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.

transistor length selection

Status
Not open for further replies.

jszair

Junior Member level 3
Joined
Sep 25, 2009
Messages
25
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Activity points
1,479
In Johns and Martins 'Analog IC Design' book, page 222 Basic Opamp Design and compensation, the authors said that reasonable sizes for the length of the transistor might be somewhere between 1.5 and 2 times the minimum transistor length available of a particular technology.

Is this 1.5 ~ 2 times minimum length a rule of thumb in practical design? What's the reason for that? I thought one should select smallest length if possible to reduce capacitance thus minimize power..

any hints?
 

I believe that the authors give that advice because of fabrication issues. The fabrication process is not very accurate and transistors usually come with size errors. The smaller the transistor the greater the error will be.

However many people, such as myself, try to use minimum lengths because it improves bandwidth. It depends on the application, if your performance depends too much on transistor sizes then you shouldn't use minimum length transistors.
 

jszair said:
Is this 1.5 ~ 2 times minimum length a rule of thumb in practical design? What's the reason for that? I thought one should select smallest length if possible to reduce capacitance thus minimize power.
Right. In agreement with fcfusion's contribution I think min. length usage in analog design is justified for transistors which do not have to match with other transistors. For those which have to match (differential stages, current mirrors), however, I use longer than min. size lengths in order to reduce mismatching (ref. to Pelgrom and process mismatch). The required max. mismatch then decides upon the length increase factor, which sometimes can be much larger than 1.5 .. 2 .
 

    jszair

    Points: 2
    Helpful Answer Positive Rating
Another reason is that the output impedance of transistors are very much smaller if minimal length are used, resulting in low gains. Short channel effects will also increase the complexity of the design.
But still, matching forms the most important consideration. It's one thing to design a circuit to write a paper, where you only need one working chip. But for a chip in production, you usually design for 3 to 5 sigma, in which you will find that even designing twice minimal length is often a luxury.
 

    jszair

    Points: 2
    Helpful Answer Positive Rating
Hi,

I have seen L=2um in TSMC 180nm technology Lib, but I don't know what is the maximum practical L in 180nm Tech ? please help me

Thanks

Added after 4 minutes:

Hi,

I have seen L=2um in TSMC 180nm technology Lib, but I don't know what is the maximum practical L in 180nm Tech ? please help me

Thanks

Added after 2 minutes:


I need to use L=1um for a transistor in 180nm. is it impractical? can I use it to design my circuit to write a paper?

Thanks
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top