Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
CMOS process is cheap,less reliable,universal process,low voltage
GaN process is expensive,high reliable,robust,foundry dependent
Both are possible,all depends on the application..
At mm-wave you have few plain-CMOS options and
the ones you do have, are very low voltage. Stacking
costs you geometrically in area and adds losses. A
SiGe device would help.
GaN and GaAs suffer from not having the ability to
do logic or power management, well. If your "application"
wants to be single-chip-does-all, and "all" is any kind
of complex, CMOS rules for the ability to integrate
"stuff" at low power and high density.
If you care only about the radio and it's a simple one,
GaN will have the easiest time at the highest frequency.
But radios today tend not to be simple or single-spec.
GaN devices evidently need some help with linearity
despite having nice fT/fmax. Your linearization scheme
sure would like some CMOS digital learning capability
I'd bet.
Now I've spent 20-plus years in high reliability CMOS
product development, and I have no idea where this
stuff about CMOS being unreliable but too-new-to-
have-datasheets (let alone time in the field) GaN
being high reliability, comes from. I could speculate
a particular orifice.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.