VP2450 Spice Model

Jim Thompson

unread,

Apr 27, 2014, 12:27:27 PM4/27/14

to

John Larkin

unread,

Apr 27, 2014, 1:01:42 PM4/27/14

to

On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:

>I'm behaviorally modeling a _system_ that includes the discrete device
>Supertex VP2450.
>
>At the Supertex/Microchip site, it says no Spice model available :-)
>
>Does anyone know of a model?
>
>Thanks!
>
> ...Jim Thompson

Why not just use some similar part that you have a model for? That mosfet is not
at all unusual, and they are not precision parts anyhow.

That's what I usually do for vanilla diodes, transistors, or mosfets: pick
something close that LT Spice already has. Exotica, like phemts or depletion
fets or jfets, justify more exact models.

Or, if you really want a VP2450 model, pick a similar mosfet model and tweak the
params.

--

John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com

Precision electronic instrumentation

Jim Thompson

unread,

Apr 27, 2014, 1:24:59 PM4/27/14

to

On Sun, 27 Apr 2014 10:01:42 -0700, John Larkin
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:

>On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson
><To-Email-Use-Th...@On-My-Web-Site.com> wrote:
>
>>I'm behaviorally modeling a _system_ that includes the discrete device
>>Supertex VP2450.
>>
>>At the Supertex/Microchip site, it says no Spice model available :-)
>>
>>Does anyone know of a model?
>>
>>Thanks!
>>
>> ...Jim Thompson
>
>Why not just use some similar part that you have a model for? That mosfet is not
>at all unusual, and they are not precision parts anyhow.

500V PMOS

>
>That's what I usually do for vanilla diodes, transistors, or mosfets: pick
>something close that LT Spice already has. Exotica, like phemts or depletion
>fets or jfets, justify more exact models.
>
>Or, if you really want a VP2450 model, pick a similar mosfet model and tweak the
>params.

The problem is nailing the GBW product of this monster.

John Larkin

unread,

Apr 27, 2014, 1:49:43 PM4/27/14

to

On Sun, 27 Apr 2014 10:24:59 -0700, Jim Thompson

<To-Email-Use-Th...@On-My-Web-Site.com> wrote:

>On Sun, 27 Apr 2014 10:01:42 -0700, John Larkin
><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>
>>On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson
>><To-Email-Use-Th...@On-My-Web-Site.com> wrote:
>>
>>>I'm behaviorally modeling a _system_ that includes the discrete device
>>>Supertex VP2450.
>>>
>>>At the Supertex/Microchip site, it says no Spice model available :-)
>>>
>>>Does anyone know of a model?
>>>
>>>Thanks!
>>>
>>> ...Jim Thompson
>>
>>Why not just use some similar part that you have a model for? That mosfet is not
>>at all unusual, and they are not precision parts anyhow.
>
>500V PMOS

Spice models of semiconductors rarely include breakdown voltage. LT Spice is
happy to run a 1N4148 at -100 kilovolts.

That part is not avalanche rated, so don't let the drain voltage hit -500.

>
>>
>>That's what I usually do for vanilla diodes, transistors, or mosfets: pick
>>something close that LT Spice already has. Exotica, like phemts or depletion
>>fets or jfets, justify more exact models.
>>
>>Or, if you really want a VP2450 model, pick a similar mosfet model and tweak the
>>params.
>
>The problem is nailing the GBW product of this monster.
>
> ...Jim Thompson

As I said, pick something close. The GBW won't be precise anyhow. Your circuit
should have a lot of tolerance for fet capacitances and such.

Hey, you are supposed to be the model creation guru, so create a model.

Jim Thompson

unread,

Apr 27, 2014, 2:07:39 PM4/27/14

to

On Sun, 27 Apr 2014 10:49:43 -0700, John Larkin

Takes data to make a model. That is lacking, and the customers
measurements are suspect.

Maynard A. Philbrook Jr.

unread,

Apr 27, 2014, 5:05:18 PM4/27/14

to

In article <p9gql9hh9ke81ua3n...@4ax.com>,
jjla...@highNOTlandTHIStechnologyPART.com says...

I suppose one could insert the "BV" parameter in a existing model, that
is, if the model being used has that.

If you use a VDMOS, you can apply the BV Vds(Break down voltage)

Oh well.

