[Moved]: intrinsic gain and (Vgs, Vds,W/L) relationsihp

Status
Not open for further replies.
S

shanmei

Advanced Member level 1
Joined
Jul 26, 2006
Messages
430
Helped
8
Reputation
16
Reaction score
8
Trophy points
1,298
Location
USA
Activity points
4,496
As shown in the bellowing figure, I plot the intrinsic gain of three nmos transistor with 180nm process.

The Vds of the three transistors are set to 0.75V, and then I sweep the Vgs voltage of all of them from 0 to 1.5V. (The Y-aix is the intrinsic gain, and the X-aix is the Vgs.) Their size are 5/0.5, 5/0.5, and 50/10, with the um unit.

From the figure, it seems that the intrinsic gains are almost the same for the nmos transistors which are biased at the same Vgs and Vds, regardless of their size.

Anyone can explain that why the gain is independent of the size? Thanks.

 

Re: intrinsic gain and (Vgs, Vds,W/L) relationsihp

Gain depends on output resistance, current and Gm:

Ro depends on your L, if you increase the length of transistor you will get higher gain.

So for a transistor with higher length you should get more gain.

---> However I see a minor problem here:

5/0.5 means W/L = 10 but 50/10 means W/L=5 so here, you have higher length but because of lower W/L ratio you obtain lower Gm so your gain does not change.
 
Re: intrinsic gain and (Vgs, Vds,W/L) relationsihp

Thanks. Yes, you are right that one transistor with high gm but lower gain.

However, when I made the three transistors with the size(um): M1: W/L=5/0.5, M2: W/L=5/2, M3: W/L=5/10, and their gain V.S. Vgs plot is shown as the following figure, which shows that the gains are almost the same.

 

Re: intrinsic gain and (Vgs, Vds,W/L) relationsihp

Your data implies that intrinsic gain is maximized at Vgs=0, which makes me suspect that something is wrong with your analysis or setup. How exactly are you defining/measuring intrinsic gain?
 
Re: intrinsic gain and (Vgs, Vds,W/L) relationsihp

Hi mtwieg,

The intrinsic gain is defined by gm*ro or gm/gds.

Although Vgs=0, there are still some current going through the channel, with nA range. In this case, the transistors are not totally off, which is biased in the subthreshold region. Small current generates small gm, however, small current also leads to large ro(small gds), so it might be a maximum value intrinsic gain with Vgs=0. Thanks.
 

Re: intrinsic gain and (Vgs, Vds,W/L) relationsihp

It is gm.ro

So: (Id/vov) X (Va/Id) it means= Va / Vov

when you increase Vgs your Vov increases so gain will decrease.
when you increase L it means you increase Va so gain will increase.

There is some thing wrong with your simulation.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…