PAiA VCO

[Jim Patchell]

René Schmitz wrote:

> Hi Ian, Jim and List.
>
> I have tried to get behind the secret of this circuit. As a first
> step I redrew it, and at some point the circuit started to look
> familiar to me. Now I see much clearer. Infact it looks like a version of
> the exponential current source in the National Semi. datasheets for the
> LM13600. Which even features an analysis of this circuit. According to
> their maths the expo scale factor is independant of T. (So the gm is really
> the tempco, not the diode.) At least theoretically this will cancel,
> although in practice thermal coupling considerations apply. (Which was the
> reason why I used four trannies on one die when I did a solid state
> compensation.)
>
> Bye,
>  René
>

    Mystery solved....Thanks René.

    And, on the other hand, this pretty much puts the circuit in the public
domain.  I must have looked at that circuit a million times, never gave it more
than a passing thought.

>
> At 08:16 08.08.00 -0600, Ian Fritz wrote:
> >Hi Jim --
> >
> >Yes, this is all very curious. I hope to find time to wire this thing up and
> >measure its temperature dependence. I notice there is no drift spec listed.
> >
> >  Ian
> >
> >----- Original Message -----
> >From: "patchell" <patchell at silcom.com>
> >To: <synth-diy at node12b53.a2000.nl>
> >Sent: Monday, August 07, 2000 8:55 PM
> >Subject: Re: PAiA VCO
> >
> >
> >>     I just took a good look at the circuit.  I also redrew it , which I
> >find
> >> helps a bit.  But after I looked at it I saw something else that I find
> >> curious.  In the FAQ it said it was using the diode in the linearizer for
> >a
> >> temp sensor.  I don't think so.  In order to use a diode as a temp sensor,
> >you
> >> need to maintain a constant current through it.  The way the circuit is
> >set
> >> up, very little if any current will be flowing through that diode.  In my
> >> opinion, the tempco would be generated by the temperature dependence of Gm
> >of
> >> the transconductance amplifier.  Even if the diode were forward biased,
> >it's
> >> tempco would be canceled out the tempco of the Gm.  I depend on this for
> >my
> >> own electronically compensated exponential converter.
> >>
> >>     Time to break out the superstrip, I guess.  I am going to get to the
> >> bottom of this puzzel.
> >>
> >
> >
> >
> --
> uzs159 at uni-bonn.de
> http://www.uni-bonn.de/~uzs159
>
>