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

criteria to choose the width of a single finger in a design

Status
Not open for further replies.
jimito13

jimito13

Advanced Member level 1
Joined
Nov 17, 2008
Messages
464
Helped
130
Reputation
262
Reaction score
113
Trophy points
1,323
Location
Greece
Activity points
4,929
Hello all,

I have a very fundamental question concerning the start up phase of a design.What are the criteria to choose the width of the single finger (Width Single Finger as referred at the properties tab of a fet in cadence) that will reproduce all the transistors in my design?Is there any rule of thumb for this choise?

I know the question is very general and maybe depends on the nature of the design (baseband/RF) or the specifications or other factors (matching etc...) but i left it to be so general in order to expect opinions from every possible direction.

I would appreciate your precious suggestions.Thanks in advance.

Jimito13
 

Hi Jimito,

for me, this depends on the floorplan of my intended analog cell: Usually, I roughly know the required x*y dimensions of the cell, then sketch a crude outline of its internal devices depending on the desired (relative) position of inputs and outputs. The needed areas for the "big" devices - input and output transistors, caps - regarding matching and common centroid aspects then decide upon their location in the cell and their aspect ratio, and from this I calculate a convenient w = number_of_fingers * Width_Single_Finger.

HTH! erikl
 
For high speed / RF, you want many narrow fingers to hold
down Rg. For analog you want wider, as much mismatch (esp
at weak inversion) can come from the channel edge (bird's
beak) or strain effects (STI). Your foundry may or may not
break down mismatch into area, w, l dependencies (often
only area is discussed; whether this is because the others
don't matter, you would like to know). You could deduce this
from looking at MM-vs-area plots' detail, whether narrow*long
and wide*short occupy different points on the map for same
area.

I don't think there's value in assigning all devices the same
finger width, in analog. Outside of a must-match group, at
least.
 
Thanks a lot guys for your helpful answer.I will still be awaiting for more opinions if possible.Thanks in advance!
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top