Not clear to me what you are looking for, but here are some pointers.
- Methodology, that is, process for extracting capacitance -- I recommend going to SupportNet for the newer documentation. You could try "Getting Started: Parasitic Extraction Using Calibre Batch Mode" in the Calibre xRC User's Manual. In the 2008 versions, the xRC instructions assume batch mode.
- How xRC calculates capacitance -- based on mode, try the "Types of Extraction" chapter. (What gets looked at depends on whether you are going transistor level, gate level, etc.) For what xRC is considering based on "C only" versus "RC" or "RCC", look at the "Producing Parasitic Models" chapter.
Another thing to keep in mind that what you get depends on your rule file or Calibre Interactive-PEX setup. For example, you may be extracting only certain nets, or excluding specific nets. Are you set up to include floating nets, like metal fill? That is specified there, too, along with a bunch of other things. (The Calibre xRC manual covers some of the possibilities in chapters 5, 6, and 7.) For the 2008 version, chances are good that your foundry-provided rule file was also still set up to ignore devices for calculating parasitics. (Device parasitics would be part of the device models.)
Thanks for your answer but this is not what i am looking for. I am looking of the bases of the methodology used for solving the Maxwell equations for electromagnitic wave used by Calibre. I just need the base aspects, what kind of fast filed solving method is used theoreticaly. I found some articles in the net about this but i need to connect it to the calibre xrc. I am trying to use the old calibre tm format to create custom profiles for MEMS technologies. But i need a theoretical discription of the field solve equation. This part has to be just a quick discription of the methodolgy. Thanks!
Ah, a question I can answer with confidence! Calibre xRC does not directly use Maxwell's or other equations; the field solver comes into play in xCalibrate, which the foundry uses to generate the PEX rules. (xCalibrate is the part that uses .tm files; xRC has no way to read them in.) At least in older versions, you could specify any field solver to use as part of the flow. Raphael came standard, but there was still a field for setting it to any other one you happened to have.
The new xCalibrate uses the MIPT format instead of .tm files. There may still be a .tm to .mipt translator in the utilities directory. I don't believe it has kept up with the updates to the MIPT format, but it should get you started. As the person who used to maintain the xCalibrate docs, I can vouch that the MIPT format is _much_ easier to set up profiles in!
The Calibre suite does offer a field solver product, xACT 3D. It is pretty recent and I only know the most general things about it. There may be some details on how it works in the Calibre xACT User's Manual, but for anything more in-depth you'll need to speak to Claudia Relyea.
I know that xCalibrate does not solve exactly the Maxwell equation.Actually this is exactly what i need, how it calculates the cap . A bit of mistake form my side when refering to xRC sorry. I suppose that the methodolgy is a Mentor Graphic secret but i don't need it in detalis as i mentioned. Just theoreticaly.
I tryed to use the mipt files for creating these profiles. But in the old version of Calibre, which i have in the university the supprot for mipt is not good. The rule files fail to xcalibrate.
For xACT 3D i will chech the document you point, thanks for that. Just one question, how to conntact Claudia Relyea?
I checked with Claudia for you, since you mention being with a university. (Normally you'd speak to an AE who could arrange the meeting.) She says there really isn't much more information we can provide at this time; the parts that aren't published are confidential, as they constitute part of our competitive advantage.
As someone who used to write manuals for xCalibrate, I'd suggest looking at the current versions on SupportNet for the best information. There were sections I had to remove that took a few years to replace with correct information; the 2008 version of the manuals may not have had the new topics. Here is some possibly helpful information that answers the original question:
"The xCalibrate Rule File Generator has both calibration models and effect models. In the xCalibrate transcript, references to "model 114" or "model 2024" refer to the calibration models. They are used with the technology description to create small layouts for the field solver. Results from the field solver are then analyzed by a curve fitter to determine coefficients for the capacitance and resistance equations."
(from "xCalibrate Models" in the "Parasitic Effects and Calibre Tools" appendix of the xRC manual)
As for how to substitute different field solvers, see the xCalibrate Batch User's Manual -- you need to set up configuration files that include a FIELD_SOLVER value. Do keep in mind that in 2008 we were only officially supporting MCS and MCS2. "Officially supported" means basically that we take the time to rigorously test the results and will correct any bugs found.
Great, i chech this appendix and this is acutally excatly what i was searching for. Just one questin left, can i refer to it in my texts? Also is it a problem to use the last version documentation?
The particular blurb I quoted is true for all versions of xCalibrate. (xRC may have been updated to stop mentioning models; I know it was on the to-do list.) The main thing that keeps changing is the complexity of the models and created rules; the overall flow has been the same from the first version. There is a slightly more explicit rendition of it in the 2008 version of the xCalibrate Interactive manual; look for table 3-1, "Calibration Process."
I believe Mentor Graphics does not object to having documentation that ships with Calibre cited, using standard reference styles and all that. (For future readers, we DO object to wholesale copying, or quoting parts of manuals for beta/specially licensed functionality.) I am neither a lawyer nor a legal representative of Mentor Graphics, however, and if this is intended for a journal you may want to have someone official give an okay.
Good luck, and glad I was able to answer your questions!
Thank you for your answers. It was very helpful.
Best of luck.