sunjiao3
Member level 5
Hi, friends. I encounted a problem now on the simulation of mismatch.
I am simulating a comparator, and I found the mismatch of the main transistor paris affect the performance greately. I wanna simulate the possible mismatch in full-scale with credible result. Would anyone of you give me some suggestion?
Well, I have came up with several approach. First, I think about including some model into my lib. However, which one is credible. I read some papers on this topic, each with a "reliable" model. I know, the modle can not be separated from the technic. So, is there a general and "universal" one?
Second, I just change the size of one of the pairs, like 5%,10% or so. However, mismatch is random in fabrication. And the geometrical mismatch is just among one of them. I think this method is limited.
Thank you in advance.
I am simulating a comparator, and I found the mismatch of the main transistor paris affect the performance greately. I wanna simulate the possible mismatch in full-scale with credible result. Would anyone of you give me some suggestion?
Well, I have came up with several approach. First, I think about including some model into my lib. However, which one is credible. I read some papers on this topic, each with a "reliable" model. I know, the modle can not be separated from the technic. So, is there a general and "universal" one?
Second, I just change the size of one of the pairs, like 5%,10% or so. However, mismatch is random in fabrication. And the geometrical mismatch is just among one of them. I think this method is limited.
Thank you in advance.