Question about cascode current mirror & subthreshold reg

[ccw27]

Cascode current mirror

Hi,

Suppose you have a cascode current mirror, M1 and M2, where M2 is the cascode device. Is it ok for M2 to go into subthreshold region over some corner particularly FF @ -30°C while M1 remains in saturation?

Thanks

 

[Apollo13]

Re: Cascode current mirror

Going into linear is ok. But subthreshold is not modeled properly. So I suggest avoid that

 

[Lantis]

Cascode current mirror

You should not let M2 level saturation any time, or the shielding property of cascode current mirror will be vanished. That is, the current is variated siginificantly with the output voltage, some nonlinearity will occur.

 

[opamp741]

Re: Cascode current mirror

i agree with apollo..u can accept that transistor going into triode..but avoid subthreshold as much as u can..but slightly going into subthreshold is not a big issue..
c ya

 

[MLR67]

Re: Cascode current mirror

On my point of view you need to have both transistors in the saturate region to get the output impedence equal to gm1×gm2×rds2.
When you are in triode region you are decreasing this output resistance , so you have now a bigger spread on your current due to VDS modulation and subtreshold is not good too (MOS very slow ).