antilog

[patchell]

Ian Fritz wrote:

> I looked at this circuit some, too. There are approximations -- note the
> equation on the right side of the page that has <approximately equal> marks.
> I think there may be a couple of approximations there. Also I wondered about
> noise, with the OTA input being attenuated to such a low level. Still,
> probably worth having a closer look. And I still like tempco resistor
> compensation. Once you realize how accurate it is and that the statement you
> keep reading that it "can only work at one temperature" is *wrong*, then it
> becomes much more attractive.
>
>   Ian
>

    I have no doubts to the accuracy of Tempco's.  I still have about 20 of the
Q81K's (Tel Labs) that I bought over 20 years ago.  And the Paia circuit won't
be any better that a Tempco because of the fact that the transistors and the
sensor are two different parts.  When I did my electronic compensation circuit,
I went though several different iterations, but the one thing I wanted was the
temp sensor had to be one of the transistors in the exponential pair.  I was
probably being anal retentive, but, I wanted the compensation to track the
temperature exactly, and it was fun designing the circuit.  And J.H. was a big
help when he pointed out that you could use the diodes in a linearized OTA to
control the gain and get the 1/T response without putting the OTA in a feed back
loop.  So far, my circuit has not been any better than what I used to get with
Tempco's, but, it hasn't been any worse.

    Even though the Paia/National circuit might be cute, and it may even work as
advertised, for my own stuff, I am going to stick with either my circuit or use
tempcos.

>
> ----- Original Message -----
> From: "Jim Patchell" <patchell at silcom.com>
> To: <synth-diy at node12b53.a2000.nl>
> Sent: Thursday, August 10, 2000 9:13 AM
> Subject: Re: antilog
>
> >
> >
> > "Fraser, Colin J" wrote:
> >
> > >
> > > The LM13600 temperature compensation scheme looks really smart and might
> be
> > > the answer.
> > > How come nobody noticed it hidden away at the end of the data sheet all
> > > these years ?
> > >
> > > Colin f
> >
> >     Speaking for myself personally, it was because it was mislabed as a
> "log"
> > converter, so that was the first reason.  Second reason, if I had read the
> text
> > discribing it, I would have realized what it really was.
> >
> >     Paia seems to feel that it is good enough for a VCO.  I am going to
> wait
> > until I get my 9720 kit to reserve judgement.  I am mostly currious as to
> how
> > well it will actually perform the exponential function.  I am somewhat
> concerned
> > because the "temp sensor" is not exactly linear.  It very well me be that
> this
> > is not going to be a problem.  They seem to be operatin that OTA over a
> very
> > small input range which should keep it pretty linear.
> >
> >     -Jim
> >
> >

--
 -Jim
------------------------------------------------
* Visit:http://www.silcom.com/~patchell/
*-----------------------------------------------
*If you didn't buy a home in Santa Barbara,
* You didn't pay enough!
------------------------------------------------