"Rat-Race" and Wilkinson combiners question.

In previous discussions in this forum on Wilkinson splitter/combiners, a key point made clear was that the Wilkinson, when used as a combiner, has to have both inputs driven in phase.

It could combine signals that had derived from a single source, and previously split (and maybe then amplified), but it would not be suitable for combining two independent signals. As a splitter of course, it will function well.

The "Rat Race Ring", like Wilkinson, is also made up of quarter and three-quarter wave sections of transmission line, and also has features like the line impedances being (Port_Impedance)*SQRT(2).

My question is, when used as a combiner, is the Rat-Race-Ring subject to the same limitations as the Wilkinson?
Do the inputs have to also be in phase to get a rat-race to work, or can it be used to combine independent signals?

Hi Darktrax,

Yes. The two signals must be in phase in order to add into the in-phase port, or in counterphase in order to add into the other port.
If the two voltages are V1 and V2, the outputs at the two ports are (V1+V2)/sqrt(2) and (V1-V2)/sqrt(2) (even- and odd-mode respectively), when "+" and "-" are vector operators.
Regards

Z

zorro said:

Yes. The two signals must be in phase in order to add into the in-phase port, or in counterphase in order to add into the other port.
Z

Click to expand...

Hi zorro
I have been thinking about this. What you say is certainly true of the Wilkinson combiner, but I was not so sure it is also true for the Rat-Race.

I think it through this way..
In a Wilkinson combiner, the two inputs have to be in phase for the combiner to function. If you take one input away, replacing it with a termination, there is no cancellation, no port isolation, and what comes out the common output is more than 3dB down. The whole thing stops working! Different when used as a splitter. Then there is no problem.

In a Rat-Race combiner, a single input splits, takes 2 routes, and meets itself to either cancel, or add, at the various ports.
A second independent input does all this again for itself. By superposition, the internal voltages may be a complicated addition of both signals, but I think they should go through the combiner separately, in much the same way one can send signals through a coaxial cable simultaneously from each end in both directions.

I have yet to properly figure out how to use a Rat-Race as a combiner. Need to do the work! I could be just so wrong.
**broken link removed**
**broken link removed**
My regards ..