Re: Dual sine wave generator with variable frequency and 90 degree phase difference

oscillator,

The phase shift oscillator is a bad idea for the OP's application. You have to control 4 time constants. The state-variable option is better because you only have to set two gains.

I wonder howcome nobody's mentioned a "quadrature oscillator" e.g.

Of course, with those low freqs, you'd need CMOS opamps and some BMF caps and resistors. :-)

Cheers! Rich

re an

Because it doesnt meet the OPs requirements.

oscillator,

I think the joke is on me. The differential of a sine wave is a cosine wave right? So how do we measure the "rate of change" of a sine wave? Ken

It's not that bad. Remember, with four sections, you only need 45 degrees phase in a section, so a single RC will do it. The transconductance amplifiers are well matched for R (as a function of the common program current), and thus the only requirement is four matched C values. One can be a tad looser, because it's only really required that C1 + C2 =3D C3 + C4 to get the required phases.

Another poster referred to 'four 90 degree shifts' which WOULD require some large number of matched components; a simple RC filter only approaches 90 degrees, never actually achieves it, so such a design would need a more elaborate filter (more components). Remember, a simple inverting amplifier is good for 180 degrees phase shift. Use that.

Transconductance amps, of course, really shine as voltage-controlled oscillators; there are designs that go from 10 Hz to

oscillator,

Avoid measuring rate of change when you can. Integration gets you from sin() to -cos() and -cos() to -sin()

e.g.http://www.ecelab.com/circuit-quadrature-oscillator.htm

"state variable" == quadrature

It is being discussed in another subthread.

The OP wants 0.05Hz to 10Hz. Over that range you need the ability to adjust more than two sections of the phase shifter if you want to work with a near constant gain.

oscillator,

If I have a voltage sine wave across a Low R, then take a Cap off the R, said Cap conveys the derivative of the voltage across said R, provided said Cap feeds a high resistance impedance "r".

Watch this ascii mess,

O__o___C___p | | R r g r o u nd is down here...

O is the oscilator, "o" measures that. Shouldn't scoping at p find o and p 90 degrees out of phase?

I'm using, current= C*d (Voltage)/dt. Regards Ken PS: I think Moose FETs a BIG meany.

check out the ad633 data sheet for a decent quad VCO

you may have to increase the C's a bit for that low a frequency

Steve Roberts

oscillator,

Circuits that make "rate of change" measurement tend to increase distortion and the odds of oscillation. When you can, you should avoid such circuit.

I don't intend to be :>

Of course. These are TRANSCONDUCTANCE AMPS with current program pins, all cascaded. The four sections all change R_output together. A single pot changes four sections of the RC phase shift all in unison.

That is a lot more parts to do it that way. You need 4 CA3080s for this version and only 2 for the state variable version.

to

Well, not CA3080 (they're obsolete); dual OTAs (NJM13700) are $0.68 each at Mouser. If it takes two, I can afford that.

It is a moral outrage to use more parts than needed :>