Wien Bridge osc.

Their capacitance. In practice you might use a "gimmick" (twisted wires to make a very small capacitance). In more complex applications (like 400Hz MIL stuff) I've used window comparators to accurately set the amplitude. ...Jim Thompson

I see. Thank you for the explanation.

Piotr

n.

t -

Thanks Phil. I don't know the MC1496, I'll give it a look.

George H.

re

FG694301...

Oh it's not crossover distorion that I'd worry about... the threshold sets where the gains starts to taper off. Different +/- thresholds equals distortion. 60dB is like a mV of harmonic in a volt signal.

George H.

George H.

Quadrature output. That's nice! I made this quadrature osc. by sticking two all-pass filters into each other... with a bit of gain. Component and other variation gave only about 1 degree of 90.

George H.

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

ide quoted text -

It's a Gilbert cell mixer. I was just thinking about making a Wien bridge oscillator out of the current source transistor and adjusting the feedback gain using the cross-connected diff pairs. I haven't tried actually doing it.

Cheers

Phil Hobbs

Nasty to control. Best is to simply control tail current of a diff-pair oscillator. ...Jim Thompson

distortion.

If you have a monolithic dual, probably. But they cost real money, whereas a 1496 costs a quarter.

Cheers

Phil Hobbs

distortion.

temperature

I have one at 500MHz. I'll ask client if I can present :-) ...Jim Thompson

distortion.

Ohm)

temperature

much...

detector

I spent a chunk of today wringing out a low noise amplifier made by cascoding an AlGaAs pHEMT with a SiGe:C NPN. Then my probe slipped and the magic smoke came out. :(

Cheers

Phil Hobbs

How about using an LM13700?

You could do that too, but it's good only to about 1 MHz. I like OTAs in general, despite their horrible noise, but there's a shortage of good ones. VTC used to make a 70-MHzish one, which I still have a few of, and there's the OPA861, which is fast but weirdly nonlinear and not differential. OTOH the MC1496 is good to 100 MHz or so, and costs a quarter.

Cheers

Phil Hobbs

quantity:

Using the SPICE model provided by Panasonic for the FG694301 and the same circuit as above, the FFT of the output swinging nearly rail to rail shows the 2nd harmonic down about 120dB:

Obviously that's pretty optimistic, but maybe worth considering.

quantity:

I have done something stupid and read that FFT wrong! That's the second harmonic that's to the right of the first, not the third. Sorry! Those results were too good to be true. The first harmonic results seem to be comparable to the unmatched diode case.

quantity:

UGH. That should read the first harmonic. First harmonic is to the right of the fundamental, then second. Anyway, the results of the SPICE run are there in the picture, regardless of my bumbling attempts to interpret them. I'm going to bed!

Ummm, fundamental == first harmonic. :^)

Tim

Yeah. Tired/somewhat intoxicated posting isn't good. Sorry!

In any case, now that I'm more rested I think it's clear that the FFT shows the results with the MOSFET circuit with complimentary fets probably won't be as good as matched diodes - the second harmonic just isn't low enough. Would've been cool if it worked.

distortion.

R=3D1k Ohm)

ge and

temperature

odes hot

(the

hat much...

detector

n bridge

en

g the

ied

=A0 =A0 ...Jim Thompson

Sigh.... I hate when I screw up and break something. My day was a bit better. I broke a 'cardinal' rule yesterday and fired up a new pcb on Friday afternoon. (Knowing I'd most likely be going back in today.) So all my goofs on the pcb have been found. Monday will just be tweaking a few values and then making it look nice.

But I rewarded everyone by getting Moo Shu Pork from the Great Wall to bring home for dinner... yum.

George H.

t -

I dont know if this is still relevant or not, but one way to control gain in applications like this is miniature NTCs. They are very small NTC beads, less than 1mm diameter=3D1/25.4 inch. Best are probably bare beads, but I last saw them as a temperature sensors in an old PC motherboard. And they were taped over. I saw them them years ago, manufacturer was something like ESi or OSI. Perhaps Digikey has them.

Hope this helps