The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> unity-gain fully differential rail-to-rail buffer design.
https://designers-guide.org/forum/YaBB.pl?num=1406472477

Message started by Jacki on Jul 27^th, 2014, 7:47am

Title: unity-gain fully differential rail-to-rail buffer design.
Post by Jacki on Jul 27^th, 2014, 7:47am

Hi,

I am designing a fully differential rail-to-rail buffer which is used to drive the 50ohm load and capacitors for the measurement. Normally people like to use the source follower but it will reduce the gain by -6dB. If I want to design a fully differential unity gain rail-to-rail buffer, can anybody give me some hints or references?
Thank you.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by Jacki on Jul 27^th, 2014, 8:26am

By the way, the buffer should have enough linearity because the harmonics are important in the design.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by loose-electron on Jul 27^th, 2014, 11:55am

What frequency and bandwidth?

At low frequencies an op-amp type structure gets used, at high frequencies it is a much more difficult problem.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by Jacki on Jul 28^th, 2014, 12:51am

loose-electron wrote on Jul 27^th, 2014, 11:55am:

What frequency and bandwidth?

At low frequencies an op-amp type structure gets used, at high frequencies it is a much more difficult problem.

Hello loose-electron,

The frequency is not high, just upto 10MHz should be enough, the -3dB bandwidth should be wider, maybe upto 100MHz because I don't want the distortion due to the bandwidth.
Can you please tell more about which kind op-amp type structure, do I need to use feedback to improve the linearity?
Thank you.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by raja.cedt on Jul 28^th, 2014, 6:20am

yes, if you are targeting good linearity then probably -ve fb is the only one option. Design a two stage opamp with class AB o/p stage(like a audio amp). If you are okay to spend some more power a simple diff pair with 50Ohm load and some degeneration might help..

Thanks,
Raj.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by loose-electron on Jul 28^th, 2014, 7:41pm

Suggest using a differential output op amp configured in unity gain feedback.

A lot of open loop amplifiers will have poor linearity unless you burn a lot of current.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by Jacki on Jul 30^th, 2014, 5:38am

Thank you for your advices, loose-electron and raj.
I try different types of continuous time buffer, the simulation results are not good, when the amplitude is below 300mV (600mV full swing), the distortion is OK. But when the amplitude is 500mV, then the distortion is too high (0.18um CMOS). Because the previous stage in front of the buffer is a SC-amplifier, I don't want to use any resistive feedback based on an opamp to be the buffer, then the load for the previous stage will change a lot.
Now I will consider to use the SC buffer with the capacitor feedback (based on an opamp, the last stage could be source follower which can give enough driving ability), I guess it could work with the large dynamic range. Also it could be easier for the fully differential architecture.
At the beginning, I thought maybe I didn't need pay too much attention on the buffer design since it only drove the output signal for the measurement. But now I think I have to change my mind.

Title: Re: unity-gain fully differential rail-to-rail buffer design.
Post by loose-electron on Jul 30^th, 2014, 1:43pm

1. You can configure a buffer opamp in unity gain without the prior stage seeing the resistors. (open up a book on op amps and start looking)

2. Set the op map up to be sampled time as well since you are looking at a discrete time stage.