Protect sensitive RF circuits from burn out by sequentially switching from transmit to receive.
By Zack Lau, W1VT
What, another T/R sequencer? Yes, I decided to make it as simple as possible, while adding +12-V outputs and reverse polarity protection. I think the reverse polarity protection is pretty nifty— who hasn't worried about hooking up batteries backwards and frying something? The 12-V outputs make it real easy to wire up my transverter designs, since they are made up of modules that run off 12 V. Hopefully, this design will spur more designers into adding reverse polarity protection.
The primary reason for using a sequencer is to protect the RF relay and the amplifiers hooked up to it. RF
225 Main Street Newington, CT 06111 email [email protected]
relays can be damaged by hot switching at high power levels. Unfortunately, what constitutes high power is rather fuzzy—I have not found any good references that adequately address this topic. Based on my experience with microwave transverters, I always put in a sequencer when switching more than half a watt. I don't bother with them in inexpensive systems running less than 100 mW.
Like the T/R sequencer designed by Chip Angle, N6CA, this sequencer uses a quad comparator to monitor the voltage on a charging or discharging capacitor.1 I looked at using an integrator to get more uniform delay intervals, but decided to stick with the simpler circuit. An integrator can gen-
1Angle, Chip, N6CA, "TR Time-Delay Generator," The ARRL Handbook for Radio Amateurs, 1997, pp 22.53-22.56.
erate a nice triangular waveform, as opposed to the exponential curve generated by an R-C network. The 10 Li resistors are used to provide hysteresis, so the outputs don't switch back and forth unnecessarily near the transition point. They are supposed to provide a little positive feedback, instead of the more common negative feedback used in other applications.
This sequencer first turns off the receiver, then activates the relay, amplifiers and finally the transmit IF outputs when switching from receive to transmit. When going back to receive, it turns off the transmit IF, amplifiers, and relay before reactivating the receiver. The idea is to introduce enough delay between these states to allow everything to settle down, and to reverse the order when switching from transmit to receive. This is especially important with a mechanical relay which may make intermittent contacts for a few ms when switched.
A big advantage to using solid-state amplifiers is that you can switch them off during receive. Not only does this reduce the possibility of hot switching, it reduces the chance of amplified broadband noise getting into the receiver. Using a PIN diode switch to cut off transmit drive also helps to prevent hot switching—I use the transmit IF signal to control this switch.
Since I've never needed to change the polarity of the output of one of my sequencers, I decided some simplification was in order. Instead of the XOR gate Chip used, I decided to use a hardwired switch based upon the principles of a transfer switch. To switch polarity with this design, you change a pair of resistors from horizontal to vertical, or vice versa. By labeling the blank area of the board between the resistors, identifying the resistors ought to be straightforward. The reduction in parts count ought to enhance reliability.
I've also taken advantage of the improvements in switching transistor technology. The International Rectifier P-channel IRF 9Z34 will easily switch 2 A, enough to power a 5-W GaAs FET power amplifier. Similarly, the Zetex ZTX 789 in a little TO-92 style case will actually switch a small SMA relay that draws a few hundred mA. The bonus to using more expensive PMOS/PNP parts is that the switched supplies are reverse polarity protected. I just needed to protect the comparators with a diode and use bipolar electrolytic capacitors which don't care about voltage polarity.
Cheap 2N3906s are used for the RX and TX IF supplies, since they typically draw under 100 mA. You could use ZTX 789s instead of the 2N3906s, with the appropriate bias resistors, for higher current.
You may want to replace Q2 with a VN10LP N-channel FET. This will allow you to hook up the PTT line without pulling the voltage down significantly. Another advantage is that you can now hook up the PTT input to a stiff voltage source without frying Q2. But these FETs aren't as easy to get as common NPN switching transistors.
Of course, it makes sense to watch out for even newer FETs, which will undoubtedly offer better performance at lower cost. There is even a trend toward lower gate thresholds, which allow better performance at lower voltages. For instance, the IRF 7104
Fig 1—Schematic diagram of the transmit-receive sequencer. 16 QEX
drops only 154 mV when sourcing 0.57 A (V = -5V). But, despite the marketing hype, you can do even better than "full enhancement." It only drops 117 mV with a Vgs of-10V.
I made the pads big enough to accommodate swaged terminals, which are a really nice way of making dc connections if you have the tooling to rivet them to a fiberglass circuit board. The board is a bit crowded—I wanted the board to fit nicely on the wall of a chassis box only two inches high. A mirror image of the etching pattern is provided—it simplifies the toner transfer process that some people use to make circuit boards. Similarly, a parts placement diagram using part values and component designations is also provided in Fig 4.
To test the board, I made a little fixture out of LEDs and dropping resistors. A separate fixture makes it easy to line up the LEDs in the proper sequence. It may be useful to slow down the sequencing by bridging the l-^iF timing capacitor with a 10-pF capacitor. This makes it easier to see the LEDs turn on and off.
Fig 2—Etching pattern for the transmit-receive sequencer circuit board.
Was this article helpful?