<@U016H5X1K62>: `open_pdks` commit `99a5cf99862a2...
# magic
t
@User:
open_pdks
commit
99a5cf99862a2d49a0fea77b0b40a828a8765a16
on opencircuitdesign.com fixes this issue.
y
@Tim Edwards that made a big difference in getting LVS to pass but there were still a couple of other things I needed to do to make it work: 1. In the library spice files the
sky130_fd_sc_hd__diode_2
diode uses the parameters
a
and
p
which don't match the extracted
area
and
perimeter
parameters 2. The
sky130_fd_sc_hd__conb_1
library cell uses subcircuit (X prefix)
sky130_fd_pr__res_generic_po
devices while the extracted devices are resistors (R prefix) I've modified my
sky130_fd_sc_hd.spice
file to match the extracted versions. Is this the best approach to take? If so I can raise an issue and/or pull request detailing it
t
Thanks; I should have noticed the
R
instead of
X
. The area and perimeter difference I have to look into to make sure that the change is the intended one. I just coded it up in python as a regexp string substitution, so fixing the installation to match your changes should be easy enough (plus, it applies them to all the digital standard cell libraries, since I found that the problems occur in all of them).
@yrrapt : The latest ngspice-34 now supports diode level 3, which it didn't when I first started writing the PDK. The level 3 diode should use
area
and
pj
as parameters (it can also use
l
and
w
which is probably preferable). I have not checked that this works.