Mixer as freuency doubler

On a sunny day (Wed, 11 Jul 2012 07:56:38 -0700 (PDT)) it happened George Herold wrote in :

Double phase rectifiers (transformer with tap and 2 diodes) make great frequency doublers.

Have you taken into account that usually the LO port needs a higher signal level than the signal port.

A diode ring mixer is intended to be a LO-controlled signal polarity switcher, which will work as a full-wave rectifier in the frequency doubling case.

For cleaner doubling, you'll need an analog four-quadrant multiplier, e.g. a Gilbert cell.

It may also be skewed due to same frequency and phase on both ports produces a DC term. ...Jim Thompson

Am 11.07.2012 16:56, schrieb George Herold:

That's feedthrough of the input frequeny. I happened to do the same thing last week and it looked the same. (TUF-1) It's probably inbalance. I wanted to get better doubling efficiency than -13 dB, but didn't succeed. It works somewhat better if I hit the Mixer with tot. 13 dBm, Funny enough, the feedthrough gets better when I maximize the input power.

You'll have to filter the fundamental away. By the way I just saw that the asymmetry is MUCH worse when the scope input is on 1 Meg instead of 50R.

less than ideal setup:

scope screen:

spectrum output of ring mixer if hit with +10 dBm@50 MHz:

hope that works. I have not used bilder-space b4.

regards, Gerhard

Mini-Circuits will sell you a proper frequency doubler for about the same price.

What everyone else said (particularly John's comment about Mini-Circuits selling you a real frequency doubler), plus the fact that you'll do best if the LO and RF ports are 90 degrees out of phase -- otherwise you're just doing synchronous rectification of the signal, which does have bumps at 2*f, but not as good as with the phase shift.

One transformer, two matched diodes, just the right drive level -- ahh, perfect doubling.

Am 11.07.2012 22:34, schrieb Tim Wescott:

Completely correct. I use the ring mixer only because it is already in the system and adding another part type would add ??? for radiation and vacuum tests. :-(

John, today I saw these laser drivers made by AD:

ADN2525, 2526, 2531.

They are CML-like, switching 80 mA in 25 ps with +-1V compliance around VCC. Should make nice pulses. I'll try that.

regards, Gerhard

Thanks Tauno... I've got some analog multipliers, nice chips, but no high frequencies.

George H.

OK Gerhard, Thanks. I thought it might also be a non-symmetery between the channels. When I use the mixer in a more traditional mode, I get something that look like ringing artifacts on each channel... I'll take some screen shots. Hmm a quick look on the 'scope and I don't see it now..

Thanks again,

George H.

ler.

Really? That's neat... so the velocity in coax is ~1/3 of c? wavlength at ~100MHz is only a meter... so 1/4 meter path length difference? Cool thanks, I've gotta stitch some coax together and try it!

OK Jan, John L. and Tim W. all agree one transformer...Say I could just break open the mini-circuits mixer and take the parts from there. (maybe)

George H.

ttdesign.com

Oh this is just a quick 'one of'. I want to extend the range of a signal generator by a bit. It goes to ~150MHz and I'd like something nearer to 200MHz.

George H.

ed text -

Old service shop grade signal generators produced harmonics, and often had a scale for the second harmonic of the highest band. Another trick is to use a fixed oscillator to mix against, to get another range of frequencies.

AFAIK the slowest you'll get in ordinary coax is 70% or maybe 65% velocity factor. If you know the type of coax you can check the data sheet. If not -- experiment. 1/4 wavelength should invert the end impedance -- an open will look like a short, a short will look like an open.

You can _try_, but Mini-Circuits really, really puts the "Mini" in their product.

Depending on how broad-band you need to be, a resonant LC may very well provide what you want.

A single schottky diode, series or shunt, can generate harmonics.

The Boff Diode, since renamed Step Recovery Diode, was discovered by accident during test of a single-diode multiplier.

More like 2/3 of c, assuming polyethylene dielectric.

What's interesting is to make a transmission line out of something with high e, like in the thousands. Coaxial ceramic resonators, for instance.

to

th

e quoted text -

Hey Tim this 90 degree trick works great. (I guess you knew that :^)

Here=92s the doubler with four feet of extra coax in one arm. (That works out to something like 2.4E8 m/s (at 50MHz.)

and the FFT

I tweaked the frequency for a minimum at the fundamental (46 MHz),

And then 46 MHz with more poop.

Thanks again for the tip. Any idea why this works? It=92s fairly obvious why the DC term goes away, but not why the fundamental becomes smaller.

George H.

Foam dielectric?

quoted text -

By -3 dB? The signals are no longer in phase and don't add.