Welcome to our site! EDAboard.com is an international Electronic 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.
He wanted to multiply the current in M19 to be 5 times the current in the transistors chain
so what is done is using a unit transistor of 1.28u/300nm and in the chain a 5 transistors in series that's like multiplying L by 5
so in the reference mirror, the chan will have a size of 1.28u/(0.3u*5) and on the mirror branch the size will be 1.28u/0.3u
which 5 times higher
the idea of using a unit transistor is to simplify layout and matching techniques (common centroid and interdigitization)
I don't think this the best way to design the current mirror with different L sizes because this will affect the current due to the 2nd order effects
(Like Vth and channel length modulation)
Integer ratio matching is superior for litho effects.
The alternative would be to use a 1X reference
and a 5X output device, but maybe there was
an interest in keeping current density higher
(such as leakage floor or low current "gate kink"
observed to be a threat to fidelity). Particularly
when VT is not so high that you are on the
junction leakage floor by Vgs=0, there could be
a loss of ratio control at some corners if current
density is let to sit too low.