The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> passive mixer linearity improvement
https://designers-guide.org/forum/YaBB.pl?num=1298934228

Message started by kelly on Feb 28^th, 2011, 3:03pm

Title: passive mixer linearity improvement
Post by kelly on Feb 28^th, 2011, 3:03pm

Hi All,

I am looking into improving the passive mixer linearity. I have come across a paper using a predistortion technique. In that paper, there is a reference to a bootstrapped switch technique. I am having trouble getting the boostrap switch paper (it's out of my subscription). Can any of you help me get a copy? It's titled " a bootsraping technique to improve the linearity of a CMOS passive mixer" by Tillman.

Also can someone comment on the differences (pros and cons) between the 2 approaches?

Your help is much appreciated. Thanks.
Kelly

Title: Re: passive mixer linearity improvement
Post by aaron_do on Feb 28^th, 2011, 6:43pm

Hi Kelly,

what kind of application are you looking at? Personally I can't imagine predistortion being used in a receiver. Also, passive mixers are already quite linear.

cheers,
Aaron

Title: Re: passive mixer linearity improvement
Post by kelly on Mar 1^st, 2011, 12:44pm

Hi Aaron,

Just regular baseband receiver. The predistortion paper I mentioned is only for the passive mixer swtches. I think all it does is connecting the mixer swtich (that's biased at strong inversion) output to another device (biased at weak inversion ), so the two combined gives you relatively constant current (which goes through a load resistor). They also have a CLC filter after that which is interesting. the paper is titled "Highdynamic range mixer in CMOS 0.18um technology for WLAN direct conversion receiver".

I think all these papers just trying to get more linearity out of the low voltage passive mixer since the LO is limited to 1V. (at least that's why I need it). I am looking for p1dB of 10dBm with margin.

I think the boostrapping method might be more reliable than the predistortion method. But I am not sure. Have you used the boostrapping method ?

Thanks.
Kelly

Title: Re: passive mixer linearity improvement
Post by aaron_do on Mar 1^st, 2011, 10:09pm

Hi Kelly,

the method you described for the pre-distortion sounds interesting. However, I still have a lot of doubts. First of all, in a passive mixer, the switch changes bias point over pretty much the full range. So it isn't really correct to talk about whether it's biased in the subthreshold region or the strong inversion region. It is more meaningful to talk about the duty cycle (in my opinion). That said, I'm not saying the idea doesn't work.

Another thing to consider is that using this technique, you are increasing the series resistance of the switch. You can compensate it by widening the devices, but this will increase the capacitance. If you are considering adding additional LC filtering at baseband, then you must be looking at a very wide signal bandwidth since inductors are usually used at very low frequencies (on-chip anyway).

Lastly, I have seen a few papers where people talk about trying to cancel the positive gm" with negative gm", but I seriously doubt their robustness to variation.

I think I recall seeing some kind of bootstrap technique using capacitors, but can't recall enough of the details to comment. It does sound more reliable, but you have to make sure its suitable for your application.

hope it works out for you,
Aaron

Title: Re: passive mixer linearity improvement
Post by kelly on Mar 2^nd, 2011, 12:07pm

Hi Aaron,

Thanks. I agreed with your assessment on the predistortion thing. It seems to be an attempt to extend what was done in the linear amplifier (such as LNA). Sounds more empirical to me anyway.

Thanks.
Kelly