I designed a set of filters, similar to the 30m version already in the beacon, based upon the series resonator design in EMRFD (page 6.76). Here's my Spice Model of the (three) filters, configured for 40, 30 and 20m operation...
The predicted frequency responses of the filters are shown in the graph below...
OK - so we've got a set of bandpass filters - how should we select between them?
Obviously mechanical switches are out of the question (the rest of the Beacon being electronically controlled) so I looked for a more subtle approach. I thought I'd found it in Doug DeMaw, w1fb's article in QST, January 1991, page 24...
... in which Doug describes switching between parallel filters using diodes.
I designed a circuit, in which the chosen band is loaded into a (74137) latching demultiplexer, which switches the biasing voltage to the appropriate diode switch - here's the important part of my design...
I even got as far as a PCB...
The system works, but only up to signal levels of ~ 100mV. Above this, the diodes introduce significant non-linearity (which negates the whole purpose of the bandpass filters!!). Unfortunately, there's around 0.3V of output from my present 30m bandpass filter in the beacon, so this circuit isn't suitable - back to the drawing board!
OK - I guess Doug only intended the design as a Rx front end band-changing scheme but I was expecting the switches to work up to a larger fraction of the diode forward voltage drop.
Today, in a recourse to good old brute force "steam hammer to crack a nut" engineering, I've ordered up a bunch of PCB relays. I'll replace those dumb, non-linear diode switches with something a little more clunky!
...-.- de m0xpd
No comments:
Post a Comment