Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

how do I calculate the channel-length modulation from proces

Status
Not open for further replies.
F

field_catcher

Junior Member level 3
Joined
Dec 27, 2004
Messages
26
Helped
3
Reputation
6
Reaction score
4
Trophy points
1,283
Activity points
332
how to calculate channel-length modulation

Dear All,

I have seen some posts regarding this issue. But no one tells how to "Calculate" the lambda from process parameters. I know how to measure Lambda by doing simulation. But I want to know how to calculate it too.

My justifcation to this question is:

Lambda is listed as a prameter in level 1 and 2 model, but not in level 3 model. There must be some equation we can derive the lambda from level 3 model parameters.

Can anyone help? No matter how complex or tricky the equation is?



Thank you all!



Jianhong
 

channel length modulation

λ=sqrt(2ε/qNa) / (2*L*sqrt(Vds-Veff+phi))

ε = permitivity of the substrate
q = electron charge
Na = impurity concentration of the substrate
L = length
Veff = Vgs-Vt
phi = built in potential between diffusion and substrate

Simulating a single transistor will give you better results though.
 

calculate channel length modulation

plot Id vs Vds for different Vgs and measure the slope of the curve in sat region and use lambda =(Id2-Id1)/(Vds2-Vds1)

hope this works
 

channel length modulation lambda calculation

Mince,

Thanks for the equation. Can you tell me where you get the equation (e.g the reference)? Thank you!



Jianhong
 

Re: channel length modulation lambda calculation

If you are using Cadence Virtuoso software, the value of Gds in the device operating parameters of the transistor can give you the lambda. because Gds = 1/r0 , where r0= 1/ (lambda*Id).
 

In a modern process at minimum channel, good luck calculating
anything. Or even getting good numbers out of your foundry
to work with. With halos and engineered drain regions there's
nothing like a plain long channel MOS approximation going on.
I'd shotgun the relevant params for length modulation until
it looks like the other model (or better yet, real curves).
 

This book on gm/Id methodology gives a good explanation of how Early voltage (1/lambda) can arise more from DIBL than channel length modulation in modern processes
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top