The error term 5 22y F 2 predicted by Eq. (4.141) for the Wilson mirror of relies on perfectly matched betas. This matching causes fi rstorder terms in 1yF to cancel each other out, leaving only the...

The error term


5 22y F 2 predicted by Eq. (4.141) for the Wilson mirror of relies on perfectly matched betas. This matching causes fi rstorder terms in 1yF to cancel each other out, leaving only the quadratic term. (a) To investigate what happens if the betas are mismatched, assume F2 5 F1 for simplicity, but arbitrary F3. Proceeding in a manner similar to that depicted in Fig. 4.61b, show that i F3 5 100, 90, 110, and 101, and comment on your fi ndings. (c) A student, worried about the error due to fi nite betas, is thinking of replacing Q3 with a Darlington pair to exploit its high equivalent beta, or even with an nMOSFET, which draws zero gate current. What advice would you offer to that student?