How to protect RF amplifier against open/short?
I have read few threads here on the subject but still didn't get the answer that i needed, so starting my own.
Please check out my thinking process and then answer or suggest the solution.
I am designing 20-1000MHz 30W SSPA which should be able to withstand open/short conditions on its output. I thought of protecting the PA using directional coupler and detecting the REV power and switching off the PA, there are two problems that I can see with this approach
1- Detection and switching time will be slow and might not be able to protect the PA
2- There is no guarantee that a open/short at output port will create a maxima at the coupler for it to couple it across
So thirdly I am thinking to use power FET that can deliver 4 times more power than what I require i.e. 120W and use it at 30W so when it sees open/short it will not burn out i.e. voltage maxima will be twice which translated to 4 times the power i.e. 120W.
These are my thought but still cant decide the exact method.
Many thanks in advance for you help.
A sniffer for reflected power or output voltage should be fast enough to disable the amplifier or at least limit it's output power but it depends to some degree on the output envelope shape. In my experience, overheating the PA devices is what kills them so the short reaction time of the electronics is of little concern.
Is it really for 20MHz to 1GHz? You might have trouble fabricating a directional coupler over that range. Maybe a combination of power sensing and DC supply current sensing would work.
Brian.
Ferrite Circulators are used to protect the Power Amplifiers.
Select an output device that has a breakdown voltage as high as possible. It's not overheating that usually kills the device, but instead is the breakdown voltage being exceeded during high VSWR situations. Your idea of using a much higher power device has merit. I would start at 2X, build and test a prototype. If it failed, I would then move to 3X, then 4X. As you know, each increase in "X" dramatically increases the cost of your product. A 5+ octave HPA is quite a challenge, especially starting at 20MHz.
Thank you all, I think I will go with a higher power device and as SLK001 said try with in steps. I have designed the PA from 20 to 1000MHz before but protection was not a concern but for this new product it is.
I could come up with a "rough" approach. Make the output power twice what you need and mount a power 3dB attenuator. At least it will reduce the worst case VSWR.
Not very smart though....
The difference in price between a 30W RF transistor and a 120W RF transistor could be more than S100.
However, nobody can guaranties for sure that the 120W transistor it will not die when short or open happen, even if is not transmitting maximum output power.
MOSFETs are biased in voltage and a VSWR protection circuit placed at the output it will work fast enough to shut down the transistor.
I am attaching a schematic example of a commercial VHF/UHF power amplifier that use simple VSWR detectors at the output which shuts down the gate bias of the PAs during open/short at the output.
We use ferrite circulators in RF as isolators as @Bigboss mentioned. The condition of open or short can be caused by antenna as well as other stuff. 30W output power makes about 45-46dbm.
With that much power going around, if you burn out something it will probably be at start-up due to heating, I would recommend a heat sink connected at all times.
At 20 MHz?
Why not ? There are many narrow-band circulators in the market.
The application is wideband, but what I meant: I'm only aware of ferrite circulators & isolators for UHF and above.
Yeah, I saw that but I think this amplifier cannot handle all X Watts over 20-1000MHz Frequency Band. ( at least it doesn't make sense )
I mean the PA can handle the carriers over a moderately broadband such as 470-790 MHz or 174-230 MHz or 127-134MHz etc..
Therefore I answered in general manner.
Really though I thought it was 200MHz, thats 1.5 decade BW. I'm not even sure that much power can be given in that band.
I guess this is for EMI testing where such amplifier bandwidth isn't unusual
https://www.arworld.us/html/find-it-...amplifiers.asp
I have designed 200W in this band but without any protection.