mismatch inputs in complex filter for wireless comminication

[776ft]

I am working on a complex filter for wireless communication, 6-stage, biquards structure, for ideal inputs, the outputs is good enough. the image suppression (-2M Hz) is 70dB, but when mismatch occurs in the inputs (0.3dB in magnitude and 1° in phase), the image suppression have reduced to 34dB,(what i need in 35dB at least) which was run in TT case, process variation could make it worse.

considering adding it to 8-stage, and i simulate it, in ideal case, the image suppression is 90dB, but when I make it 0.3dB & 1° mismatch, the image suppression is still 34dB, i don't know whether it is proper.

and how should I improve the image suppression in mismatch inputs?

 

[elbadry]

Re: mismatch inputs in complex filter for wireless comminica

Maybe you can refer to:

"A Pseudo-differential cpmplex filter for bluetooth with frequency tuning", Ahmed A.Emira and Edgar Sanchez, JSSC October 2003.


He doesn't speak about mismatches but he considers an IRR of 45 dB ( image suppresion of 30 dB ) as acceptable.

 

[776ft]

thank you
as you said, he didn't mention about the mismatch case, which i care about much,
i will read that paper in detail.