Differential Amplifier

[ahmad_abdulghany]

differential amplifier above rail

Hello all ..

I want to use ADS to design this easy problem as a tutorial for it ... but i don't know where to start and how !

Please if anyone can help me to design such diffrential amplifier ..

It is a differential amplifier gain stage with the following c/cs:
- Works with resistive loads.
- The required gain is 30 dB and the maximum output signal swing is 3.4V peak-to-peak differential.
- The supply voltage is 3V.
- The current consumption should not exceed 200 uA (micro Amps).
- The design is to be implemented in 0.35 µm technology ( I had attached a SPICE model for it .. but don't knwo how to do similar model for the ADS )

If insteed you can give help in OrCAD PSPICE this will be great
Please try to help with many thanks ...
Note that i am poor in using the ADS

 

[ahmad_abdulghany]

ads differential amplifier

Where are you ?!
Why no one need to reply ?!!
Please reply as fast as possible
Thank you ..

 

[|IAngel|]

Ok - I assume @DS = Agilent's ADS?

Firstly, you need to import those models. You need to load them into ADS. I forget the command for that, but it's very well marked. I can look it up later if you are having issues. I'd suggest splitting up your PMOS, NMOS and RES into different model files to be read in. When I used a 0.13u process in ADS, I split it into every consituent component and read them all into ADS seperately (huge list of devices on the left hand side).

Then, put down your PMOSFET and NMOSFET layout for the amplifier. Under "model" for these, choose your MOSFET model names that you imported. Do this for your resistors as well.

Use a 3V source up top and 0V ground (or +1.5/-1.5V, however you want to do it). Use the standard diff pair equations to ballpark your sizes for gain, CMR, output swing, current.

Also - 3.4V output swing? From a 3V supply? Unless you are designing a complicated opamp, that's not going to happen. Did you mean 2.4V? That's more than reasonable with these specs...

Good luck. If you need more help let me know. If I have time tonight I may just draft it for you and you can do the models...

 

[ahmad_abdulghany]

Ok - I assume @DS = @gilent's @DS?

Yes you are right that is what i mean actually ..

I forget the command for that, but it's very well marked. I can look it up later if you are having issues. I'd suggest splitting up your PMOS, NMOS and RES into different model files to be read in.

Ok .. many thanks .. and i hope you to look for it for me as soon as possible ( remember that i am newbie in ADS and didn't use it before.

Use a 3V source up top and 0V ground (or +1.5/-1.5V, however you want to do it). Use the standard diff pair equations to ballpark your sizes for gain, CMR, output swing, current.

Sorry but i don't understand what will be diffrence in peformance if i had used either 3V up top and 0V ground or +/-1.5 V? and I hope you comment on this point .

Also - 3.4V output swing? From a 3V supply? Unless you are designing a complicated opamp, that's not going to happen. Did you mean 2.4V? That's more than reasonable with these specs...

No, I mean 3.4V not 2.4V, and i didn't think it will be very difficutl.. I not willing to do very complicated OpAmp, but also I may use simple add-ins like simple current mirror, cascoded pair, etc. I assume also that i have an ideal current source from which i will mirror currnt to the ampilifier circuit
I don't need the circuit to be very trivial ( plz don't get angry from me )

Good luck. If you need more help let me know. If I have time tonight I may just draft it for you and you can do the models...

Yeah .. in fact i need a lot of help if you can, although you helped me so much but sorry.. i need more ..

Really, Thank you very much ... You don't imagine how much did you help me !

 

[|IAngel|]

Hmmm interesting. I doubt it's just me, but Capital a is turning into @. WTF is that?

Anywho:

Ok - firstly - take your models and split them into each section. This means - everywhere you see a new "definition" (for example : Model NMOS0) copy and paste this entire sections (for example: End NMOS0) into a file called NMOS.txt. Do this for every device needed (if it isn't already nicely split). If you post your models for me, I can do it for you, but that may (depending on the model source) be a copyright infringment or something (doesn't matter to me - just making you aware of that). I'll look up the command in a few minutes for adding models to an ADS sim.

