Designer’s Guide Community

kidman

Community Member

Offline

Posts: 42

Question on current mirror implementation
Jan 31^st, 2009, 3:52pm

A current mirror that consists of only 2 transistors won't copy the exact reference current due to channel length modulation. The solution to that is cascoding as well know to shield the mirror from VDS variation and copy the exact reference current.

My question is, is the normal current mirror that's not cascoded used in any practical design? I'm asking this because I'm wondering how does a designer get his circuit working after the PVT variations and so on using the basic current mirror that even without PVT variations will not copy the current accurately.

Back to top

IP Logged

loose-electron

Senior Fellow

Offline

Best Design Tool =
Capable Designers

Posts: 1638
San Diego California

Re: Question on current mirror implementation
Reply #1 - Feb 1^st, 2009, 12:28pm

They are used all the time. Many applications for a current source do not require either good absolute value, or good PVT stability. Example - the current source under a differential pair can usually function properly with a lot of variance (2X over PVT without problems)

If you are looking for tight tolerance on any circuit, it is going to require adjustment circuits or the architecture needs to be reviewed to get rid of tolerance needs.

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

ywguo Community Fellow Offline Posts: 943 Shanghai, PRC	Re: Question on current mirror implementation Reply #2 - Feb 1^st, 2009, 7:51pm Kidman, The simple current mirror are often used. That depends on the performance requrements. Yawei
Back to top	IP Logged

raja.cedt

Senior Fellow

Offline

Posts: 1516
Germany

Re: Question on current mirror implementation
Reply #3 - Feb 2^nd, 2009, 3:07am

hi kidman,
Based on the application type mirror being used is decided,for example in differential amplifiers if the i/p variations are not much then simple mirror is the first choise,but if u want really high CMRR then u have to go for cascode.again cascode consume a lot of headroom for voltege,in that case high swing mirror is the best choice.This information u can get from razaavi.

Thank you.

Back to top

IP Logged

kidman Community Member Offline Posts: 42	Re: Question on current mirror implementation Reply #4 - Feb 5^th, 2009, 5:08am thanks for your i/p guys
Back to top	IP Logged