I deviated from the recommended configuration (some on purpose and some by neglect) and so I'd like your advice.
1. The part is shown powered from ground to +3.3V, I powered it from -3.3V to ground exchanging all pins as expected (RF inputs/outputs are AC-coupled). Power supplies are all properly decoupled.
2. The part asks that the pad under the package be connected to ground (in my case that would be -3.3V).
3. The part is specified from 4GHz to 18GHz.
a.) I expect powering the part with a level-shifted power supply has no real impact as long as decoupling is good. Right?
b.) I neglected to connect the pad under the chip to anything. It is soldered-down to a floating pad. The pad is 10mils above an AC ground plane through Roger's 3003 material. I ohmed-out the pins on a chip and see that the pad is shorted to all GND pins. Question. Would you expect the part to still operated--just have worse RF performance? I may be able to modify my PCB to repair this error, but I'd like to first see it function somewhat.
c.) The low-end 4GHz limit. Is this defined by the slow slew-rate of an anticipated sine-wave input or, if I used a fast-edge limiting amplifier on the inputs, would this operate slower?