Designer’s Guide Community

love_analog

Senior Member

Offline

Posts: 101

analog design 28nm and below - poll
Feb 18^th, 2012, 5:45pm

I am curious how people are finding their transistor parameters for analog design 28nm and below

Do you still use gm=2*I/(Vgs-Vt). and I=K*W/L*(Vgs-Vt)^2.

Working with recent college grads - they just put the structure together and parameterize all the transistors and run optimization routines.

This for an opamp with gain, BW specs.

Back to top

loveanalog.blogspot.com
The Power of Analog

IP Logged

raja.cedt Senior Fellow Offline Posts: 1516 Germany	Re: analog design 28nm and below - poll Reply #1 - Feb 19^th, 2012, 1:49am hello, this equations even don't work at 180nm as well...better trust the simulator for finding parameters called gm,gds... and use this parameters to find opamp gain and BW...
Back to top	IP Logged

loose-electron Senior Fellow Offline Best Design Tool = Capable Designers Posts: 1638 San Diego California	Re: analog design 28nm and below - poll Reply #2 - Feb 21^st, 2012, 4:28pm For CMOS Academic equations have never worked properly. Academic simplifications. Run a set of bias curves, use those for approximate bias locations (geometry scaling to set Id aftger setting up Vds and Vgs) After that use a simulator.
Back to top	Jerry Twomey www.effectiveelectrons.com Read My Electronic Design Column Here Contract IC-PCB-System Design - Analog, Mixed Signal, RF & Medical IP Logged

weber8722

Community Member

Offline

Posts: 95

Re: analog design 28nm and below - poll
Reply #3 - Apr 20^th, 2012, 7:11am

love_analog wrote on Feb 18^th, 2012, 5:45pm:

I am curious how people are finding their transistor parameters for analog design 28nm and below

Do you still use gm=2*I/(Vgs-Vt). and I=K*W/L*(Vgs-Vt)^2.

Working with recent college grads - they just put the structure together and parameterize all the transistors and run optimization routines.

This for an opamp with gain, BW specs.

Typically - and independant on process - I am doing normally this:

- If you know you need highest speed, use min L, and find W e.g. to achieve a certain Vdsat or drive-strength.

- If you know matching is important, then better use non-minimum-L and start with meaningful L, find suited W for good matching, Vdsat and noise. Best make some technology investigations on these parameters vs Id, L, W.

- Especially the for current source better use never min-L. Look mostly at Vdsat and matching.

- Of course also the Id needs to be selected, just to fit power budget, noise, drive strength, bandwidth, etc.

- The more experiences you have the closer usually the starting point on Id, L, W is to the optimum values. Number of fingers, etc. I set usually to make the transistor not looking too extreme, best not far away from quadratic. More care is needed for special applications like RF and power-amps.

Bye Stephan

Back to top

IP Logged