1

I am currently working on an RF project and using a high power RF SPDT switch. The control voltages of this switch are between 0V/-40V. If the control voltage is 0V, then the corresponding path is ON and if it's -40V then the path is OFF.

I wish to control this switch using a microcontroller. I know I would need a driver between the microcontroller GPIO pin and the RF SPDT control pin. Assuming I have a -40V rail available in the system, how could one convert the microcontroller GPIO signal (for example, working between 0V to 3.3V) to those required by this switch?

One option that I imagined is to use a level shifter. However, level shifters working in these voltage ranges are not available off the shelf. So I thought of building one using high power MOSFETs. But this solution would need several MOSFETs, thereby costing some PCB area.

Are there any other smart solutions for this problem?

Thanks.

Niteesh Shanbog
  • 711
  • 7
  • 21
Nik
  • 11
  • 2

1 Answers1

2

consider this

schematic

simulate this circuit – Schematic created using CircuitLab

analogsystemsrf
  • 33,703
  • 2
  • 18
  • 46
  • thanks for the schematic. When the Emitter of Q2 is lets say 1.8V, then we have around 2A of current through the diode path. I understand that you put the diodes to clamp the base to emitter voltage of Q1. But how do we prevent such high currents ? Or did I understand this wrong? – Nik Feb 19 '19 at 17:37
  • When left end of R1 is at 1.7v (maybe 10mA thru Q2), the emitter of Q2 will be about 0.7v and we'll see 1.0 volt across R1, thus the 10mA. That 10mA flows out the collector of Q2 (which needs to rated for 40Volts), and down to the top of R2. About 1.5mA flows thru R2, maybe 1mA thru D2 into the base, and the rest (7.5mA) thru D1 into the collector. The 2 diodes provide a Baker Clamp. Schotty diodes are best here. – analogsystemsrf Feb 19 '19 at 18:03
  • Perfect it works as expected. Thanks for the suggestion – Nik Feb 21 '19 at 16:09
  • Realize the output time constant with 100 pF load, and the 10Kohm pullup, will be 1 microsecond. Does your system tolerate that slow a control signal? – analogsystemsrf Feb 24 '19 at 10:16