The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Semiconductor Devices >> Plotting Vth variations
https://designers-guide.org/forum/YaBB.pl?num=1173197046
Message started by Croaker on Mar 6th, 2007, 8:04am

Title: Plotting Vth variations
Post by Croaker on Mar 6th, 2007, 8:04am
Hello all, I'm wondering if anyone knows how to do a plot of how Vth varies due to process mismatch...  I guess I'm looking for a graph similar to a bell curve to show what Vth values I'm likely to get.

Thanks!  :)

Oh, I'm using Cadence.

Title: Re: Plotting Vth variations
Post by sheldon on Mar 6th, 2007, 8:32pm
Croaker,

  Not sure I understand exactly what you are looking for. Do you want
to plot distribution of delta Vt? Do you want to plot the correlation between
Vt and mismatch [delta Vt]? Do you want to plot correlation between Vt
and delta L, delta tox, ...? Do you want to plot correlation between delta
Vt and deltaL, delta tox, ...?  It should be possible plot the results, just
need a understand what you would like to plot.

                                                              Best Regards,

                                                                 Sheldon

Title: Re: Plotting Vth variations
Post by Geoffrey_Coram on Mar 7th, 2007, 5:39am
In some processes, the VTH0 parameter is set by calls to a Gaussian distribution function.  In this case, you can just take the mean and std dev of the Gauss call and plot the distribution directly.

In some cases, though, TOX and VFB (which indirectly change Vth) are varied, and then you'd need to define what you mean by "Vth" ... and probably run some M-C simulations to "measure" it for your devices.

Title: Re: Plotting Vth variations
Post by SATurn on Mar 18th, 2007, 7:00am
The question is that how it is possible to plot (or print) the VTH for example in Cadence?

Saturn

Title: Re: Plotting Vth variations
Post by Geoffrey_Coram on Mar 19th, 2007, 4:23am

SATurn wrote on Mar 18th, 2007, 7:00am:
The question is that how it is possible to plot (or print) the VTH for example in Cadence?


What is VTH?  With subthreshold current, Id != 0 for Vgs < VTH, so you need to decide what you want to use as a definition of VTH: is it based on the value of drain current, the slope of the Id/Vg curve, ...  BSIM3/4 have a VTH output parameter, which is computed from VTH0 and other parameters, but it does not include things like GIDL that might actually change where a circuit designer would say the "threshold voltage" is based on the Id/Vg curve.

Title: Re: Plotting Vth variations
Post by Croaker on Mar 19th, 2007, 8:59am

SATurn wrote on Mar 18th, 2007, 7:00am:
The question is that how it is possible to plot (or print) the VTH for example in Cadence?

Saturn


I do know how to plot Vth...or are you looking to find out?

Title: Re: Plotting Vth variations
Post by Croaker on Mar 19th, 2007, 9:03am

sheldon wrote on Mar 6th, 2007, 8:32pm:
Croaker,

  Not sure I understand exactly what you are looking for. Do you want
to plot distribution of delta Vt? Do you want to plot the correlation between
Vt and mismatch [delta Vt]? Do you want to plot correlation between Vt
and delta L, delta tox, ...? Do you want to plot correlation between delta
Vt and deltaL, delta tox, ...?  It should be possible plot the results, just
need a understand what you would like to plot.

                                                              Best Regards,

                                                                 Sheldon


Hi Sheldon, I want to have something like a pdf so I can see what my chances of getting certain Vths are.  

Title: Re: Plotting Vth variations
Post by Croaker on Mar 19th, 2007, 9:04am

Geoffrey_Coram wrote on Mar 19th, 2007, 4:23am:

SATurn wrote on Mar 18th, 2007, 7:00am:
The question is that how it is possible to plot (or print) the VTH for example in Cadence?


What is VTH?  With subthreshold current, Id != 0 for Vgs < VTH, so you need to decide what you want to use as a definition of VTH: is it based on the value of drain current, the slope of the Id/Vg curve, ...  BSIM3/4 have a VTH output parameter, which is computed from VTH0 and other parameters, but it does not include things like GIDL that might actually change where a circuit designer would say the "threshold voltage" is based on the Id/Vg curve.


For the purposes of this thread, I will say Vth is the operating point value that spectre spits out.

Title: Re: Plotting Vth variations
Post by Geoffrey_Coram on Mar 19th, 2007, 11:15am

Croaker wrote on Mar 19th, 2007, 9:04am:
For the purposes of this thread, I will say Vth is the operating point value that spectre spits out.


If that's what you want, and you know how to access (plot) Vth, then I would expect you could do a Monte-Carlo and set your expression to be that Vth.  See "spectre -help montecarlo"

Title: Re: Plotting Vth variations
Post by Croaker on Mar 19th, 2007, 1:46pm

Geoffrey_Coram wrote on Mar 19th, 2007, 11:15am:

Croaker wrote on Mar 19th, 2007, 9:04am:
For the purposes of this thread, I will say Vth is the operating point value that spectre spits out.


If that's what you want, and you know how to access (plot) Vth, then I would expect you could do a Monte-Carlo and set your expression to be that Vth.  See "spectre -help montecarlo"


Yes, I've done that but it just shows a bunch of Vth values.  I was looking for something more histogram-ish.   :-/

Title: Re: Plotting Vth variations
Post by ACWWong on Mar 19th, 2007, 3:19pm
in the cadence monte-carlo GUI, histogram ploting of your outputs (i.e. vth expression) is available under results, as are loads of other useful plot types.

Title: Re: Plotting Vth variations
Post by Croaker on Mar 20th, 2007, 5:44am
OK thanks, I'll look into that!  Sounds like what I wanted.

The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.