mr.anderson
Junior Member level 1
I need to design an wideband amplfier with the primary characteristic being low distortion (e.g. all 2nd & 3rd order products below 55 - 60 dBc).
The gain is approx. 30 dB, bandwidth (-3 dB) LF - 1,5 GHz, where LF is a sufficiently low frequency (say 5 - 10 MHz).
The amplifier will be used to drive a differential ADC input with power of approx. -3 dBm, so the amplifier or output stage should probably be balanced.
The problem with these is that manufacturers of MMICs rarely specify anything apart from IP3 measured at 0 dBm or so and P1db, leaving out 2nd order distortions performance and how the MMIC performs with small signals.
The design I'm thinking of is following:
single-ended MMIC amp -> rf balun -> push-pull MMIC amp -> ADC
As for MMICs, the SiGe HBT MMICs from Sirenza seems to be good for this, plus Sirenza has some IP2 data on push-pull & single-ended performance of their MMICs published in appnotes.
By the way, how these days nonlinear simulations are done? E.g. there are software that does that but none of vendors publish any data except S-parameters.
Any ideas how this could be done?
Regards,
The gain is approx. 30 dB, bandwidth (-3 dB) LF - 1,5 GHz, where LF is a sufficiently low frequency (say 5 - 10 MHz).
The amplifier will be used to drive a differential ADC input with power of approx. -3 dBm, so the amplifier or output stage should probably be balanced.
The problem with these is that manufacturers of MMICs rarely specify anything apart from IP3 measured at 0 dBm or so and P1db, leaving out 2nd order distortions performance and how the MMIC performs with small signals.
The design I'm thinking of is following:
single-ended MMIC amp -> rf balun -> push-pull MMIC amp -> ADC
As for MMICs, the SiGe HBT MMICs from Sirenza seems to be good for this, plus Sirenza has some IP2 data on push-pull & single-ended performance of their MMICs published in appnotes.
By the way, how these days nonlinear simulations are done? E.g. there are software that does that but none of vendors publish any data except S-parameters.
Any ideas how this could be done?
Regards,