315MHz Oscillator

And all that nuisance bias? If you must, use 3904's as diff pairs.

Yay! Yay! Let's hear it for boutique chips... keeps Q45's in the garage ;-)

Seriously... it will be inexpensive.

...Jim Thompson

Hello Jim,

But that would add 1.5 cents to the BOM budget. Per stage :-(

That would be nice. I could see some apps in I/Q detection.

You can - it is a fairly expensive route, but the products are available pretty much off the shelf.

Douglas Dwyer has posted here on the subject, but he seems to have stopped posting on the 18th May 2005.

Something vaguely similar that I have tried was to use a crystal oscillator module near 100MHz and then triple it by making a square wave and filtering out everything except the third harmonic. To make the square wave, I passed the oscillator signal through some CMOS logic inverters (one inverter as an input stage driving several in parallel for the output stage). I found that the fastest 3.3V logic variety is quite a bit faster than old AC series and works much better. Then I filtered the output to get a fairly clean signal around 300 MHz. You can simulate the LC filter in SPICE and if you use surface mount inductors and allow for some parasitic capacitance in the SPICE simulation, then the real circuit actually works pretty much like the simulation. I think I put a zero at the fundamental and another one at the second harmonic and another one at the fifth harmonic. It ended up being very small and cheap, apart from the 100MHz crystal oscillator module which was not cheap and not particularly small.

Chris

Dunno if you can still get them, but the Motorola MC13176 is perfect for this. It has an internal 32x multiplier, so a 9.84375 MHz xtal will give you 315. The power out is a little higher at about 4 dBm, but that's easy to attenuate.

Barry Lennox