Confusing measurment results from Non-RF Eng.

We have a Narda 2 to 1 coupler/combiner (2372A) that we are using as a combiner. If we apply a single transmitter at a single input, 5 watts @ 2230 MHz, at either input we get 2.5 watts out (as expected by us). However, if we put two 5 watt transmitters (one at each input, 2230 and 2260 Mhz) we get 7.5 watts out. More than the mathematical sum of the two individual outputs. Varying the input frequencies seems by 10-40 Mhz seems to have no effect.

One or two folks a guessing that it has something to do with the phase relationship of the two signals, but since changing frequencies has no effect, we are left scratching our heads to understand why we got these results.

I'm sure the answer is in the RF 101 book for someone, but we aren't RF engineers by trade.

If someone could educate us, it would be most appreciated. A link to a reference (for the non-believers in our group) would be great.

Thanks in advance.



I should add that the power measurements were obtained with a Bird Wattmeter.

hi 7.5W is not more but you have gotten less. you got to figure out why you getting less. power combiner will give you 3dB gain and power splitter will give you 3dB loss. but combining arms (1 & 2) and spliting arm (3) are different and you have to be of this when you inject the input signal.

5watt --> 36.9dBm
2.5watt --> 33.9dBm
7.5watt --> 38.7dBm

when you split the signal, you got perfect 3dB loss so you using correct port to inject the signals but when you combiner, you have only gotten 1.8dB which is 1.2dB lesser then what is suppose to be. are you using correct ports to inject two input signal?

[QUOTE]hi 7.5W is not more but you have gotten less. you got to figure out why you getting less. power combiner will give you 3dB gain QUOTE]
No. The combiner is lossless only if both signals are coherent (in phase and magnitude). With two 5W signals of different frequency, each signal will be attenuated by 3 dB at the combiner output, total power is 5W, theoretically.

What kind of detector are you using? I guess, it's not a true power detector.

Power splitters/combiners typically introduce some minor loss. At 2.5 or 5 GHz, I would expect 0.5 dB.
These devices are matched to 50 Ohms, and require the source(s) and load(s) to be matched to the same impedance. To find the sensitive point, use adjustable two-pad matching coax device (tuner) to vary the impedance. You will see that one or both your signal sources "transmitters" are a bit mismatched, and the tuner can help to achieve a higher power than specified.
Maybe also your Bird power sensor is a bit mismatched and using the tuner you can improve its power reading.

some interesting input. The insertion loss is spec'd as 0.45 dB. That would easily explain the 10 w (5x2) in 7.5 w output. It's at least in the ball park, when combined with other factors.

But where I'm really scratching my head is when only a single 5w (as measured at the input to the coupler) input is turned on (and the other input is terminated), resulting in a 2.5 w output -3dB.

I thought, that perhaps the missing 2.5 w was going out the terminated port, but after measuring it, on 0.5 watts appears at the unused port. Hmm.

No, it's absorbed inside the power divider by design, see https://en.wikipedia.org/wiki/Wilkinson_power_divider
0.5 W at the isolated port is however considerably above specified 2372A isolation of 18 dB, suggests some mismatch at the other port.

You didn't yet answer the question about power meter type.

Sorry, saying that we used a Bird wattmeter is like saying I drive a Ford. It's a Bird #43 thru-line. Don't know what kind of a detector it uses.

Bird model 43, thanks. I see you have mentioned it. As far as I understand, it's a forward/reflected power meter with diode detector. So it's most likely not measuring the combined power of two transmitters with 30 MHz carrier frequency offset correctly.