High Power combiners

[trash]

I'm looking at ways of combining several amp modules to make a single high power amp. I've been consideing a few different methods each on it's own merits. Normaly at lower powers, it would not be an issue. But at high powers, small mistakes can create big fires
I'm wondering if anybody else has come up against this problem or has suggestions. I've been looking at using 3dB hybrids as combiners, but they are a little scarce on the ground to test to the point of destruction.

 

[DTB Radio]

What amp modules do you want to combine? Are you trying to do what Texas Star does for their bigger amps?

 

[trash]

The power modules themselves really aren't of any concern, they could be mitsubishi modules or transistors etc.
In my case they are 50 watt 1.3GHz SSPA's, so the power is high to start with.
I want to combine them 4,6 or 8 fo an output of up to 400 watts minus losses.

The problem I have with things like wilkinson combiner dividers is that I can easily set them on fire ! They're just not up to the task of high power. But 3dB hybrids are.

I've used hybrids at work on commercial broadcast combiners and very high power satellite systems. But knowing how things are done and how to make them work correctly for myself is another thing.

I'm not familiar with the Texas Star amps. Do you have a picture or circuit/block diagram of one ?

 

[FL Native]

 

[rlnthndr]

I believe he was refering to the 16 transistor units that use a transformer for teh final combiner.

DTB maybe you can answer a question I have had about those. Doesn't combining the amplifiers in that method make the output port 1/2 of the input. I.E. both 8 transistors outputs are 50 ohm and the combined impedance would be 25.

So would they not have to hav the impedance of teh 8 transistor amps higher say 100 ohms or are they using capacitors to offest that(not to sure if that is possible to do or not).

Always been curious.

 

[trash]

Yes, the torriod transformers in that circuit are wilkinson dividers. The kind I would normally build at SHF frequencies, though they don't use ferrite, they're just stripline.

The dividers aren't the problem.
It's the combiner. You'll notice that if you just treat each push pull circuit as an amplifier module it brings you down to the same circuit that I'm using (equivallent). But at SHF.

Now if you look carefully at the output from 'e' and the other from 'd' that go to the input of you combiners you will notice they are slightly different lengths even in the circuit diagram.

At HF/VHF, a couple of centimetres, the phase difference is not an issue. But at SHF, millimetres can mean much bigger phase differences. Again, this might not be an issue at low power, but at high power, the potential across those 100 ohms resistors and you have .... FIRE !

Even when you're careful to keep the modules phased, there is always some degree of error. Imagine 200 watts across that 100 ohm resistor. The resistor has to be small at SHF else it's inductive component is too great.

3dB Hybrids are the component for the job, I just don't understand how to build them correctly or the advanced theories behind their applications.