RX sensitivity of a differential receiver

Hi, I have a differential input receiver and I want to calculate its sensitivity. So with a spectrum analyzer, peak detector, display units dBuV , I connect to each of the differential inputs. I see the center carrier and two sidebands (modulation signal). At each of the differential inputs I measure the sidebands [dBuV]. Now how do I add all four numbers (two sidebands for each of the differential inputs) to get receiver sensitivity in Vrms ?

Do you know what the sensitivity is ?

Ok, maybe I explained it wrong. It doesn't strictly follow the definition of sensitivity. But, I measure the sidebands the way I explained above at the minimum detectable signal according to our criteria. So how do I add all the numbers to get the result?

If this thing is an AM receiver, its sensitivity would be measured some thing as XμV @40% mod for a Y S+N/N ratio.
If the differential input device is a transformer then working it single ended, means that if it does not use a centre tap then the "other" must be decoupled to earth.
If the input stage is a true differential solid state amp, then the gain in differential mode is just twice of working it singled ended. If its some other value then its not a differential amplifer.
Frank

The receiver can work as AM or BPSK receiver. At each of the differential inputs we have a carrier and two sidebands. I have only a single ended probe so I have to measure each input separately. I have different criteria for sensitivity than S/N . Ok, calling it sensitivity can be confusing, lets say it is just an input voltage.
I am not quite sure how to add the sidebands. If I read from spectrum analyzer XdBμV (peak detector) , and then convert XdBμV to YμV, then YμV*0.707 to get AμVrms , than can I just add all of them Sens=sqrt(AμV^2+BμV^2+CμV^2+DμV^2) to get the result?

You confuse many things..
Sensitivity is relative and referenced measurement and you define a "treshold" for that.
For instance...
You define a treshold for the sensitivity-for example- S/N ratio at the output of the receiver or BER of a digital receiver or EVM... OK ?
After this definition, while decresing the signal amplitude at the input you find a worse treshold value after a certain level of amplitude.At that point, it's the sensitivity of your receiver..

Sensitivity is not measured at the input...

I do not confuse things.. I think calling it sensitivity can confuse people (although it is technically a sensitivity level the way we define it in my company). I know, you would normally feed the signal to the input over a coax cable....but our receiver has no input connector and we cannot tell directly the input signal level.
How to sum the voltage from sidebands and differential input pins ?

If there is no input connector there should be an antenna... shouldn't be ?
All you need is to measure the sensivity by using a reference antenna a distance in a chamber.The calibration must very well be done before measurments with a reference receiver.

There is an antenna but it will not pickup any external signal. It is quite unusual receiver and setup.
I only want to know how to calculate input voltage from the numbers I read from spectrum analyzer.