I'm designing an AFE for 100 Hz to 5 MHz bandwidth input signal. ADC is 18-bit 15 MSPS SAR with differential input 8.192 Vpp. I have completed my signal amplifying & conditioning circuitry, the only thing that now concerns me is the ADC driver. I need very low noise, very fast setlling, high slew rate, high bandwidth single-to-differential amplifier and filtered down to ADA4922-1 and AD8139. My input to driver is 4Vpp single-ended signal, from PGA and would require very high input impedance of driver to maintain good PGA performance. Driver has unity gain translation to differential signal. The problem with ADA4922-1 is that I have no VCM input, so my single-ended input signal needs to have offset for the 2.048V to have differential 8.192 Vpp output (0 - 4.098V each output). I cannot do this offset easily because my input has programmable gains and input signal voltage amplitude varies in different configurations. AD8139 has a VCM pin, but don't like its quiscent current - 25mA! For +/- 5V power supply, this is really a lot of heat dissipation. Also, it doesn't have high input impedance like AD4922-1.
Some other concerns are: AD4922-1 is stated as rail-to-rail output amplifier on AD pages, but in datasheet it's not mentioned at all, and in output voltage section for +/-5V supply provides information of typical +/-3.6 Vpp output. Why is there no maximum stated? I want to be sure it can provide up to 4.092V ouput with +/-5V supply. Also, settling time is stated for 0.01%, but I would need 1.9 ppm on 8 Vpp. How can I be sure it will suffice with ADA4922-1 / AD8139?
Are there any advices I could use in solving this issues/uncertainties?