Jamie

RobertMacy

unread,

Apr 27, 2014, 10:52:11 PM4/27/14

to

On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson

<To-Email-Use-Th...@on-my-web-site.com> wrote:

I'm behaviorally modeling a _system_ that includes the discrete device
Supertex VP2450.

At the Supertex/Microchip site, it says no Spice model available

Does anyone know of a model?

Thanks!

...Jim Thompson

I haven't found many of Supertex's models to converge well, or even seem
to replicate the part very well. Sometime they'd supply something with
text that 'looked' like a .cir file, Had to check all the default
prarameters and then 'correct' the forms, but that was a few years ago.

Do you have this one? the VN2450.

*
* SEPTEMBER 24, 2001
*
* COMPOSITE MODEL FOR VN2450
*
.SUBCKT VN2450 1 2 3 4
*
* NODE 1 = DRAIN
* NODE 2 = GATE
* NODE 3 = SOURCE
* NODE 4 = BODY
*
MOS1 11 2 3 4 NE_DMOS L=3.0E-06 W=58E-3
JFET 1 3 11 JMOD 1
DBODY 4 1 DMOS
R 1 11 1E+6
*
.MODEL NE_DMOS NMOS
+ LEVEL=3 UO=240 VTO=2.527 NFS=7.448e+11
+ TOX=5e-08 NSUB=3.59e+15 NSS=0 VMAX=1e+06
+ RS=1 RD=1e-06 RSH=5000 CGDO=3.12e-09
+ CGSO=2.5e-09 CGBO=0 CJ=8e-11 CJSW=0
+ MJ=0.5736 MJSW=0.33 IS=6.472e-13 PB=1.2
+ FC=0.5 XJ=1e-05 LD=0 DELTA=0
+ THETA=0 ETA=1e-06 KAPPA=1e-06
*
.MODEL DMOS D
+ IS=12e-12 N=1.218 RS=10
+ BV=500 IBV=1.0e-3 TT=1e-6
+ CJO=60e-12 VJ=1.2 M=0.5736
*
.MODEL JMOD NJF
+ VTO=-10 BETA=8.872e-3 IS=1e-12
+ RD=10 LAMBDA=0
.ENDS

RobertMacy

unread,

Apr 27, 2014, 10:57:41 PM4/27/14

to

On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson
<To-Email-Use-Th...@on-my-web-site.com> wrote:

> I'm behaviorally modeling a _system_ that includes the discrete device
> Supertex VP2450.
>
> At the Supertex/Microchip site, it says no Spice model available :-)
>
> Does anyone know of a model?
>
> Thanks!
>
> ...Jim Thompson

I didn't see it anywhere in the 35MB of Alex Bordodunov's models, but he
might have one. Did you email him?

Maynard A. Philbrook Jr.

unread,

Apr 28, 2014, 8:08:21 PM4/28/14

to

In article <op.xe0aifd22cx0wh@ajm>, robert...@gmail.com says...

I am no model expert, let alone understanding some of the spice
parameters that are not even seen in the data base models, but I came up
with this using supertex PDF file, as vague as it was. Lots of items not
specified in it and some that are, I don't see where to apply them in
spice, because like I said, I ain't no model expert.

I ran this using their test jig and the charts to do some compares.

.model VP2450 VDMOS(pchan Rg= 3 Rd=15 Rs=15 Vto=-2.0 BV=-500 Kp=320
Cgs=190p Is=1e-30 CgdMax= 75p)

There are items missing for the body diode for capacitance etc..

Then I apply it in an inductive test app and seems to look ok.

Oh well.

Jamie

Jamie

Kevin Aylward

unread,

Apr 29, 2014, 3:13:00 PM4/29/14

to

"Maynard A. Philbrook Jr." wrote in message
news:MPG.2dc8b7754...@news.eternal-september.org...

In article <op.xe0aifd22cx0wh@ajm>, robert...@gmail.com says...
>
> On Sun, 27 Apr 2014 09:27:27 -0700, Jim Thompson
> <To-Email-Use-Th...@on-my-web-site.com> wrote:
>
> > I'm behaviorally modeling a _system_ that includes the discrete device
> > Supertex VP2450.
> >
> > At the Supertex/Microchip site, it says no Spice model available :-)
> >
> > Does anyone know of a model?
> >
> > Thanks!
> >
> > ...Jim Thompson
>
>
> I didn't see it anywhere in the 35MB of Alex Bordodunov's models, but he
> might have one. Did you email him?