It changes all the aspects of operating your ciruit. I'm assuming 3V supply means 0-3V : stick with that. Otherwise you will have a -ve to +ve differential range, and other things. This is a common IC configuration, but for your own simplicity use 0V-3V.

I think I misunderstood the output swing part. You mean AC signal Vpp, right? Like 1.7V - (-1.7)V, right? That's possible. I thought you meant the output swing had to be 0V-3.4V DC. You can design for a DC swing above rail, but that's very complicated (compared to what you seem to want). So yeah - that's not an issue.

Do you know the standard equations for a diff. amp? Like for sizing your loads, your input pair, your current sink/source, etc?

Pmos or Nmos input stage?

Is this an assignment, or something you can draft some specs on? I ask because if you have some leaway, it could similify your life.

As for the circuit yourself, I'd suggest the following. Use MOSFETN (under MOS) in the models box for all the fets. Use an Independant current source under SOURCES in the models box. Use resistors from PASSIVE in the model box.

Let me know whatever else you need.

 

[ahmad_abdulghany]

Hmmm interesting. I doubt it's just me, but Capital a is turning into @. WTF is that?

sorry but i don't understand what is ment by WTF.. anyway, i think it isn't that important word

. If you post your models for me, I can do it for you, but that may (depending on the model source) be a copyright infringment or something (doesn't matter to me - just making you aware of that). I'll look up the command in a few minutes for adding models to an @DS sim.

don't worry no copyright there .. it just a faculty small report..

It changes all the aspects of operating your ciruit. I'm assuming 3V supply means 0-3V : stick with that. Otherwise you will have a -ve to +ve differential range, and other things. This is a common IC configuration, but for your own simplicity use 0V-3V.

Ok. Let's work with 0-3V this is what i need

I think I misunderstood the output swing part. You mean AC signal Vpp, right? Like 1.7V - (-1.7)V, right?

yes .. that is right

That's possible. I thought you meant the output swing had to be 0V-3.4V DC. You can design for a DC swing above rail, but that's very complicated (compared to what you seem to want). So yeah - that's not an issue.

I like this and i want to know how (but please let it be later but really i need to know how)

Do you know the standard equations for a diff. amp? Like for sizing your loads, your input pair, your current sink/source, etc?

Of course i know the basic standard equations of CMOS current, gain,input/output impedance, ... etc. but I don't understand what do you mean with sizing equaitons, current sink/source,...etc so please clearify this again

Pmos or Nmos input stage?

I don't know what will be the difference, but if there is no much difference let's assume it's NMOS

Is this an assignment, or something you can draft some specs on? I ask because if you have some leaway, it could similify your life.

I don't understand what do you mean .. anyway it is a report in collage.

As for the circuit yourself, I'd suggest the following. Use MOSFETN (under MOS) in the models box for all the fets. Use an Independant current source under SOURCES in the models box. Use resistors from PASSIVE in the model box.

Ok, I think i can use the Independant current source, but I have to use the MOSFE, and Resistance moedels .. can't I use the MOSFETN from the models box and edit it to be like my current model ? I think it can be but dunno how ..

Let me know whatever else you need.

I wanted to know diffrent types and c/cs of current mirrors and what are parameters with which i can compare..?? such as maximum o/p swing and so on .. I noticed that you used the simplest current mirror in your schematic

Again, Thank you very much for your help ..
Ahmad

 

[|IAngel|]

Firstly: The 'models' issue.

Placing a "MOS" NMOS model won't give you any model data. You have to assign a model to it. What you do is: Go to file -> import -> Choose your model type -> Choose "Save as ADS netlist" -> give your ADS netlist a name. THEN goto "DynamicLink" -> ADD netlist file include. Place it on the layout.

