Issue with two stage opamp operating in weak inversion for very low power dissipation

[amitjagtap]

In a conventional two stage opamp operating in weak inversion for ultra low power dissipation, does all transistors operate in weak inversion OR only differential pair transistor and tail transistor operate in weak inversion.
Plz, Need help!
Thanking You.

 

[transbrother]

subthreshold opamp

Usually, we try to avoid putting the current sources in the weak inversion because you'll get a bigger mismatch (since an offset between MOS devices will result in a exponential delta mismatch in currents). Hence, current sources will have to be in saturation.

The input pairs could be in weak inversion to make it high gm and high area to reduce the random offset and have a high bandwidth.

 

[sissi]

weak inversion opamp

hi, transbrother
as you said,The input pairs could be in weak inversion to make it high gm and high area to reduce the random offset and have a high bandwidth. why?
as i know,when the transistor operate in weak inversion, gm is small.

 

[yschuang]

subthreshold opamp for bandgap circuit

I think transbrother is right about gm and offset caused by local mismatch.
gm=Id/(n*Vt) in weak inversion and is good with limited current.

 

[RATUL KUMAR BARUAH]

subthreshold circuit design

when input pairs are in weak inversion gm/id is maximum..pls refer "cmos analog ckt degn" by Philip E. Allen, Douglas R. Holberg.

when all transistors are in weak inversion speed will be very low...so it is advantageous to keep some transistors in moderate inversion for low power ckts.

 

[obi1]

op amp in subthreshold

It depends on the application; i used such an OpAmp in a bandgap circuit, to get the smallest error and it worked great; if the SPEED is not an issue, then go for subthreshould, to get higher gain and low power.

 

[leehying]

operational amplifier weak inversion

hi obi1, what do you mean about 'such an OpAmp ', all MOS in subthreshold, or part of them?

 

[pixel]

op amp subthreshold

In the weak inversion gm~Id and in strong inversion gm~√Id, which means that gm/Id ratio is better for weak inversion, but gm in strong inversion has higher value.

Input stage of differential amplifier in weak inversion is able to compress offset caused by mismatch in the drain. Because of exponential characteristics (same as bipolars) offset is smaller. As previously noted current mismatch could be higher.
For weak inversion you need Vgs-Vth~100-200mV, which could be made increasing W/L ratio and decreasing current Id, which is generally slow.