The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Mar 28th, 2024, 12:45pm
Pages: 1
Send Topic Print
VDSAT more than VGS-VTh (Read 4177 times)
cmosa
New Member
*
Offline



Posts: 9
IIT
VDSAT more than VGS-VTh
Mar 27th, 2018, 1:55pm
 
Simulating sub 28nm planner technology node IO 1.8V devices.
Device Sizes: W=1u; L=1.5u
Operating points from Spectre sims.

VGS= 730mV
Vth=455mV
VDS= 400mV
VDsat= 330mV


Clearly,  VGS-Vth is 275mV whereas Vdsat is 330mV
I have a very little margin on VDS/VDsat.
If my Vdsat is lower (~275mV as expected), I have enough margin.

I have read and heard that due to channel length modulation effect and/or velocity saturation effect Vdsat is lower than expected VGS-Vth.
however I see VDSAT more than Vgs-Vth !!!

Any explanation for Vdsat > than Vgs-Vth?
Could it be due to metal gate technology?
Could it be some error in the technology file setup since its a relatively new setup for us?


Thanks
Back to top
 
 
View Profile   IP Logged
AnalogDE
Senior Member
****
Offline



Posts: 137

Re: VDSAT more than VGS-VTh
Reply #1 - May 9th, 2018, 12:42am
 
Saturation condition for short channel devices isn't going to match square law.  I suggest you look at whatever small signal parameters are of interest to you and make sure those are still OK at your biasing point.
Back to top
 
 
View Profile   IP Logged
Nyquist
New Member
*
Offline



Posts: 5

Re: VDSAT more than VGS-VTh
Reply #2 - Jul 10th, 2018, 5:11pm
 
Hi,

I think of VDSAT and VGS-TH as two different parameters. VGS-VTH is the overdrive voltage available to me while VDSAT is the voltage that signals the onset of saturation.

Having VGS-VTH < VDSAT means you are operating your transistor in weak inversion/moderate inversion region.

As a rule of thumb, you could also annotate your gmoverid to see if it is > 15. This value might differ for your technology, but it gives a feel for you on where you are operating your transistor.

Can you post the picture of the circuit, you are trying to simulate ?
Also, May I know why the W > L ?
Back to top
 
 
View Profile   IP Logged
A Kumar R
Community Member
***
Offline



Posts: 72
India
Re: VDSAT more than VGS-VTh
Reply #3 - Jul 23rd, 2018, 11:07pm
 
I want to understand this topic a little more deeply.

Why is it that a potential difference between Gate-to-Drain(VGD) < Vth of the device ALONE can't saturate the current through the device?

My point is that once we set VGD < Vth, we have created a pinch-off region near the drain-end and this is the on-set of saturation. So, >= 275mV of VDS is enough to saturate then why should we be concerned about a VDSAT of 330mV at all?

are we saying till we hit a VDS of 330mV device won't saturate, why?
Back to top
 
 
View Profile   IP Logged
Nyquist
New Member
*
Offline



Posts: 5

Re: VDSAT more than VGS-VTh
Reply #4 - Jul 24th, 2018, 12:34pm
 
Hi Kumar,

I get your point. Even I felt the same when I started. VGS-VTH = VDSAT is based on the square law. For short channel devices that do not follow square law, VDSAT is no longer equal to VGS-VTH.

you can refer to the discussion here https://www.edaboard.com/showthread.php?321080-What-are-betaeff-and-vdsat-parame....
Back to top
 
 
View Profile   IP Logged
A Kumar R
Community Member
***
Offline



Posts: 72
India
Re: VDSAT more than VGS-VTh
Reply #5 - Jul 25th, 2018, 8:53am
 
Can I take it as VDSAT being something independent of VTH of the device?
Back to top
 
 
View Profile   IP Logged
Nyquist
New Member
*
Offline



Posts: 5

Re: VDSAT more than VGS-VTh
Reply #6 - Aug 6th, 2018, 11:48am
 
Hi Kumar,

I am not an expert in device technology. But I would take VTH and VDSAT as independent parameters, just like you mentioned for any technology <0.5um

Back to top
 
 
View Profile   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.