Garbage in=garbage out

[Roff]

Nigel may have a field day with this, and I'll predict that anything he says will be well deserved.
This is a sim on SwitcherCAD III. I got the transistor model from On Semiconductor. Notice that the collector voltage goes above the supply voltage when the transistor is on!
Just proves that you need to do a reality check on any sim you do.
I think the garbage is in the model, because I have never seen SwCAD do this before.
Any comments?

 

[bloki]

You got inverted waveform and a peak value greater than 5V. Impossible. Use some other simulation software. SWCAD is supposed to be used prefferably for switcher regulators simulation.

 

[Roff]

I've sent a query on this off to the SwitcherCAD III forum, which has several experts on the software who are very active in monitoring the forum.

Bloki, I've been using it for years (in my hobbyist mode), and it has always worked exceptionally well. I'm not ready to write it off just yet.

 

[audioguru]

SwitcherCad is driving down a completely clear road, when all of a sudden a huge tree jumps out from the side and its car runs straight into it! "No officer, I haven't been drinkin' tonight".

Maybe SwitcherCad hasn't seen a slow old MJ2955 power transistor before and gets confused when its massive c-b capacitance is trying to cancel (integrate?) its base signal.
The sharp edges of the output pulses are a lot faster than I expect from such a slow transistor.

We should have a separate forum where folks post screw-ups from sims.
Nigel and I can visit it whenever we need a good laugh. :lol:

 

[Optikon]

I've sent a query on this off to the SwitcherCAD III forum, which has several experts on the software who are very active in monitoring the forum.

Bloki, I've been using it for years (in my hobbyist mode), and it has always worked exceptionally well. I'm not ready to write it off just yet.

Yeah, I wouldnt right it off yet either. Inccocent until proven guilty. You know, it very well might just be an erroneous model. Not like that's ever happened before!

Can you post the model for the tran? I assume it's a subcircuit and not a "Q" intrinsic model.

I'm interested to know what the SWCAD guys say..

 

[audioguru]

I'm interested to know what the SWCAD guys say..

They're laughing so hard they can't say anything yet. :lol:

 

[Optikon]

I couldnt resist.. tried it in my non SWcad simulator...

Same problem! It's not switchercad, it's the model.

I simulated your same circuit with the onsemi model.

Also noted, all the spice models for TIP2955 from onsemi, pass the same paramaters so it's not like the spice3 is different than the spice versions and so on....If you care enough to bother, you should email Onsemi.

 

[Optikon]

I dont know what the problem is but here is the BAD model:

**************************************
* Model Generated by MODPEX *
*Copyright(c) Symmetry Design Systems*
* All Rights Reserved *
* UNPUBLISHED LICENSED SOFTWARE *
* Contains Proprietary Information *
* Which is The Property of *
* SYMMETRY OR ITS LICENSORS *
* Modeling services provided by *
* Interface Technologies www.i-t.com *
**************************************
.MODEL Qtip2955 pnp
+IS=9.73033e-13 BF=139.321 NF=0.705063 VAF=537.022
+IKF=11.9427 ISE=8.51339e-09 NE=1.71065 BR=13.9321
+NR=1.5 VAR=185.952 IKR=3.70468 ISC=1e-16
+NC=1.31102 RB=20.1463 IRB=0.1 RBM=0.1
+RE=0.0507364 RC=0.253682 XTB=0.1 XTI=3.67928
+EG=1.05 CJE=9.53946e-08 VJE=0.426507 MJE=0.675433
+TF=1e-08 XTF=1.35754 VTF=0.998574 ITF=0.999498
+CJC=4.44708e-10 VJC=0.400409 MJC=0.409494 XCJC=0.803124
+FC=0.71051 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
* Model generated on Feb 8, 2004
* Model format: PSpice

I guess MODPEX is some auto model generator.. well that was a mistake.
Here is a GOOD onsemi '2955 model that simulated correctly:

