Frequency down-conversion with passive circuitry?

On a sunny day (Tue, 11 Jan 2011 22:31:23 +0100) it happened Jeroen wrote in :

Na, perhaps *tunnel* diode oscillator tuned to f/2, and sort of hard locked?

Sure--parametric amplifiers. You can also make parametric dividers using varactors.

Cheers

Phil Hobbs

Of course you could also use the diodes as a bridge rectifier, powering a tunnel diode oscillator at any frequency you like. ;)

Cheers

Phil Hobbs

...or use square-law inductors (mag amps) and build a bi-stable...

With ideal components, I think not.

Tricks might use saturating magnetics or diodes with negative-resistance breakdown behavior.

John

I wish I understood parametric drive better. The only reference I know is "Vibrations vol. I" by Pippard. I'll have to read that again.

My only experience is when someone suggested I take our optical pumping apparatus and move the RF magnetic field modulation axis. It's typically perpendicular to the DC magnetic field and light axis, and make them all colinear. The signal were similar, but the power dependence was much different. I've never figured it out. (I've always assumed this was a parametric drive... the RF modulates the DC field, which causes a small change in the resonant freq.)

George H.

Use a diode and a large enough capacitor - DC is as low a frequency as you can get!

Rectumfrier driving tank tuned to f/2 ?? ...Jim Thompson

A very special rectumfrier... what a brain fart :-(

To halve (for a boom box I did years ago) I used a "smart" phase-reversing switch, followed by active filters. ...Jim Thompson

Yes, it can generate subharmonics making use of the nonlinear capacitance in a diode. A quick google search came up with D. Leenov and A. Uhlir, "Generation of Harmonics and Subharmonics at Microwave Frequencies with P-N Junction Diodes" Proceedings of the IRE, October 1959.

Pere

As linear components it is mathematically impossible to do so. Linearity dictates that no new frequency components are created. Of course any component is only linear in a certain range so it is not at all impossible to do so with the components you have suggested. One just has to find the right way to use them and take advantage of their non-linearity. The result may not be all that great though...

Diodes aren't linear--they're nonlinear both in conductance and in susceptance. A parametric amplifier produces signal and idler frequencies that sum to its pump frequency.

Cheers

Phil Hobbs

That would be a subharmonic parametric oscillator. You could probably do that with an LC tank, a diode, and some DC bias. That wouldn't be hard to Spice.

The same thing should work with real-world crap-dielectric ceramic caps.

John

Is a diode considered a passive device?

Now _that_ would be a fun project--a kilohertz parametric amp with just Ls and Cs.

Paramps are pretty quiet, too...I wonder if a Y5V paramp would make a good lock-in front end, or something like that.

Cheers

Phil Hobbs

Yes, but you need some nonlinear passive component.

Saturable magnetics work. Classic telephony gear used these extensively, since they don't wear out. The classic Western Electric ring generator (48VDC in, 88V 20Hz out) is an example.

John Nagle

With a high-power, high-frequency pump, and some output diodes, you could build an audio power amp with caps as the gain elements.

I have a paper around here somewhere on a kilovolt NLTL pulse sharpener based on nonlinear ceramic caps.

Probably so, but it would be slow. For the same capacitive loading, you could parallel a dozen BF862s.

How about using a photodiode for both its PD function and its nonlinear capacitance? Fun and likely useless, like the parametric audio amp.

I think you can get a *lot* of gain out of PIN diodes, too, used as RF switches. Heck, ideal diodes can do that.

John

Of course, it is non-linear though.

Audiophool fame is just around the corner. I wonder if it would sound like a tube amp?

Easier than saturating ferrites. I wonder if it's as fast, though.

That would need about 250 mA of supply current. (Hmm, I wonder if you can use AC-coupled stage stacking to get lower noise....) The paramp would have to be tuned, so there wouldn't be any capacitance at the centre frequency. That would be medium ridiculous, but fun to try out.

You can use the junction capacitance to tweak the tuning of a parallel LC, which is a cute method for doing narrowband measurements.

Running a PD as a paramp might be kind of fun too...I can't think of a practical use for it, though. The series resistance goes up pretty steeply when the diode isn't fully depleted, so you'd lose the high frequency response as soon as you got the varicap action going. You need a really high pump frequency to get any gain from a paramp, and it's usually the PD that's the slowpoke anyway.

Doing that with a laser diode might be an interesting way to make an optical frequency shifter, but it would be hard to keep it from injection locking.

You need the charge storage to get gain from a PIN switch, I think. Ideal diodes require DC bias equal to the peak RF current, so your gain could never be more than -3 dB.

Cheers

Phil Hobbs