Root cause of RF power amplifier failure?

[bjbdts]

Hi, everyone:
We had a RF power amplifier with 500W CW output @ around 2GHz. The last stage of the amplifier is an isolator that's supposed to handle 1000W power. The output connector is a N type female. We use a semirigid cable RG401 with N Male connector to connect the PA output power to a microwave cavity.
The failure happened after we had a high reflected power due to impedance mismatching. The contacts of both N-female from PA and N male from cable stuck together and fell from the connectors. The connectors got burned (turned black inside) of course. We sent the power amplifier to the manufacture to have the output connector replace, but when the manufacturer opened the power amplifier box they found the output and input of the isolator got burned (the part of the boards that connect to the input and output of the isolator turned black) too. Several Power transistor on one of the boards that connects to the input of the isolator failed. Based on these the manufacturer claimed the power amplifier failure is due to the failure "occured at the solder connection made on the cable side of the N connector". But we think that isolator is supposed to protect the internal circuit.
Sorry it's a long message. Do you have any idea/experience to help locate the root cause of this PA failure?

Thanks,

bj

 

[BrianG]

With isolators, as with everything else, there are design limits. High reflected power can develop very large voltages which can fry isolators, couplers, make waveguides arc and generally kill other microwave components.

Better designed output stages or power sources often have a refelected power / VSWR detector circuit which shuts things down when this sort of fault condition occurs.

 

[bjbdts]

Hi, BrianG:
Thanks for your response. I still have confusion why didn't the 3rd port( should be matched load)of the isolator get burned. And since the temperature of the isolator is above 50 degree C, is it possible that the ferrite material of the isolator got demagnetized so that the reflected power went to input port instead of the matched load port?
Can you advice in more detail about the detector circuit? Since as a customer we cannot change the design of the power amplifier. Is it possible that we can put this circuit outside the PA unit to protect the power amplifier?

Thanks,
bj

 

[BrianG]

Hi Bj,

I would have thought that the Curie point of the ferrite material would be very much higher than 50 deg C.(120-150C?)
so that is unlikely to be the problem. Are you sure the matched load port is actually terminated, i.e. have you checked that the load is not open circuit or wrong value?

Reverse power detectors usually require some sort of directional coupler: forward direction is "straight through" power path (with some insertion loss obviously) and the reverse direction gives a measure of the reflected power into a suitable detector circuit which can drive a the power reduction circuit. This is actually just like your isolator, so a detector could be coupled to the third port, however with an isolator the reverse power also appears here and must be dealt with.

The type I have more experience with (at 9.5GHz) are not magnetic but use tuned slots for the reverse sample port and have a definite built in loss in the reverse direction, e.g. 20 or 30dB, thus cannot actually handle any reverse power.

Depending on the r.f. source this detector circuit could just reduce the supply voltage to maintain power within safe limits, or it may have to actually trip the supply.