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.

Larger current mirror for copying large current

Status
Not open for further replies.

melkord

Full Member level 3
Full Member level 3
Joined
May 18, 2018
Messages
151
Helped
0
Reputation
0
Reaction score
1
Trophy points
18
Activity points
1,772
I am designing a stage source follower with bias current around 90uA.
It seems that I need large size, W and L, to get good accuracy.
Basic formula teaches us that increasing W leads to more id but somehow it does not work here.
What effect explains this? Am I missing something?

Here is a comparison with smaller current mirror which cannot give me anything near 90uA.
1646044109562.png
 

If you want to improve the "accuracy" of the current mirror, you have to do the following two methods.

1) Apply cascode to your current mirror. (improve systematic offset)
This method can reduce the short channel effect in your circuit,
as @Dominik Przyborowski said: "Drain to source voltage. Make it equal".

2) Increase the length of current source, not the width. Make sure the MOSFET operate in saturation region.
(improve random offset)
The reason is that the current mirror mismatch is approximately "(gm/id)AVt",
where the parameter of AVt is proportional to 1/sqrt(W*L),
and the parameter of gm is proportional to sqrt(W/L).
Thus, current mirror mismatch is only proportional to 1/L.
 

2) Increase the length of current source, not the width. Make sure the MOSFET operate in saturation region.
(improve random offset)
The reason is that the current mirror mismatch is approximately "(gm/id)AVt",
where the parameter of AVt is proportional to 1/sqrt(W*L),
and the parameter of gm is proportional to sqrt(W/L).
Thus, current mirror mismatch is only proportional to 1/L.
This is not the issue seen in the described problem. The current is not lower because of mismatch reasons, rather current is lower because of lower output impedance for the device having a small channel length of 100nm.

Channel length plays a major role while deciding the current mirror ratio. Here, the device having 100nm channel length suffers from channel length modulation more, compared to other device having more than 2um channel length. Lower output resistance, resulting in poor current mirror rationing.
 
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top