[sdiy] quadrature sine wave oscillator question

[Czech Martin]

I always wonder why jfets are used for AGC,
and not LDRs.
Perhaps someone came up with a jfet servo loop
and all the others just copied it?

The distortion the LDR causes should be way lower
than those of the jfet. Also diode forward conduction
problems simply do not exist.

Perhaps the useable range of the LDR is too small?

Applies for quadrature as well as Wien bridges.

m.c.

-----Original Message-----
From: patchell [mailto:patchell at silcom.com]
Sent: Dienstag, 25. März 2003 16:11
To: Sdiy (E-mail)
Subject: Re: [sdiy] quadrature sine wave oscillator question


    There is an article in the Nov/Dec 1980 issue of polyphony, and when
I built this circuit, I got about 0.5% distortion, although, that was
mostly due to the CA3280's that were used to control the frequency.

    If you only need a fixed frequency, seems like I recall I managed to
get about 0.1% distortion using zener diodes as shown in the Analog
Devices Nonlinear Circuits Handbook.  Some other experiments I did with
controlling the amplitude of oscilation susgests that it is best to get
the circuit as close to oscilation as posible without the AGC, and when
the AGC is connected in, it has the smallest amount of influence.

Czech Martin wrote:

> has anybody dealt whith those?
>
> This looks like a state var filter, but the Q
> regulation is set up so that a sine wave can build
> up.
>
> To prevent heavy saturation and distortion a amplitude
> regulation is needed, like in all low distortion designs
> be it Wien bridge or else.
>
> Unfortunately the gain/phase characteristics change
> a lot if the frequency is changed a lot, so
> the regulator must work fast.
>
> Usually this is done via saturation devices (zeners etc.).
> This will cause some distortion.
> Or by some peak detector and very slow lowpass filter.
>
> If the quadrature outputs are squared und summed up,
> this term will be DC (ideal) and can be taken for
> instanteneous amplitude regulation:
>
> SQR(sin(x)) + SQR(cos(x)) = 1
>
> I know there was a single chip or hybride package for
> this. Has anybody tried to build an audio range quadrature
> VCO using this method?
>
> Cheers
> Martin Czech
>
> CAD/CAE Library Group
> Micronas GmbH
> Hans-Bunte-Str 19
> D 79108 Freiburg
> Tel.: +49-761-517-2422
> FAX.: +49-761-517-2258
> http://www.micronas.com
>

--
 -Jim
------------------------------------------------
* Visit:http://www.silcom.com/~patchell/
*-----------------------------------------------
* Come to the Pacific Southwest DIY Meet
* April 12, 2003 10am-5pm
* Santa Barbara, CA
* http://www.silcom.com/~patchell/synthdiymeet.html
* for details
*-----------------------------------------------