PIN T/R Switch


The final board in my transmitter is the T/R switch. Since my power level is below 100 W, it’s feasible to use ordinary power diodes as RF switching devices. Wes Hayward, W7ZOI, wrote a nice article on this in the May, 1995 issue of QEX, “Electronic Antenna Switching.” If you have copy of his excellent book, Experimental Methods in RF Design, it’s included on the CDROM. I simply used the final design in that article, with some minor modifications.

The one thing I added, as suggested in the article, is a shunt switch at the receiver port. This made a big reduction in leakage from transmitter to receiver at frequencies above about 10 MHz.  I included an onboard power supply made from a couple of small 7.5 VAC transformers I had in my junkbox.

As usual, I fabricated my filter board on double-sided copper clad with a full ground plane using the Sharpie method.

A Solid-State T/R Switch Using PIN Diodes

The 9 V bias supply has to be well filtered. In the article, Wes used regulated 12 VDC from the transmitter supply. I may switch to that, because hum and noise on my 9 V supply does leak into the receive signal.  Using a 10,000 uF filter cap makes it negligible compared to band noise, but still above the receiver’s noise floor.

Here is the transmit to receive isolation with and without the final shunt PIN switch on the receiver port. For reference, an isolation of 50 dB at 7 MHz is equivalent to 0.75 pF. That means the “off” capacitance of the diodes is quite low, and the stray capacitance is also very low.


Passband losses

From antenna port to receiver port            From transmitter port to antenna port                                            antenna port

Losses are acceptably low, especially for the HF bands. At 3.5 MHz, most of the loss is from the 22 nF coupling capacitors that I used. Paralleling a couple would have improved that.

Switching time is in the sub-microsecond domain when one of the transistors is turning on, which is the direction that brings a pair of PIN diodes into conduction. When the transistor turns off, the 100k resistor must charge one or more 22 nF capacitors, with a time constant of 2.2 ms. However, the diodes are effectively off almost immediately since the forward bias disappears long before the reverse bias has built up. The only reason for high reverse bias is to reduce leakage capacitance to the lowest possible level. So this system also turns off in a few microseconds.

The above data was taken with the T/R switch on the bench under small-signal conditions. When used with the completed transmitter and with the power amplifier running full power (about +48 dBm), power at the receiver port was as higher than predicted. For instance, at 14 MHz, the receiver sees -21 dBm, or an isolation of 69 dB. Other stray coupling methods were at work. Still, that’s only 10 microwatts of leakage, which any receiver can handle gracefully.