Don't forget that in the lab people measure with 50 ohm load.
In my experience a good solution is to design a buffer that doesn't degrade the IC performance (you do not care about current consumption as it is only for test purpose), mainly the linearity.
Design the buffer in order to be power matched and differential (unless your output frequency will be very low). On the board it will be easy to put a balun.
I hope it can help.
Max
Well I am planning to use a source follower for each output of the mixer and then outside I will have another amplifier which will convert differential to single ended and match to 50 ohms. So as far as the source follower is concerned it will only see a high impedance node and capacitative loading. Does this look plausible?
Yes
it's a good choice, only if the external amp has higher linearity respect the circuit you are measuring and do not add noise.
This is a common used approach to measure signals from few kHz to few MHz (it's difficult to find a balun in this range).
Used amps are max4444 and AD8130
I hope it can help.
Mazz
1) The buffer can be a simple differential pair with large bias current driving low value resistors. But you have to know the performance of the buffer in order to derive the actual performance of LNA.
Hi, dsjomo, can you post the draft illustration you mention on web **broken link removed**