The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> Mixer linearity
https://designers-guide.org/forum/YaBB.pl?num=1214549712

Message started by vivarf on Jun 26^th, 2008, 11:55pm

Title: Mixer linearity
Post by vivarf on Jun 26^th, 2008, 11:55pm

Dear all

I am designing a direct conversion mixer working at 1 GHz.
As pointed out at thomas Lee paper: http://www-smirc.stanford.edu/papers/cancun97s-chet.pdf

The linearity of sub-micron transistor no longer depends on the ratio W/L, but the bias DC current.
However in my design, I dissipate a lot of DC current, the linearity is not improved. Am I missing something here, any comments would be appreciated.

Title: Re: Mixer linearity
Post by didac on Jun 27^th, 2008, 2:01am

Hi,
mmm, not sure but I think that exists an optimum DC consumption where linearity stops to improve since it's dominated by the switching pair, I've used "bleeding" to put more current into the transconductor and reduce the amount that goes to the switching pair.
Hope it helps,
PD:You can check :"Intemodulation distortion in curren-commutating CMOS Mixers",Journal of Solid-State Circuits,vol.35,no.10,october 2000,Manolis T. Terrovitis, Student Member, IEEE, and Robert G. Meyer, Fellow, IEEE. Figure 20 separates the linearity between transconductor and switching pair.
PD2:you can look for linearitzation techniques also, like tanh or source degeneration.

Title: Re: Mixer linearity
Post by vivarf on Jun 28^th, 2008, 12:11pm

Hi didac, thank you very much for you respond

I checked the paper, you were right that the switches will define the linearity when dissipating a lot of current.

The others techniques you suggested (tanh or source degeneration) might be not efficient since the gain is sacrificed. And yet, this still a question for me: why those techniques are used ? Reducing the gain of transconductance itself (reducing transistor width: W) will improve the linearity as much as those techniques do.

Title: Re: Mixer linearity
Post by didac on Jun 28^th, 2008, 12:22pm

Hi,
Mmmm, I don't see as obvious that you directly reduce gain in the same proportion using linearitzation or reducing the size of the transistor so maybe it's necessary to compare the performance in terms of dynamic range(SFDR-Spurious Free Dynamic Range-) and see how well it behaves the mixer... I will try to think a little bit more about this.
Hope it helps,(although that I doubt that it helped )

Title: Re: Mixer linearity
Post by aaron_do on Jun 28^th, 2008, 7:50pm

Hi vivarf,

the way i understand it, reducing W (and current proportionally) will not improve input-referred linearity at all (assuming the nonlinearity is coming from the input transistor). On the other hand, source degen can provide a much greater reduction in the nonlinear components than the fundamental.

cheers,
aaron