Sub-threshold saturation--is that ok?

[knights]

I did an opamp and found when input some voltage,the input differential transistors
was in sub-threshold saturation.It mens Vgs<Vth,Vds>Vdsat.And the AC simulation shows the DC gain and PM is the same as other input voltage which is in
the input voltage range.
Is that ok for the opamp?
The architechture is simple folded cascode architechture.Thanks.

 

[DenisMark]

Sub-threshold (aka weak inversion) region is normal for diff. pair of OPAMP.
In this region u've maximized transconductance efffectiveness Gm/Id. This gives u an advance in terms: gain (Gm*Ro=>(Gm/Id)*Va), GBW (Gm/Cc), Input Refered Noise, Input Offset/Matching (offset close to threshold voltage mismatch). At the another side the slewrate is decreased (requere slew-rate boosting technique) and PM because high gain and increased GBW.
Somebody'll add what i forget.

 

[pfd001]

it is ok

 

[knights]

Sub-threshold (aka weak inversion) region is normal for diff. pair of OPAMP.
In this region u've maximized transconductance efffectiveness Gm/Id. This gives u an advance in terms: gain (Gm*Ro=>(Gm/Id)*Va), GBW (Gm/Cc), Input Refered Noise, Input Offset/Matching (offset close to threshold voltage mismatch). At the another side the slewrate is decreased (requere slew-rate boosting technique) and PM because high gain and increased GBW.
Somebody'll add what i forget.

Thanks first. You mean that the transistors in sub-theshold region can get the larger Gm than that in normal saturation region with the same size?And then
result in high gain and large GBW and so on.

 

[DenisMark]

Read it first.



Yes, Gm increased for same current, u can keep size (W*L) by increasing ratio W/L to some limit. Keep in mind that short chanel devices will have low Rout and suffer from increased mismatching. Very large GBW gives an issue with PM, because high frequency poles and zeros are apeared. It's subject of trade-off.

 

[jeepx5]

It depends on your application

 

[gte582w]

In some rr OpAmp design, people intentionally bias the input MOS in sub-vt region for better and simple gm variation control.

 

[lghhust]

gm=I/(nkt),so gm may be nearly constant,that's why Gain &PM remain the same.