Quick jump to:
Main Page | News Group | Course Notes | Instructor Information | Course Information | Schedule | Projects | Resources

Changes made to the extraction procedure:

"ext2spice" has been modified so that extracted transistors have source/drain areas and perimeters equal to zero.  It is assumed that all of the capacitances are accounted for as lumped capacitors.  The new ext2spice binary file and source code are located in /home/ff/ee141/ext2spice/.

Technology files (mmi25.tech and mmi25.tech27 in /usr//local/pkg/mmi_020213/max/tech/mmi25/) have been updated with the following (located at the end of each file):

extract
style mmi25
    noplaneordering

    cscale 1
    lambda 1
    step 1000
    sidehalo 0

    resist poly 100
    areacap poly 0.000101
    perimc poly space/active 0.000001

    resist ndif 100
    areacap ndif 0.001724
    perimc ndif space/active 0.826

    resist pdif 100
    areacap pdif 0.0019
    perimc pdif space/active 0.345

    resist m1 100
    areacap m1 0.000034
    perimc m1 space/m1 0.007

    resist m2 70
    areacap m2 0.000017
    perimc m2 space/m2 0.047

    resist m3 70
    areacap m3 0.000011
    perimc m3 space/m3 0.041

    resist m4 70
    areacap m4 0.000006
    perimc m4 space/m4 0.028

    resist m5 20
    areacap m5 0.000005
    perimc m5 space/m5 0.000001

    fet nfet ndif 2 nfet GND! 0 0
    fet pfet pdif 2 pfet Vdd! 0 0

end

This sets the resistances and capacitances for the different layers.  A note from the nice folks who put together this hack:

These values are taken from TCMC 0.25u run T25T. Since there is a mix-up with dimensions in max extract, the values are scaled so that they give proper order of magnitudes in the spice file (as calculated by hand: Area of diffusion times areacap for diffusion + perimeter of dif. times perimc + … . (By mistake, resistances of the first four layers is incorrect, should be 4400, 4700, 3500, and 70, respectively (this is in miliohms per square). This is of no importance for the present problem, but will give wrong results for resistance of long wires. The calculation of resistances of long wires was not checked, it may be necessary to correct the order of magnitude.)

- Leland Chang 10/30/02

    Thanks to Aleksandra Pavasovic, University of Belgrade.