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.

Advanced matching in layout?

Status
Not open for further replies.

alecsander

Junior Member level 3
Joined
Jun 7, 2007
Messages
29
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Activity points
1,473
layout matching techniques

Hi.

Does anyone know a book / paper that describes avdanced matching techniques from layout perspective?

This is the question. Now, details.

When designing a layout for a current mirror, for example, there is a large number of MOS transistors involved. The books i read briefly describe matching of these transistors. I give you the issue that conviced me to open this topic; for a common centroid layout, din all the books, the following pattern for matching is described:

ABAABA
BABBAB
BABBAB
ABAABA

Looks fine. BUT, i just found out that it is NOT good. A better one is:

ABABAB
BABABA
BABABA
ABABAB

In the first one, A dog-bone arrangement is the same in the left part and in the right part, but A-MOS and B-MOS do not have the same stress gradient.
In the second, A dog-bone in the left equals the B dog - bone in the right.

This is only an example. My problem is that i cannot find theory explaining these issues.

So, does anyone know books like this? Thank you very much.
 

I dont know where you got the second layout from but I've never seen it before. From an aesthetic point of view, I prefer the first one. The gradient are more evenly spread across the transistors in the first one than in the latter.
 
I'm sorry, but i really believe the second one is better, and i shall prove it, soon i shall post a picture for better understanding.
 
Is that book, the art of analog layout of alan hastings? in my opinion, the sequence depends on the function and the devices surrounding the pair devices. The most important thing is for the devices experience the same environment.
 
It's a hard to decide what solution is better. But the first one utilizes common centroid technique only and it seems to me this solution is more robust.
In referenced paper authors've used some model to compare different layout matching techniques. U can try to make some investigations of your problem in the same maner.
 
I think, that common centroid layout mean absolutely symmetry relatively Axis of symmetry. Axis of symmetry must intersect center of layout. See example
 
section 1 is better when the process gradients dominate.
But the disadvantage is that the need to cross-connect the devices increases the separation between matched devices.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top