Isolated RF switching

A simple switching diode like a 1N4148 should be suitable. Drive it through isolating resistors of maybe 1k each from a +/- 30 to 50V source so that the diodes are passing around 20mA when conducting and are reasonably reverse biased when off. It would still work with zero volts in the off condition, but reverse bias will reduce the capacitance. You will also need a couple of series isolating capacitors. John

That's a good easy solution. Chokes in series with the resistors should give even more RF isolation and allow lower resistor values so that it could be driven from a lower voltage (it may have to be field portable).

A reed relay could be even better. They take about a millisecond to open or close, though the actual makes and breaks happen a lot faster.

A mercury wetted reed relay doesn't have any contact bounce, which could be an advantage.

A slightly bizarre way of managing the magnetic field to open and close the relay would be to put two permanent magnets on wheels on either side of reed,and rotate the two wheels in opposite directions at 200Hz.

Why not use a relay?

The Fujitsu FTR parts are excellent 3 GHz switches.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

The RF world solution is roughly the 1N4148 circuit implemented with a PIN diode:

.

Joe

The high on/off ratio of a realy isn't needed and the 200 c/s switching will wear it out. Also, I need to synchronously detect the signal and any delay in switching will upset the null point.

I don't need a particularly high on/off ratio and running a relay at 200 c/s for long periods isn't a good idea. Further down the line the signal will be synchronously detected, so any delay in switching will upset the null, which is critical for the system to work properly.

Bias voltage should be greater than anything (ac peak) on the line.

RL

If you're planning on using diode switching, I'd suggest doing it at the receiver end of the feedline. That'll make it easier to protect from surges, and let you do a better job of filtering the reverse bias voltage--otherwise, any duty cycle asymmetry in your lock-in will let low frequency junk into your front end.

Cheers

Phil Hobbs

Would it have to run long-term?

A little relay will switch in about a millisecond.

If you don't mind the complexity, use a logic-level isolator and a diode. Just be careful about the EMI issues.

A PV optoisolator driving a phemt would be fun. The SAV541 goes to about 2 ohms with +0.7 on the gate.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Are you making a switchable reflector?

People make isolated mosfet gate drivers with built-in dc/dc conversion, but they will have oscillators inside so may be emi hazards.

You could use my dual optoisolator totem-pole thing with a battery, driving a fet or phemt.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

It definitely will be, this is for receiving faint RF signals.

Yes, a bit like the Lorenz blind-landing system but on receive, not transmit.

The dipole is the reflector of a Yagi-Uda receiving array, so normally there would be no connection to it at all.

So you're trying to find the exact null to get rid of a strong signal coming from the reverse direction? Or what? Inquiring minds want to know. ;)

Cheers

Phil Hobbs

I want to have two reflectors, one each side of the usual position, and short them alternately so the lobe of peak sensitivity shifts from side to side. By synchronously detecting the carrier level, it should be possible to make a direction-finding system based on the point of maximum sensitivity. This may give better results on weak signals than the usual system which is based on a null.

Basically a Lorenz blind-landing system in reverse.

I guess you could have a feedline of the proper number of wavelengths, and apply the switched short at the end. That puts all the electronics in one place, out of the rain.

I suppose I should Spice that.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Why not switch between two receive antennas?

There are lots of cool RF switch chips. We use one that switches cleanly in 5 ns.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Sure thing. Bright-field interference works a lot better than dark field in optics too. The trouble with surfing a null is that the SNR is zero there, making the tracking sloppy.

Phase-sensitive detection fixes that in electronics, but that’s harder to do without having a phase reference from the transmitter.

Cheers

Phil Hobbs