mailing list