File: ext2spice.1

package info (click to toggle)
magic 8.1.218+ds.1-1~bpo9+1
  • links: PTS, VCS
  • area: main
  • in suites: stretch-backports
  • size: 16,604 kB
  • sloc: ansic: 160,782; sh: 7,068; tcl: 3,982; lisp: 2,554; makefile: 918; cpp: 587; csh: 149; awk: 140
file content (152 lines) | stat: -rw-r--r-- 6,493 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
.TH EXT2SPICE 1
.UC 4
.SH NAME
ext2spice \- convert hierarchical \fIext\fR\|(5) extracted-circuit files
to flat \fIspice\fR\| files
.SH SYNOPSIS
.B ext2spice
[
.B \-B
] [
.I "extcheck-options"
] [
.I -M|m
] [
.I -y num
] [
.I -f hspice|spice3|spice2
] [
.I -J hier|flat
] [
.I -j device:sdRclass[/subRclass]/defaultSubstrate
]
.I root

.SH DESCRIPTION
Ext2spice will convert an extracted circuit from the hierarchical
\fIext\fR\|(5) representation produced by Magic to a flat spice
file which can be accepted by spice2, spice3, hspice and other
simulation tools.
The root of the tree to be extracted is the file \fIroot\fB.ext\fR;
it and all the files it references are recursively flattened.
The result is a single, flat representation of the circuit that is
written to the file \fIroot\fB.spice\fR .
.LP
The following options are recognized:
.TP 1.0i
.B \-o\ \fIoutfile\fP
Instead of leaving output in the file \fIroot\fB.spice\fR, leave it
in \fIoutfile\fP.
.TP 1.0i
.B \-B
Don't output transistor or node attributes in the spice file.
Usually the attributes of a node or a device are output as
special comments **fetattr and **nodeatrr which can be processed
further to create things such a initial conditions etc.
.TP 1.0i
.B \-F
Don't output nodes that aren't connected to fets (floating nodes).
Normally capacitance from these nodes is output with the comment
**FLOATING attached on the same line.
.TP 1.0i
.B \-t\fIchar\fR
Trim characters from node names when writing the output file.
\fIChar\fR
should be either "#" or "!".
The option may be used twice if both characters
are desired. Trimming "#" and "!" is enabled by default when the format is
hspice.
.TP 1.0i
.B -\fIM|m\fR
Merge parallel fets. \fI-m\fR means conservative merging of fets that have
equal widths only (useful with hspice format multiplier if delta W
effects need to be taken care of). -M means aggressive merging: the fets
are merged if they have the same terminals and the same length.
.TP 1.0i
.B \-y \fInum\fR
Select the precision for outputting capacitors. The default is 1 which means
that the capacitors will be printed to a precision of .1 fF.
.TP 1.0i
.B \-f \fIhspice|spice2|spice3\fR
Select the output format. Spice3 is the the format understood by the
latest version of berkeley spice. Node names have the same names as they
would in a \fIsim\fR(5) file and no special constructs are used.
Spice2 is the format understood by the older version of spice (which
usually has better convergence). Node names are numbers and a dictionary
of number and corresponding node is output in the end.
HSPICE is a format understood by meta-software's hspice and other
commercial tools. In this format node names cannot be longer than 15
characters long (blame the fortran code): so if a hierarchical node name
is longer it is truncated to something like x1234/name where x1234 is
an alias of the normal node hierarchical prefix and name its hierarchical
postfix (a dictionary mapping prefixes to real hierarchical paths is output
at the end of the spice file). If the node name is still longer than 15
characters long (again blame the fortran code) it is translated to something
like z@1234 and the equivalent name is output as a comment. In addition since
hspice supports scaling and multipliers so the output dimensions are in
lambdas and if parallel fets are merged the hspice construct \fIM\fR is used.
.TP 1.0i
.B \-J \fIhier|flat\fR
Select the source/drain area and perimeter extraction algorithm. If
\fIhier\fR is selected then the areas and perimeters are extracted
\fIonly within each subcell\fR. For each fet in a subcell the area
and perimeter of its source and drain within this subcell are output.
If two or more fets share a source/drain node then the total area and
perimeter will be output in only one of them and the other will have 0.
If \fIflat\fR is selected the same rules apply only that the scope of
search for area and perimeter is the whole netlist. In general \fIflat\fR
(which is the default) will give accurate results (it will take into
account shared sources/drains) but hier is provided for backwards
compatibility with version 6.4.5. On top of this selection you can
individually control how a terminal of a specific fet will be extracted
if you put a source/drain attribute. \fIext:aph\fR makes the extraction
for that specific terminal hierarchical and \fIext:apf\fR makes the
extraction flat (see the magic tutorial about attaching attribute labels).
Additionally to ease extraction of bipolar transistors the gate attribute
\fIext:aps\fR forces the output of the substrate area and perimeter for
a specific fet (in flat mode only).
.TP 1.0i
.B \-j  \fIdevice:sdRclass[/subRclass]/defaultSubstrate\fR
Gives ext2sim information about the source/drain resistance class of the
fet type \fIdevice\fR. Makes \fIdevice\fR to have \fIsdRclass\fR source
drain resistance class, \fIsubRclass\fR substrate (well) resistance class
and the node named \fIdefaultSubstrate\fR as its default substrate.
The defaults are nfet:0/Gnd\! and pfet:1/6/Vdd\! which correspond to the
MOSIS technology file but things might vary in your site. Ask your local
cad administrator.

.PP
The way the extraction of node area and perimeter works in magic the total
area and perimeter of the source/drain junction is summed up on a single node.
That is why all the junction areas and perimeters are summed up on a single
node (this should not affect simulation results however).
.PP
\fISpecial care must be taken when the substrate of a fet is tied to a
node other than the default substrate\fR (eg in a bootstraping charge
pump).
To get the correct substrate info in these cases the fet(s) with
separate wells should be in their own separate subcell with ext:aph attributes
attached to their sensitive terminals (also all the transistors which share
sensitive terminals with these should be in another subcell with the same
attributes).


.PP
In addition, all of the options of \fIextcheck\fR\|(1) are accepted.

.PP
The awk filter spice2sim is provided with the current distribution for
debugging purposes.

.SH "SEE ALSO"
extcheck\|(1), ext2spice\|(1),
magic\|(1), rsim\|(1), ext\|(5), sim\|(5)

.SH AUTHOR
Stefanos Sidiropoulos.

.SH BUGS
The areas and perimeters of fet sources and drains work only with the
simple extraction algorithm and not with the extresis flow. So you have
to model them as linear capacitors (create a special extraction style)
if you want to extract parasitic resistances with extresis.