The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> why the pole of polyphase filter apart away ?
https://designers-guide.org/forum/YaBB.pl?num=1189391848

Message started by narcissus on Sep 9^th, 2007, 7:37pm

Title: why the pole of polyphase filter apart away ?
Post by narcissus on Sep 9^th, 2007, 7:37pm

Dear All,

I have designed the two stage polyphase filter used for IRR (image-rejection mixer). the two pole of PPF is located in 4MHz and 6MHz in the simulation result. But when I measure the some sample chip , these two pole apart away to 1.8MHz and 8.6MHz.

first I suspend the mismatch from mixer , then I add some mismatch to mixer ,the simulation result show two pole will apart away a little ,but the image-rejection ratio will degrade with mismatch level. even when the mismatch of mixer to some unbelievable level in order to meet the two pole's location in measurement. the image-rejection ratio of simulation down to 12-14 dB. but image-rejection ratio of measure result still keep in 21-23 dB. so may be have other reason for this. but Resistor and capacitor 's process variation will not so much, right?

thanks for any sugggestions will give me to debug this problem?

Title: Re: why the pole of polyphase filter apart away ?
Post by sheldon on Sep 16^th, 2007, 9:50am

Naricissus,

It is a little hard to follow your append. It appears that you designed a
polyphase filter for image rejection filter and there is less image rejection
than you expected.

Questions/Comments:

1) Is the filter an active or a passive filter?

2) Did you perform corner analysis on the filter design?
What does simulation predict the filter rejection is across
process?

3) What process are you using and did you perform LPE simulation?
Device sizes change when printed on a wafer due to various effects:
resistors are sensitive to overetch, ... Ideally, you would perform
LPE, layout parameter extraction, and use the actual device sizes
for simulation.

4) Did you perform RCX and include the layout parasitics in your
simulation?
RCX, parasitic RC extraction, extracts the layout parasitics

NOTE: LPE is not the same as RCX, you need to account for both
effects when designing

5) Do you have TEG data for the lot?
If you skew the process based on the TEG data, are the results
consistent with measurement?
For example, if the the target for polysilicon sheet resistance is 20Ohms/sq
and the measured value is 15Ohms/sq., scale the resistance values and
re-simulate r_actual = (20/15)* r_target

Best Regards,

Sheldon

Title: Re: why the pole of polyphase filter apart away ?
Post by narcissus on Sep 18^th, 2007, 7:01am

Dear Sheldon,

Thank you for your suggestions. because last time I test the PPF 'performance by external LO signal(by signal generator), but this week I using the internal LO signal(from VCO), the test result is ok . I guess that root cause may be the external LO signal path to DIV2 and mixer is not symmetrical.

Thanks again.

BR,
narcissus