The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> diode-connected NMOS's equivalent resistance issue
https://designers-guide.org/forum/YaBB.pl?num=1185953977

Message started by trashbox on Aug 1^st, 2007, 12:39am

Title: diode-connected NMOS's equivalent resistance issue
Post by trashbox on Aug 1^st, 2007, 12:39am

Hi guys,
When I was deriving a formula from a simple circuit, I run into the diode-connected NMOS's equivalent resistance issue. The details are:

In the attached figure, the formula Vref1=(1+R1/R2)Vgs1-Vgs2 is obvious for the left circuit when two NMOSs M1/M2 are both in saturation and sink the most of current because of large R1 and R2. However, if using diode-connected NMOS M5/M6 instead of the passive resistance, I can not derive the Vref2 from the right circuit. The simulation shows the formula for Vref2 is right. Would you please give me any advice?

P.S., If I use 1/gm5=R1 and 1/gm6=R2 to derive Vref2 from Vref1, the result is not correct. I think 1/gm is a small-signal resistance and here the equivalent resistance should be a DC resistor.

Title: Re: diode-connected NMOS's equivalent resistance i
Post by boe on Aug 1^st, 2007, 2:21am

Hi trashbox,
The gm is indeed a small signal parameter; for your circuit you need the large signal resistance R = VDS / IDS(VGS).
BOE

Title: Re: diode-connected NMOS's equivalent resistance i
Post by trashbox on Aug 1^st, 2007, 2:53am

Hi Boe,
Yes, you are right. Thanks for your reply. I have solved this issue now. :-)

Best wishes,
Trashbox

Title: Re: diode-connected NMOS's equivalent resistance issue
Post by hande.vinayak on Mar 25^th, 2011, 3:02am

Hi trashbox and boe,
Even after substituting R = VDS / IDS(VGS), I could not get the final expressions for Vref2 as shown in above figure. Please, can you share the method to solve the problem?
Thanks.

Title: Re: diode-connected NMOS's equivalent resistance issue
Post by loose-electron on Mar 25^th, 2011, 1:13pm

Why not just run a DC bias curve and take the resistance from the slope of the line?