*TopSPICE library: Models\Vendor\ONSEMI.MDB
*PART NUMBER: QMJD2955
*MODEL NAME: QMJD2955
*SYMBOL: QPNP
*SYNTAX: Qname c b e QMJD2955
*
* ON Semiconductor BJT
*
.MODEL QMJD2955 PNP (
+ IS=2.44387E-010 BF=84.7605 NF=1.03718 VAF=10.0169 IKF=5.96743
+ ISE=6.25674E-012 NE=3.40545 BR=3.10971 NR=1.12389 VAR=2.37417
+ IKR=6.56305 ISC=2.5E-013 NC=3.9375 RB=6.6898 IRB=0.1
+ RBM=0.1 RE=0.000445417 RC=0.0940744 XTB=0.100111 XTI=1.00001
+ EG=1.19324 CJE=1E-007 VJE=0.4 MJE=0.640465 TF=1E-008
+ XTF=1.35714 VTF=0.995088 ITF=1.00012 CJC=5E-010 VJC=0.4
+ MJC=0.411572 XCJC=0.803125 FC=0.8 CJS=0 VJS=0.75
+ MJS=0.5 TR=1E-007 PTF=0 KF=0 AF=1 )

 

[bloki]

QMJD2955 model OK
QTIP2955 model WRONG

 

[Roff]

Thanks for all the good feedback, guys. Special thanks to Optikon for the working model. Can you provide a URL for that?
Here's the reply I got back from the guy at Linear Technology who wrote SwitcherCAD III:

Ron,

It's an error in the model. NF should not be
less than 1. FYI, LTspice gives the same very
strange result as other SPICE programs for
this screwy model, so when you complain to
the concern that supplied this model, tell
them it's not an LTspice problem.

--Mike

 

[Optikon]

Thanks for all the good feedback, guys. Special thanks to Optikon for the working model. Can you provide a URL for that?
Here's the reply I got back from the guy at Linear Technology who wrote SwitcherCAD III:

Ron,

It's an error in the model. NF should not be
less than 1. FYI, LTspice gives the same very
strange result as other SPICE programs for
this screwy model, so when you complain to
the concern that supplied this model, tell
them it's not an LTspice problem.

--Mike

The one I used came with my other spice sim as a built in vendor (ONSEMI) library.

But, searching ONSEMI found this one:

**************************************
* Model Generated by MODPEX *
*Copyright(c) Symmetry Design Systems*
* All Rights Reserved *
* UNPUBLISHED LICENSED SOFTWARE *
* Contains Proprietary Information *
* Which is The Property of *
* SYMMETRY OR ITS LICENSORS *
*Commercial Use or Resale Restricted *
* by Symmetry License Agreement *
**************************************
* Model generated on Jan 18, 02
* MODEL FORMAT: PSpice
.MODEL Qmjd2955 pnp
+IS=1.02735e-10 BF=444.172 NF=1.12616 VAF=20.5632
+IKF=0.326428 ISE=1e-08 NE=2.1144 BR=16.0318
+NR=1.16416 VAR=205.632 IKR=2.65642 ISC=3.1701e-11
+NC=2.96563 RB=3.74917 IRB=0.1 RBM=0.1
+RE=0.0145233 RC=0.0726167 XTB=0.1 XTI=1
+EG=1.05 CJE=1e-13 VJE=0.99 MJE=0.85
+TF=1e-08 XTF=4.28474 VTF=3462.75 ITF=0.00625931
+CJC=5e-10 VJC=0.4 MJC=0.85 XCJC=0.999987
+FC=0.8 CJS=0 VJS=0.75 MJS=0.5
+TR=2.13927e-11 PTF=0 KF=0 AF=1



The URL is:

https://www.onsemi.com/PowerSolutio...955&searchType=others&tabbed=Y&clearFilters=Y


It's another MODEPEX special but at least the NF > 1

I havnt tried it..

 

[audioguru]

The TIP2955 model that works is from Texas Instruments.
The MJD2955 that's good is from Motorola.
The MJD2955 that's bad is from Toshiba.

If you want another good laugh, check Toshiba's datasheet of their copy of National Semi's LM337 regulator.
They have the protection diodes backwards and the English is very funny. :lol:

 

[Optikon]

The TIP2955 model that works is from Texas Instruments.
The MJD2955 that's good is from Motorola.
The MJD2955 that's bad is from Toshiba.

If you want another good laugh, check Toshiba's datasheet of their copy of National Semi's LM337 regulator.
They have the protection diodes backwards and the English is very funny. :lol:

:shock:

Believe it or not, I've seen worse Jinglish.