Doubleclick your include and: give your Insance a name. It can be anything. I called mine "MOS_Models". Then add the path the model file you made can be included in. and hit apply. Then choose the file name itself. I called mine "MOSModels.net" ->hit apply. UsePreprocessor - > Yes.

NOW change the name of your NMOS "MOS" models to the name of the model in your model's netlist. This is something like "nch", "pch", "pnp10" for example. You will have to look into your netlist file and get the names of each model yourself.

If this seems too hard to understand, send me your model files and I can do it for you. I'll import them, test them and that should give you a good primer for ADS. After that I'll explain circuit concepts, but it's up to you

That's a good starting place for now. You'll need the models to get anything done. Last question - what is the desired common mode range for your amp? Does it matter?

Cheers!

Added after 1 hours 42 minutes:

Here's your starting point. Look at the cmos7tm.mod and cmos7tm.net files. They tell you your model names and parameters. Then look at what I've done. I've placed an NMOS fet on the circuit. I've assigned it the MODEL: MODN. If you read the cmos7tm.net file, that's the name of one of hte nmos models. Now if you noticed, I placed a VDC source and used a variable in it. I placed a VAR(variable) object and defined Vbias as a variable. I placed a DC simulation on it and swept the Vbias from 0-3V with 301 steps. I placed a probe at the drain to see the current through the diode-connected mosfet.

If you run the sim (or just open the latest results), you will see I've provided a sample graph. This shows you the quadratic gain of the diode connected mosfet. There you have it. The model works, and you can use it for simulation now. I had to edit a couple of the values to make the model file work, but they are less than 0.00000000001% deviated, so you won't notice a difference.

Fiddle around with it and see how you do. If you have any other questions, let me know. This should be all you need to start your circuit.

Finally - to get the ZAP file open, save it and rename it EDABOARD.zap. It needs a zap extension, only that wasn't allowed to be posted with such. Enter ADS. Choose "unarchive project" and select the file you downloaded. This will let you open it. Good luck.

 

[ahmad_abdulghany]

Hello ..
I have some misundertanding..
I created a new project -> then New design units in microns
then, from the new design window : file->import-> "import file name(source)" : i put ther the path where the .mod file containing NMOS model for example exists then hit ok -> Pspice netlist import complete

then, I opened the dynamicLink menu -> add netlist file include :
let's stop here : where is the .net files ? and sorry I don't know what to do with the .zap file you give to me ?

so. I had attached the .mod files separated in one .rar file to kindly do what you say for me

About the CMRR .. i think it is up to me right now but i want to maximize it as possible ..

Thank you for your great help

 

[ahmad_abdulghany]

Sorry..
I could finally open your .zap file .. i didnt see your final sentence about how to open..
but also, I have a silly quistion, does it differ if i made import from the first window or the design window? i.e. may the error be due this mistake..

finally, when i made run to you zap file the shown error message appeared .. do you know what does it mean ? and how to solve ? and did you get results from the circuit ?

I hope you reply in the nearest chance.
Many thanks..
Ahmad

 

[|IAngel|]

That error is a licensing issue, not a design issue. Check your licenses, flexlm and try again.

You shouldn't have to 'import' the zap file.

You have to download it to your work directory. There can't be any spaces in it.

Then you must click File -> Unarchive project in the main ADS window.

then you must choose the EDABOARD.zap file and your directory for it to go to. Then it will uncompress and automatically open. I tried unarchiving my project on here, and it worked beautifully.

Let me know what happens...

 

[ahmad_abdulghany]

Hello |IAngle|
I am sorry for the too late reply, but i was too busy later and i was sick also for some other days, so i couldn't reply for long time.. anyway I finished the design but with OrCAD not ADS but don't get angry for that it was just time limitations. I hope i do many things under your supervision later..
Here is the final results obtained:

I hope you all also share this topic :


Thanks,
Ahmad,