Hi,
I'm simulating a digital full-custom design using Spectre (20.1.0.298.isr9) in ADE Assembler (ICADVM20.1-64b.500.21).
The layout is DRC and LVS clean. The extraction was done with Cablire (xACT, xACT 3D, xACT 3D with 200 or 600.. same problem).
The following screenshot shows the waveforms of the pre-layout simuation (schematic view). As there is no pass-transistor gates in the
the design, the output signal S<7:0> shows no "drop" or "ground bounce" as expected.
The following screen shot shows the output signal of the post-layout simulation (calibre view):
This effect don't appear in other (smaller) designs. So I assume that the extraction options are correct.
I tried the following:
* Different errpreset, spectre, APS, with and without ++aps, w and w/o postlayout (hpa, upa,...),etc.
* Different solvers. Trap causes oscillations so I stayed with gear2only for changing the solver didn't change anything.
The only thing that changed anything was:
* Setting rabsshort to 100 "fixed" the problem with the "drop" but the output still doesn't reach 0V. The following screenshot shows the effect:
But most of the extracted resistances are below 100ohm, so this can't be the right way.
The following lines are from the output of the spectre-run using rabsshort=100:
Global user options:
ignorezerovar = yes
psfversion = 1.4.0
vabstol = 1e-06
iabstol = 1e-12
temp = 27
multithread = on
gmin = 1e-12
rforce = 1
rabsshort = 100
vthmod = vthcc
ivthn = 3e-07
ivthp = 7e-08
ivthw = 0
ivthl = 0
maxnotes = 5
maxwarns = 5
digits = 5
cols = 80
dc_pivot_check = yes
pivrel = 0.001
sensfile = ../psf/sens.output
checklimitdest = psf
vdsatmod = gds
save = allpub
reltol = 0.0001
tnom = 27
scalem = 1
scale = 1
Scoped user options:
Circuit inventory:
nodes 36507
bit 16
bsimimg 1332
delta_gate 1102
juncap200 6
vsource 8
capacitor 72470
resistor 34675
Important parameter values:
start = 0 s
outputstart = 0 s
stop = 5 ns
step = 5 ps
maxstep = 50 ps
ic = all
useprevic = no
skipdc = no
reltol = 10e-06
abstol(V) = 1 uV
abstol(I) = 1 pA
temp = 27 C
tnom = 27 C
tempeffects = all
errpreset = conservative_sigglobal
method = gear2only
lteratio = 10
relref = sigglobal
cmin = 0 F
gmin = 1 pS
rabsshort = 100 Ohm
My question is: is this a numerical effect/artifact of spectre or is the problem on the calibre extraction side? If it's the first case, how can I fix it?
Best regards,
Iqbal
(long-time-reader, first-time poster)