gmID Methodology design example

[AMSA84]

Hi guys,

Can someone tell me where can I find some references and examples to design a OTA amplifier using the gm/ID methodology? Specially some examples.

Kind regards.

 

[erikl]

Here are several papers and tutorials which may help you: View attachment gm_over_Id_help-files.zip

[Note to moderators]: No IEEE files included.

 

[AMSA84]

Thanks erikl.

Do you have any experience using this methodology? Did you got good results designing for example amplifiers with it?

Can you give me some tips on how to start or where should I start?

Regards.

 

[erikl]

Do you have any experience using this methodology? Did you got good results designing for example amplifiers with it?

Yes, I did.

Can you give me some tips on how to start or where should I start?

Just study the articles I gave you. You can get more papers on gm/Id usage and W/L calculation if you have access to IEEE papers.

There's a book which helped me a lot to understand and use this method:
David M. Binkley: Tradeoffs and Optimization in Analog CMOS Design.

 

[AMSA84]

Thanks for the reply. I'll take a look at that.

By the way erikl, do you know how to do the slew rate test as well as the icmr and offset on a OTA? Can you give me a tip on how to do that?

Regards.

 

[erikl]

Too many questions, AMSA,

I'd suggest you read some analog circuit design textbooks, or search for your questions in this forum - all of them already have been treated here several times.

 

[AMSA84]

erikl, thanks for the reply.

Regarding the gm/ID, I have read almost everything that you have posted on that .rar file. However, I am a bit confused. How can I get those curves that we can see all over the references that you have posted as well as on other references that I have read too?

I see the procedure, I see some calculations, I see some tips, etc, but what I don't see and I think it's the most important is how to get those curves? Geez.

Thanks in advance.

 

[erikl]

... what I don't see and I think it's the most important is how to get those curves?

Which ones do you think of? gm vs. Id is explained several times. The other curves go by sweeping the abscissa parameter; the ordinate parameter can be evaluated with help of the calculator (in the Cādence simulation environment), or by other calculation tools like Matlab or Excel.

 

[IntuitiveAnalog]

AMSA84 you can find all the test benches in the attachment.

 

[AMSA84]

Hi zulqar. Thanks for the input, however this attached file that you have posted is useless for what I am looking for.

erikl, for example, gm/gds. Where I get the gds? Another example, Cgg?

Moreover, I wonder myself how this can work very well, since all those parameters change with the bias point as well as with the transistor sizes. So I ask, which transistor size I must choose to get all the curves? What bias conditions? Do you understand what I mean?

 

[erikl]

... for example, gm/gds. Where I get the gds? Another example, Cgg?

From a DC analysis, in the transistor parameters. Study the simulation basics!

... since all those parameters change with the bias point as well as with the transistor sizes. So I ask, which transistor size I must choose to get all the curves? What bias conditions? Do you understand what I mean?

I hope so. You run a DC analysis and sweep 1 or 2 parameters, for a special transistor with fixed W and L. For gm/Id you sweep Vgs (or Veff). Otherwise keep the bias constant. Same for frequency sweeps (ac analysis).

And study the textbooks! It's all in there.

 

[IntuitiveAnalog]

how to do the slew rate test as well as the icmr and offset on a OTA? Can you give me a tip on how to do that?

Regards.

I posted that attached file for this query only and if its useless , then I wonder why.

 

[Hongyuan Zhao]

Thank you erikl for your information. It is very helpful for a fresh like me.