The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
May 26th, 2019, 7:13am
Pages: 1
Send Topic Print
Discrepancy between spectre/diode and ahdlLib/diode_sch (Read 598 times)
cheap_salary
Senior Member
****
Offline



Posts: 162

Discrepancy between spectre/diode and ahdlLib/diode_sch
Jan 11th, 2019, 5:48am
 
Even if model parameters are all default values, DC characteristics is fairly different between spectre/diode and ahdlLib/diode_sch in Cadence Spectre.

Why can not I get coincidence ?
I think diode equations of Cadence Spectre are different from ahdlLib/diode_sch.

ahdlLib/diode_sch gives same results as http://www.designers-guide.org/VerilogAMS/semiconductors/diode/diode.va

Code:
// Generated for: spectre
// Generated on: Jan 11 14:51:59 2019
// Design library name: My_Tools_TestBenches
// Design cell name: test_Diode
// Design view name: schematic
simulator lang=spectre
global 0

//include "/hoge/Ddummy.scs"
model Ddummy diode compatible=spice2

// Library name: My_Tools_TestBenches
// Cell name: test_Diode
// View name: schematic
Vin (net06 0) vsource dc=0 type=dc
D0 (net06 0) Ddummy
D1 (net06 0) diode_sch
simulatorOptions options psfversion="1.1.0" reltol=1e-3 vabstol=1e-6 \
    iabstol=1n temp=25.0 tnom=25.0 scalem=1.0 scale=1.0 gmin=1e-12 \
    rforce=1 maxnotes=5 maxwarns=5 digits=5 cols=80 pivrel=1e-3 \
    sensfile="../psf/sens.output" bin_relref=yes highvoltage=yes
dcOp dc write="spectre.dc" save=allpub maxiters=150 maxsteps=10000 \
    annotate=status
dcOpInfo info what=oppoint where=rawfile
dc dc dev=Vin param=dc start=-1 stop=1 lin=1001 save=allpub \
    oppoint=rawfile maxiters=150 maxsteps=10000 annotate=status
save D0:a D1:vanode
saveOptions options save=allpub saveahdlvars=all
ahdl_include "/hoge/ic610/IC6.17.20/tools/dfII/samples/artist/ahdlLib/diode_sch/veriloga/veriloga.va" 

Back to top
 

2019-0111-1500-05.png
View Profile   IP Logged
Andrew Beckett
Senior Fellow
******
Offline

Life, don't talk to
me about Life...

Posts: 1708
Bracknell, UK
Re: Discrepancy between spectre/diode and ahdlLib/diode_sch
Reply #1 - Jan 12th, 2019, 7:58am
 
I'm not sure why you would think they should be the same? Even the diode_sch VerilogA model and the diode.va model on this site are different (they have a lot of similarities, but they are different).

You wouldn't expect mos1 to be the same as bsim3v3 to be the same as bsim4, so why should different diode models behave the same way?

Regards,

Andrew.
Back to top
 
 
View Profile WWW   IP Logged
cheap_salary
Senior Member
****
Offline



Posts: 162

Re: Discrepancy between spectre/diode and ahdlLib/diode_sch
Reply #2 - Jan 13th, 2019, 5:39am
 
Andrew Beckett wrote on Jan 12th, 2019, 7:58am:
You wouldn't expect mos1 to be the same as bsim3v3 to be the same as bsim4,
so why should different diode models behave the same way?
Such variations don't exist for PN-Diode and Gummel-Poon-BJT.

There are two options for model selector of PN-Diode in Cadence Spectre at most.
Quote:
The DIODE level-1 model is based on the junction (Berkeley-spice) model
and the level-2 model is based on the Fowler-Nordheim model.

1 level=1 ; Model selector.
 1 = Junction,
 2 = Fowler-Nordheim,
 3 = Junction + additional metal and polysilicon capacitances.

4 compatible=spectre ; Spice compatibility flag.
 Possible values are
 spectre, spice2, spice3, cdsspice, hspice,
 spiceplus, eldo, sspice, mica, tispice, and pspice.


