juzi wrote on Sep 13th, 2006, 8:44pm:I wanna to design a 2.4GHz CMOS PA on chip, now I have the TSMC 0.35um SiGe BICMOS model,
the RF MOSFET Vbox=0.75V.
when the input voltage bigger than 1V, the parasitical diode between the source and the bulk is forward based, and the SiO2 is breakdown .
Vbox=0.75V is because the tox is only 7.35nm, but if the tox is increasing, the fT of the MOSFET will be decreasing, which is unacceptable for a 2.4G circuit!!!
can someone give me any advise?
Hi,
Firstly I don't see how you are forward biasing the source bulk diode. If its NMOS then the bulk is at ground. The source would need to go negative of ground for forward bias (in which case you are doing a very strange PA). If its PMOS then the you might suffer DRAIN to bulk diode forward bias due to over voltage swing on the drain. Is this what you mean ?
Secondly tox=7.35nm is thick, so Vbox should be a few volts. Can't see how it is only 0.75V. Maybe some mistake in the kit.
I am using tox ~2nm and Vbox > 1.5V in a digital CMOS process (so SiO2).
cheers
aw