Stats

  • Locked Locked
  • Replies 2
  • Subscribers 125
  • Views 1228
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

atft model (Level=15) does not give Id=0 although sigma0, MUBAND, and IOL are made 0 (zero)

Jena Huh

Tool: Spectre sub-version  19.1.0.455.isr11

According to the equations provided in spectremod.pdf file (Model reference), aTFT RPI model calculates Ids = I_leakage + I_ab

If we make IOL = 0 and sigma0=0, then I_leakage=0. Similarly, making MUBAND=0 will make I_ab=0. So, Ids should be 0. However, I get a constant current output around 1e-11 for Vd=10V and 1e-13 for Vd=0.1V (100 times the Vd is giving 100 times the constant remaining Id).

I am wondering why this is happening. I cross-checked the code with another program called AIM-Spice which also implements LEVEL=15 atft model. AIM-Spice did not reproduce this behavior and Id was 0 for the above parameters. Is this because of the drain side bias voltage source (Vd) introducing a current?

PS: Circuit file and output graphs are attached for checking purpose.

Spectre output:

Spectre output (Id != 0)

AIM-Spice output:

Linear Id-Vg shows Id=0

Checking Log Id-Vg for values near 1E-15 to 1E-10 shows nothing as expected

Code:

IGZO TFT
simulator lang=spectre insensitive=yes

parameters vdd=0

vg (1 0) vsource dc=0
vd (d 0) vsource dc=vdd

model tft atft type=n \
iol=0 \
sigma0=0 \
lambda=0.0 \
muband=0

atft1 (d 1 0 0) tft l=10u w=10u
vgvdsweep sweepVgVd

subckt sweepVgVd()
SweepVdDC sweep param=vdd start=0.1 stop=10 lin=1 {
SweepVgDC dc start=-15 stop=20 dev=vg step=0.1
}
ends sweepVgVd

Parents

  • What you are seeing is the gmin conductance that is added across junctions to improve convergence when junctions are off. Given that circuit simulators cannot really handle complete off devices as this leads to an ill-conditioned matrix (put simply, a floating node means that you have an infinite number of possible solutions for that node), spectre handles this by placing a conductance (using the option gmin) across potentially off junctions. The default value of gmin is 1e-12 Siemens (so 1e12 Ohms) and this is responsible for the leakage current you're seeing. 

    You can reduce it (or even set it to 0) by using something like:

    myOpts options gmin=1e-15

    but bear in mind that this might degrade convergence when devices are off (particularly if set to 0 or a very low conductance). Typically this gmin "leakage" would be less than real leakage in the technology if set appropriately.

    Andrew 

Reply

  • What you are seeing is the gmin conductance that is added across junctions to improve convergence when junctions are off. Given that circuit simulators cannot really handle complete off devices as this leads to an ill-conditioned matrix (put simply, a floating node means that you have an infinite number of possible solutions for that node), spectre handles this by placing a conductance (using the option gmin) across potentially off junctions. The default value of gmin is 1e-12 Siemens (so 1e12 Ohms) and this is responsible for the leakage current you're seeing. 

    You can reduce it (or even set it to 0) by using something like:

    myOpts options gmin=1e-15

    but bear in mind that this might degrade convergence when devices are off (particularly if set to 0 or a very low conductance). Typically this gmin "leakage" would be less than real leakage in the technology if set appropriately.

    Andrew 

Children

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.