The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Semiconductor Devices >> Model QA
https://designers-guide.org/forum/YaBB.pl?num=1141350200

Message started by ayow on Mar 2^nd, 2006, 5:43pm

Title: Model QA
Post by ayow on Mar 2^nd, 2006, 5:43pm

I would like to build up an in-house model QA ability for my Lab. The model issues we usually have are in-continuity, difficult convergence and in-accurate in some W/L region..etc.

Could anyone share your idea about model QA process ?

Thanks a lot.
Ayow

Title: Re: Model QA
Post by Paulin on Jun 9^th, 2006, 2:56pm

Hi,
Some modeling tool like BSIMPro+ from Cadence, has the capbility to automatically Q/A the model. If you want to see some documentation regarding compact model Q/A, you could visit www.eigroup.com/cmc

Hope this help!
Paulin

Title: Re: Model QA
Post by Geoffrey_Coram on Jun 12^th, 2006, 3:38am

I've heard some people talk about a program called MQA (model QA) from Accelicon.

http://www.accelicon.com/mqa.html

Title: Re: Model QA
Post by loose-electron on Aug 1^st, 2006, 8:36pm

MQA automates a lot of the hand check stuff, but I was not super impressed. They had a booth at DAC this year.

Here are a list of common model problems:

http://www.effectiveelectrons.com/foundrymodels.html

Read the list down under the foundry model audit stuff.
That is a list I developed from one pass at a single CMOS (it was 0.13 ) foundry.

IBM, Jazz, TI, National, ST Micro get their models right, :) the rest of the world can be debated. :-/

Lots of issues as you can see.

Jerry

Title: Re: Model QA
Post by Fred_Lai on Aug 23^rd, 2006, 6:35am

In addition to MQA, there is another well-prove SPICE compact model diagnosis tool after modeling.
Its name is ModeLoo developed by SeveTek Inc. in Taiwan. (www.sevetek.com)

I suggest you could try to contact with them for further information if interested.

Title: Re: Model QA
Post by loose-electron on Aug 23^rd, 2006, 8:39am

Also, if you are on a particular foundry, post it here and people can give feedback on the model quality.

Here is my "poke and sniff" method:

http://www.chipdesignmag.com/display.php?articleId=438&issueId=16

That approach is not rocket science, but more common sense than anything.

Jerry