>I am no model expert, let alone understanding some of the spice
>parameters that are not even seen in the data base models, but I came up
>with this using supertex PDF file, as vague as it was. Lots of items not
>specified in it and some that are, I don't see where to apply them in
>spice, because like I said, I ain't no model expert.
> I ran this using their test jig and the charts to do some compares.

.model VP2450 VDMOS(pchan Rg= 3 Rd=15 Rs=15 Vto=-2.0 BV=-500 Kp=320
Cgs=190p Is=1e-30 CgdMax= 75p)

Ahmm.... Is is a tad small, and may cause convergence problems. A 1uX1u
diode junction will be of the order of < 1e-18, usually

Bigger discrete devices might be hitting 1e-15, or thereabouts

Kevin Aylward B.Sc.
www.kevinaylward.co.uk
www.anasoft.co.uk - SuperSpice

Jim Thompson

unread,

Apr 29, 2014, 5:07:06 PM4/29/14

to

A published model I found had AD, AS=1. Scratched butt, guesstimated
1u, it then simulated OK.

Maynard A. Philbrook Jr.

unread,

Apr 29, 2014, 9:19:19 PM4/29/14

to

In article <zuCdnUA8O7QkZMLO...@bt.com>,
Extractke...@kevinaylward.co.uk says...

True, but like I said, I am not an expert at this but, the PDF for the
component claims a -1.8V forward drop on the body diode, if you can
believe that, unless I misunderstood something..

http://www.supertex.com/pdf/datasheets/VP2450.pdf

So at 100mA 1e-30 31 was hitting close to home after running some sims
and comparing values.

Maybe some one else can get it closer..

Jamie

Tim Williams

unread,

Apr 30, 2014, 2:59:00 AM4/30/14

to

"Maynard A. Philbrook Jr." <jamie_...@charter.net> wrote in message
news:MPG.2dca19a46...@news.eternal-september.org...

> True, but like I said, I am not an expert at this but, the PDF for the
> component claims a -1.8V forward drop on the body diode, if you can
> believe that, unless I misunderstood something..
>
> http://www.supertex.com/pdf/datasheets/VP2450.pdf
>
> So at 100mA 1e-30 31 was hitting close to home after running some sims
> and comparing values.
>
> Maybe some one else can get it closer..

Hmm. Silicon is silicon. It's still going to turn on around 0.6V -- it
would be amazing physics otherwise. That would suggest ESR around
1.2V/100mA = 12 ohms max. I forget if the MOS model includes RD/RS
parameters, but these can be added externally if not. IS can then be
specified based on leakage and lower current behavior (it should be fairly
"ideal diode" under 10mA).

Not much to say about one datapoint unfortunately. And it's a maximum,
not typical, and certainly not plotted. How lazy. Probably a good idea
measuring a few.

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

Kevin Aylward

unread,

Apr 30, 2014, 1:16:29 PM4/30/14

to

"Maynard A. Philbrook Jr." wrote in message

news:MPG.2dca19a46...@news.eternal-september.org...

>
>> .model VP2450 VDMOS(pchan Rg= 3 Rd=15 Rs=15 Vto=-2.0 BV=-500 Kp=320
>> Cgs=190p Is=1e-30 CgdMax= 75p)
>
>> Ahmm.... Is is a tad small, and may cause convergence problems. A 1uX1u
>> diode junction will be of the order of < 1e-18, usually
>
> Bigger discrete devices might be hitting 1e-15, or thereabouts
>True, but like I said, I am not an expert at this but, the PDF for the
>component claims a -1.8V forward drop on the body diode, if you can
>believe that, unless I misunderstood something..

They possibly made the 101 error of not setting N in exp(Vbe/Vt.N) to its
correct value, from its default of 1, or possibly they set the diode
resistance incorrectly.

For example, to model LEDs, N might be 2.5. to get the several volts that it
has.

"Is" is the reverse leakage current of a diode. No way is it 1e-30 amps!

Maynard A. Philbrook Jr.

unread,

Apr 30, 2014, 10:38:16 PM4/30/14

to

In article <crSdnQkJJKt3svzO...@bt.com>,
Extractke...@kevinaylward.co.uk says...

Ok.

.model VP2450 VDMOS(pchan Rg= 3 Rd=5 Rs=9 Vto=-2.0 BV=-500 N=2.3 Kp=
320 Cgs=195p CgdMax=75p)

I think there as a reason where there is no model out there for this.