gilbert mixer RF stage

[theboss4]

i'm sort-off (been around a few years, but just recently started designing) new to RFIC design. The issue is my gilber mixer design has a higher 2nd order spur (2xRf) on the output, much higher than is expected, since typically the 2nd harmonic is expect to be >80dBc at the mixer output. I'm trying to isolate the source of the 2nd harmonic by (1) using an ideal LO source (tried it but didn't see that much degradation in the problem spur) and (2) using a ideal mixer RF stage (not sure how best to do this - will appreciate any help on this part). the mixer design is attached: need help with RF stage design, is the any reason why it would produce high second order spur? what about an ideal source?

 

[theboss4]

Gilbert Mixer RF stage causing spur problems

Hi,

i kind of new to RFIC design. The problem i'm having is related to a Gilbert mixer designed for high linearity. my problem in particular is higher even order spurs (2*RF...4*RF) due to the mixer Rf stage (I think). I have tried using an ideal LO but i'm still seeing the higher spurs on the output. The problem (even) spurs are about 50dBc when they should be >100dBc. Can you help me with this issue?

i can provide additional information (i.e. schematic) if requested.

thanks!

 

[FvM]

The shown topology can be expected to generate even harmonics due to the asymmetrical current mirror RF input. Should try a standard Gilbert topology with a differential RF input. Or reduce the RF input level.

 

[theboss4]

thanks for the response, FvM. Sorry it took so long....

unfortunately, a critical system requirement is that the mixer have high IP3 and P1dB, but the standard Gilbert topology (RF stage) is very limited since it a Class A. This is approach (showed in the picture) is a class ab, but the bias current and Re (feedback) resistor have to be chosen for the best linearity and potentially H2d rejection.