about fully differential OP input offset voltage

[chungming]

Is fully differential OPA's input offset voltage very very small (about pV) ?
because it is totally symmetric.
Is it right??
thanks~~~~!!

 

[leichen_gy]

Not true.

there are two main contributors to the offset,
the systematic offset and the random offset,

by considering a fully differential amplifier, everything is symetrical, you may get a small enough systematic offset,

whereas random offset depends on transistors' matching, so it's never 0 nor very small to pV range, it's still in the mV range for current CMOS technologies.

sum up, you still get mV kind of input offset

 

[chungming]

leichen_gy

thanks~~
But how i simulate it with spectre?

 

[pfd001]

u can simulate the offset with monte carlo analysis. u must get support from the fab.

 

[ocarnu]

Monte Cralo is one solution (the best).
The other way is to put DC sources at each pair of transistor. Get matching data from the fab and see the influence.

 

[eldo]

In CMOS OPAMP, the offset is around mV range.

 

[pfd001]

edlo, did u have test experience for opamp offset?
is 1mv possible?