RF pin diode switch design & parts

[benedettis]

Hello, i am in need of a SPDT type RF pin diode switch. I am implementing this into a 900mhz TDD amplifier, running 10-20 watts of RF power. What i need to know is what kind of RF pin diodes do i need to use that will handle 10-20 watts of RF at 900mhz? There will be 2 switches implemented, one on the RF input, and one on the RF output. And what kind of pin switch should i use. I have seen several. There will be a detection circuit implemented so i can do the switching of the RF switches.

 

[tony_lth]

The SPDT switch VSW2-33-10W+ from Mini-Circuits can match your requirements.

 

[enjunear]

Hello, i am in need of a SPDT type RF pin diode switch. I am implementing this into a 900mhz TDD amplifier, running 10-20 watts of RF power. What i need to know is what kind of RF pin diodes do i need to use that will handle 10-20 watts of RF at 900mhz? There will be 2 switches implemented, one on the RF input, and one on the RF output. And what kind of pin switch should i use. I have seen several. There will be a detection circuit implemented so i can do the switching of the RF switches.

Use tony's suggestion and find a pre-packaged device. Otherwise, you can spend a lot of time designing your own switch. If you want to make your own, start by reading this: Microsemi - THE PIN DIODE CIRCUIT DESIGNERS' HANDBOOK. See chapter two, around page 20. This whole document is great, so read the intro sections to really understand how PINs work.

Most of the PIN-based switches I've designed were series-shunt, requiring positive and large negative bias voltages (2x peak voltage of the RF waveform at P,max... keeps the diode from being forward biased during worst-case VSWR conditions). The biggest trick is to keep insertion loss low while simultaneously keeping isolation high... they are basically inversely related... one goes up, the other goes down, so look at your specs and choose an architecture that will suit your needs. Low R(on) is highly recommended, so the diode doesn't soak up power that should be going to your load. Those are the high points to remember. There lots of articles and papers written about PIN diode switches if you search the web.

 

[biff44]

are you planning on switching the switch position while it is seeing 20 watts?

 

[jaehlee99]

Hello enjunear,
I found you the thread of RF pin diode switch design & parts.

Would you comment my topic as follows?

When defining power handling capability of Tx/Rx switch module, what is the specification?
For example, peak power= 100W, Pulse width = 20uS, Pulse Repetition Rate =5kHz, mutiply pulse width by pulse repetition rate to get duty cycle. Duty will be 10%.

So,when I describe the power handling capability, peak power 100W or average 10W, which is right?

 

[enjunear]

Hello enjunear,
I found you the thread of RF pin diode switch design & parts.

Would you comment my topic as follows?

When defining power handling capability of Tx/Rx switch module, what is the specification?
For example, peak power= 100W, Pulse width = 20uS, Pulse Repetition Rate =5kHz, mutiply pulse width by pulse repetition rate to get duty cycle. Duty will be 10%.

So,when I describe the power handling capability, peak power 100W or average 10W, which is right?

You'll need both pieces of information for the design. In terms of power dissipation, heating is a generally thought of as a cumulative, long-term effect, so average power is useful for that part of the design (switch insertion loss * incident power = power dissipated). When analyzing the voltage ratings of the parts, you need to know what the peak RF level will be, so you will need to break that down to the peak power/voltage across the parts. Don't forget to include VSWR considerations in your T/R switch analysis. In the worst case, an infinite VSWR, the peak voltage in the circuit could be 2x the nominal voltage. In the case of 100W peak in a 50 ohm system, that'd be 100V,pk (nominal). If you ran that into an open circuit, you'd see 200V peaks in your circuit, due to all the energy being reflected back to the source.