Designer’s Guide Community

neoflash Community Fellow Offline Mixed-Signal Designer Posts: 397	Does TSMC Spice model include impacted ionization effect? Jan 28^th, 2011, 10:18pm In simulation, people see the I-V curve will have similar behavior when the Vds is very large. Not sure if this is due to the modeling of Ib due to impacted ionization?
Back to top	IP Logged

loose-electron

Senior Fellow

Offline

Best Design Tool =
Capable Designers

Posts: 1638
San Diego California

Re: Does TSMC Spice model include impacted ionization effect?
Reply #1 - Jan 29^th, 2011, 2:00pm

probably not - but you need to get a lot more specific in which foundry process, what models are used, etc, etc -

When you get to that level of detail, you can then look at the model parameters and figure it out yourself.

Time to be an engineer!

Genrally speaking, TSMC models are not that good, a lot of places use the foundry and develop their own models for it.

Back to top

Jerry Twomey
www.effectiveelectrons.com
Read My Electronic Design Column Here
Contract IC-PCB-System Design - Analog, Mixed Signal, RF & Medical

IP Logged

neoflash Community Fellow Offline Mixed-Signal Designer Posts: 397	Re: Does TSMC Spice model include impacted ionization effect? Reply #2 - Jan 29^th, 2011, 3:01pm Thanks for the heads up. I just want to get some quick answer before I get into sleep last night. I did some simple google search and yes, this effect is included in the bsim model.
Back to top	IP Logged

vivkr

Community Fellow

Offline

Posts: 780

Re: Does TSMC Spice model include impacted ionization effect?
Reply #3 - Jan 31^st, 2011, 6:23am

neoflash wrote on Jan 29^th, 2011, 3:01pm:

Thanks for the heads up. I just want to get some quick answer before I get into sleep last night.

I did some simple google search and yes, this effect is included in the bsim model.

What do you mean neoflash? Sure! impact ionization can be modelled using the BSIM4 model, but only if the foundry has bothered to measure the relevant parameters and keyed them into the model file you are using. You can check your model file for parameters: alpha0, alpha1 and beta0 (For a description, do spectre -help bsim4, and for the entire expression for impact ionization current, check the BSIM4 manual). Usually, the alpha parameters are set to zero when the effect is not characterized by the foundry. Of course, the foundry could also have dummy non-zero numbers which are not accurate, or do not show the true variation etc.

So, you can either look through all the details if this effect really matters to you, or else engineer your way around it. There are many ways of doing that.

Vivek

Back to top

IP Logged

modelman

New Member

Offline

Posts: 5

Re: Does TSMC Spice model include impacted ionization effect?
Reply #4 - Feb 1^st, 2011, 1:37pm

loose-electron wrote on Jan 29^th, 2011, 2:00pm:

Genrally speaking, TSMC models are not that good, a lot of places use the foundry and develop their own models for it.

Do you know in general terms what are the things lacking in TSMC models? I am looking at making some updates to their models and I am wondering if there are any major gaps I should be looking for in terms of usability for analog mixed signal design?

Back to top

IP Logged

neoflash

Community Fellow

Offline

Mixed-Signal
Designer

Posts: 397

Re: Does TSMC Spice model include impacted ionization effect?
Reply #5 - Feb 1^st, 2011, 3:37pm

vivkr wrote on Jan 31^st, 2011, 6:23am:

neoflash wrote on Jan 29^th, 2011, 3:01pm:

Thanks for the heads up. I just want to get some quick answer before I get into sleep last night.

I did some simple google search and yes, this effect is included in the bsim model.

What do you mean neoflash? Sure! impact ionization can be modelled using the BSIM4 model, but only if the foundry has bothered to measure the relevant parameters and keyed them into the model file you are using. You can check your model file for parameters: alpha0, alpha1 and beta0 (For a description, do spectre -help bsim4, and for the entire expression for impact ionization current, check the BSIM4 manual). Usually, the alpha parameters are set to zero when the effect is not characterized by the foundry. Of course, the foundry could also have dummy non-zero numbers which are not accurate, or do not show the true variation etc.

So, you can either look through all the details if this effect really matters to you, or else engineer your way around it. There are many ways of doing that.

Vivek

Thanks Vivek.
I usually deal with such kinds of things in "shallow" approaches. I will just email the TSMC client support instead of spending hours on simulation and analysis. LOL.

Back to top

IP Logged