[Magic-dev] Help

Jeff Sondeen sondeen at ISI.EDU
Tue Sep 14 20:51:20 EDT 2004


well, you are right, the
ftp://ftp.isi.edu/pub/sondeen/magic/new/beta/2002a/SCN6M_SUBM.10.light.tech27
techfile, as well as all related techfiles, are missing extraction
factors for m5->mi. 

here is the techfile 'extract' section showing factors for M6->M5 but
nothing for metali.:


  #M6->M5
   overlap (m6,pad)/metal6 (m5,m5c,m345c,m456c,m6c)/metal5   0.350
   sideoverlap (m6,m6c,m456c,mic,pad)/metal6 ~(m6,m6c,m456c,mic,pad)/metal6 (m5,m5c,m345c,m456c,m6c)/metal5   5.700 ~space/mi
   sideoverlap (m5,m5c,m345c,m456c,m6c)/metal5 ~(m5,m5c,m345c,m456c,m6c)/metal5 (m6,m6c,m456c,mic,pad)/metal6   5.700 ~space/mi

  #metal7

  #metali

  #Mi->M5

  #Mi->M6

  #fets



notice that there's NO value for MIM caps published by Mosis, for example in:

 http://www.mosis.org/cgi-bin/cgiwrap/umosis/swp/params/tsmc-018/t44e_mm_non_epi-params.txt

but let's say you believe the factor for Mi->M5 is 1 FF/UM**2 (i don't
really know), then you would first convert from units of FF/UM**2 into
AF/lambda**2, 

so Mi->5 would be .01 FF/lambda**2 = 10 AF/lambda**2

(works since 1FF*1um*1um = 1FF, and .01FF*10lambda*10lambda= 1FF)
(with lambda = .10 um)

so you could add this to the techfile:

#Mi->M5

 overlap (mi,mic)/metali (m5,m5c,m345c)/metal5   10

the enclosed example (notice, not DRC legal since #contacts is way too
small)

which defined a 40x40 lambda Mi box, gives 16.1 fF as follows:

% m mim.mag
magic
tech scmos
timestamp 1095214777
<< m5contact >>
rect -12 -8 -7 -3
<< metal5 >>
rect -8 -3 40 40
rect -7 -8 40 -3
<< metal6 >>
rect -20 10 10 11
rect -20 0 0 10
<< micontact >>
rect 0 0 10 10
<< metali >>
rect -4 10 36 36
rect -4 0 0 10
rect 10 0 36 10
rect -4 -4 36 0
<< labels >>
rlabel metal6 -16 5 -16 5 3 m6
rlabel metal5 -6 -7 -6 -7 1 m5
rlabel metali -3 -3 -3 -3 1 mi
<< end >>

% m mim.ext
timestamp 1095214777
version 7.2
tech scmos
style TSMC0.18um(tsmc18)from:t24i
scale 1000 1 10.000000
resistclasses 6900 7800 946000 1 8000 8000 80 80 80 80 80 40
node "mi" 0 314.76 -20 0 m6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 330 82
equiv "mi" "m6"
node "m5" 0 652.38 -12 -8 v4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 25 20 2324 200 0 0
cap "m5" "mi" 16142.6

% m mim.spice
* SPICE3 file created from mim.ext - technology: scmos

C0 m5 mi 16.1fF
C1 mi GND 0.3fF **FLOATING
C2 m5 GND 0.7fF **FLOATING


however, i believe the TSMC design manuals specify that the MIM cap
should be modeled with a rather large spice subckt that captures all
the parasitics (that magic can't extract).

/jeff




Roohollah Aminroaya writes:
 > Hi
 > who can help me please ?
 >  
 > I want to layout a MIM capacitor in magic tool, in 0.18um TSMC thech. files.
 > I tried it using M5 as bottum plate, Mi as top plate, and contact mi to m6. But the magic tool didn`t extract any capacitance.
 > I will be very thankfull if you help me about how I can do it. (design rules)
 >  
 > Thank you
 > Amin
 > 
 > 
 > 		
 > ---------------------------------
 > Do you Yahoo!?
 > vote.yahoo.com - Register online to vote today!
 > <DIV>Hi</DIV>
 > <DIV>who can help me please ?</DIV>
 > <DIV>&nbsp;</DIV>
 > <DIV>
 > <DIV>I want to layout&nbsp;a MIM capacitor in magic tool,&nbsp;in 0.18um TSMC thech. files.</DIV>
 > <DIV>I tried it&nbsp;using M5 as bottum plate, Mi as top plate, and&nbsp;contact mi to m6. But the magic&nbsp;tool didn`t extract any capacitance.</DIV>
 > <DIV>I will be very thankfull if you help me about how I can do it. (design rules)</DIV>
 > <DIV>&nbsp;</DIV>
 > <DIV>Thank you</DIV>
 > <DIV>Amin</DIV></DIV><p>
 > 		<hr size=1>Do you Yahoo!?<br><a
 > href="http://vote.yahoo.com">vote.yahoo.com</a> - Register online to vote today!
 > _______________________________________________
 > magic-dev mailing list
 > magic-dev at vlsi.cornell.edu
 > http://vlsi.csl.cornell.edu/mailman/listinfo/magic-dev


More information about the magic-dev mailing list