The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Strange input pair in cascode OPAMP
https://designers-guide.org/forum/YaBB.pl?num=1177753727

Message started by cchen on Apr 28^th, 2007, 2:48am

Title: Strange input pair in cascode OPAMP
Post by cchen on Apr 28^th, 2007, 2:48am

The schematic of OPAMP was shown in the attached figure. Besides the main two input transistors, there are two extra cross-coupled transistors with smaller sizes. Does anyone know the function of them? It seems they will reduce Gm of OPAMP.

Thanks for your advice!

Title: Re: Strange input pair in cascode OPAMP
Post by carlgrace on Apr 29^th, 2007, 7:16am

Those two NMOS devices clamp the output to increase the slew rate of the op amp. You can see a discussion of their use in Johns and Martin.

Carl

Title: Re: Strange input pair in cascode OPAMP
Post by RobG on May 3^rd, 2007, 5:28pm

To expand on Carl's statement, they keep the cascode transistors biased up even when the opamp is slewing.

Normally, during slewing, all of the current would be to one side, debiasing the cascode devices. The cross coupled device steers some current back to the cascode device.

Title: Re: Strange input pair in cascode OPAMP
Post by eng on May 5^th, 2007, 9:56pm

I wonder if this cross-coupled pair reduces the output inpedance by generating a negative resistance?

Title: Re: Strange input pair in cascode OPAMP
Post by nobody on Jan 3^rd, 2009, 7:01am

I just look around what in my favorites and found that I can answer the question. The extra cross-coupled pair is for neutralization reducing miller effect. More details of neutralization can be found in the textbook by Paul, Gray, and Meyer.

Title: Re: Strange input pair in cascode OPAMP
Post by raja.cedt on Jan 4^th, 2009, 8:32pm

hi,
i didn't understand why they did biasing like that..can any body explain ..
Thank you.

Title: Re: Strange input pair in cascode OPAMP
Post by superken on Feb 28^th, 2009, 11:59pm

This is a telescopic amp with gain-boosted n-cascodes. The purpose of level-shift circuit is to allow more headroom for the main amp.. Without it, additional Vt will appear in the headroom requirement of the input diff pair. Instead of 2Vdsat, we would need 2Vdsat+Vt to keep input diff pair and tail current source of the main amp in saturation, hence additional Vt reduction in the available output swing.

The progressive sizing of the bottom two triode transistors in "nVreff generator" on the far left is to account for backgate bias for the stacked transistor with size (1/3) to make it slightly stronger than the one below it due to its larger Vt. That's why they got the same Vds = 400mV.

And I agree with the functions of the small cross-coupled pair as explained by the previous posts above.

Title: Re: Strange input pair in cascode OPAMP
Post by nobody on Mar 3^rd, 2009, 10:29pm

I just realized the opamp is actually from the following paper.
A 3.3-V 12-b 50-MS/s A/D converter in 0.6-μm CMOS with over80-dB SFDR

Title: Re: Strange input pair in cascode OPAMP
Post by nobody on Mar 6^th, 2009, 10:13pm

I do not quite understand why the level shifter can increase output swing. The level shifter can set the voltage @ the drain of input pair to be 2Veff and a biasing CKT can do that too. I was wondering why a term Vt comes up without a level shifter.