I choose compatible=spice2 and level=1, so diode model has to be same as spice model.

Andrew Beckett wrote on Jan 12th, 2019, 7:58am:
Even the diode_sch VerilogA model and the diode.va model on this site are different (they have a lot of similarities, but they are different).
They give completely same DC characteristics.
Only Spectre diode model give different DC characteristics.

Discrepancies exist for forward small current region anf large current region.

Latter might be due to
"3 bv_enable=1 ; Flag to enable the breakdown of diode.
 1=breakdown effect enabled,
 0=breakdown effect disabled."
Back to top
 
 
View Profile   IP Logged
Geoffrey_Coram
Senior Fellow
******
Offline



Posts: 1959
Massachusetts, USA
Re: Discrepancy between spectre/diode and ahdlLib/diode_sch
Reply #3 - Jan 14th, 2019, 5:20am
 
Look at the documentation for the "dskip" parameter; that may explain the forward low-current region.

For the high-current region, I think some diode models linearize after a certain current (imax?), and the simulator usually prints a warning about it.

I haven't seen bv_enable used; usually, specifying bv (which defaults to infinity) is the way to turn that on. But that would be something at large negative voltage.
Back to top
 
 

If at first you do succeed, STOP, raise your standards, and stop wasting your time.
View Profile WWW   IP Logged
cheap_salary
Senior Member
****
Offline



Posts: 162

Re: Discrepancy between spectre/diode and ahdlLib/diode_sch
Reply #4 - Jan 15th, 2019, 1:20am
 
Synopsys HSPICE gives completely same DC characteristics for ahdlLib/diode_sch and http://www.designers-guide.org/VerilogAMS/semiconductors/diode/diode.va without any special care.
.MODEL Ddummy D


Code:
** Generated for: hspiceD
** Generated on: Jan 15 09:56:43 2019
** Design library name: My_Tools_TestBenches
** Design cell name: test_Diode
** Design view name: schematic


.PROBE DC
+    I(d0) isub(xd1.vanode)

.DC vin LIN 1.001e3 -1.0 1.0

.TEMP 25.0
.OPTION
+    ACCURATE
+    ACOUT=0
+    LIST=3
+    ARTIST=2
+    INGOLD=2
+    INTERP
+    METHOD=GEAR
+    NOMOD
+    PARHIER=LOCAL
+    PROBE
+    PSF=2
+    RUNLVL=6
+    TNOM=25
*.INCLUDE "/hoge/Ddummy.spi"
.MODEL Ddummy D

** Library name: My_Tools_TestBenches
** Cell name: test_Diode
** View name: schematic
vin net06 0 DC=0
d0 net06 0 Ddummy
xd1 net06 0 diode_sch
.hdl "/hoge/ic610/IC6.17.20/tools.lnx86/dfII/samples/artist/ahdlLib/diode_sch/veriloga/veriloga.va"
.END 



Geoffrey_Coram wrote on Jan 14th, 2019, 5:20am:
Look at the documentation for the "dskip" parameter; that may explain the forward low-current region.

For the high-current region, I think some diode models linearize after a certain current (imax?)


On the other hand, Cadence Spectre requires special care, even if I set "compatible=spice2".
model Ddummy diode compatible=spice2 jmax=1000 dskip=no
Back to top
 

2019-0115-0959-04.png
View Profile   IP Logged
cheap_salary
Senior Member
****
Offline



Posts: 162

Re: Discrepancy between spectre/diode and ahdlLib/diode_sch
Reply #5 - Jan 15th, 2019, 1:25am
 
If I define Code:
model Ddummy diode compatible=spice2 jmax=1000 dskip=no 

, Cadence Spectre can give same DC characteristics for ahdlLib/diode_sch and http://www.designers-guide.org/VerilogAMS/semiconductors/diode/diode.va.
Back to top
 

2019-0115-0855-27.png
View Profile   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2019 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.