EDAboard.com | EDAboard.de | EDAboard.co.uk | WTWH Media

Artefact in the John Chad model of a hysteretic inductor in

Ask a question - edaboard.com

elektroda.net NewsGroups Forum Index - EDA CAD Electronics - Artefact in the John Chad model of a hysteretic inductor in

Bill Sloman
Guest

Fri Apr 20, 2012 1:22 pm   



I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"


--
Bill Sloman, Nijmegen

Bill Sloman
Guest

Fri Apr 20, 2012 2:38 pm   



On Apr 20, 1:22 pm, Bill Sloman <eac...@gmail.com> wrote:
Quote:
I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

Sorry. False alarm. The current spikes were coming from another part
of the circuit - the OPA27 has input protection diodes, and the
current was coming from there.

It's embarrassing that it took as long as it did for the penny to
drop.

--
Bill Sloman, Nijmegen

Robert Macy
Guest

Fri Apr 20, 2012 2:47 pm   



On Apr 20, 4:22 am, Bill Sloman <eac...@gmail.com> wrote:
Quote:
I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

--
Bill Sloman, Nijmegen

haven't used this form, but will check.

why not post the .txt form of the .asc file?

Bill Sloman
Guest

Fri Apr 20, 2012 5:12 pm   



On Apr 20, 2:47 pm, Robert Macy <robert.a.m...@gmail.com> wrote:
Quote:
On Apr 20, 4:22 am, Bill Sloman <eac...@gmail.com> wrote:

I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

haven't used this form, but will check.

why not post the .txt form of the .asc file?

Since my much cleverer friend in London didn't recognise the trick I
was using (which I invented back in 1975 and haven't used since then)
it may be patentable, which kind of discourages open publication.

In any event, there doesn't seem to be anything wrong with John Chad
model, and I've fixed the real problem in my simulation, which is not
crunching away very, very slowly - 100psec per second when I last
looked - though it has started to pick up speed. It should go faster
once the circuit settles down into the start where it's doing what I
want it to, but start-up can be a bitch.

--
Bill Sloman, Nijmegen

Robert Macy
Guest

Fri Apr 20, 2012 7:20 pm   



On Apr 20, 8:12 am, Bill Sloman <eac...@gmail.com> wrote:
Quote:
On Apr 20, 2:47 pm, Robert Macy <robert.a.m...@gmail.com> wrote:





On Apr 20, 4:22 am, Bill Sloman <eac...@gmail.com> wrote:

I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

haven't used this form, but will check.

why not post the .txt form of the .asc file?

Since my much cleverer friend in London didn't recognise the trick I
was using (which I invented back in 1975 and haven't used since then)
it may be patentable, which kind of discourages open publication.

In any event, there doesn't seem to be anything wrong with John Chad
model, and I've fixed the real problem in my simulation, which is not
crunching away very, very slowly - 100psec per second when I last
looked - though it has started to pick up speed. It should go faster
once the circuit settles down into the start where it's doing what I
want it to, but start-up can be a bitch.

--
Bill Sloman, Nijmegen

doesn't
..tran 0 1 UIC
..IC n1=1 n2=3 ....
and current through inductors on their line
help?

Jamie
Guest

Sat Apr 21, 2012 12:36 am   



Bill Sloman wrote:

Quote:
On Apr 20, 1:22 pm, Bill Sloman <eac...@gmail.com> wrote:

I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"


Sorry. False alarm. The current spikes were coming from another part
of the circuit - the OPA27 has input protection diodes, and the
current was coming from there.

It's embarrassing that it took as long as it did for the penny to
drop.

--
Bill Sloman, Nijmegen
Why should it be embarrassing ? You should be used to that by now!


Jamie

Bill Sloman
Guest

Sat Apr 21, 2012 9:50 am   



On Apr 20, 7:20 pm, Robert Macy <robert.a.m...@gmail.com> wrote:
Quote:
On Apr 20, 8:12 am,BillSloman<eac...@gmail.com> wrote:









On Apr 20, 2:47 pm, Robert Macy <robert.a.m...@gmail.com> wrote:

On Apr 20, 4:22 am,BillSloman<eac...@gmail.com> wrote:

I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

haven't used this form, but will check.

why not post the .txt form of the .asc file?

Since my much cleverer friend in London didn't recognise the trick I
was using (which I invented back in 1975 and haven't used since then)
it may be patentable, which kind of discourages open publication.

In any event, there doesn't seem to be anything wrong with John Chad
model, and I've fixed the real problem in my simulation, which is not
crunching away very, very slowly - 100psec per second when I last
looked - though it has started to pick up speed. It should go faster
once the circuit settles down into the start where it's doing what I
want it to, but start-up can be a bitch.

doesn't
.tran 0 1  UIC
.IC  n1=1 n2=3 ....
and current through inductors on their line
help?

Less than you'd think. I've tried setting the initial current through
the inductor at 50mA (roughly it's peak value when the circuit settles
down0, and the routine that sets the initial DC conditions reduces the
actual starting current to microamps when I'm using the John Chan
inductor model.

--
Bill Sloman, Nijmegen

Bill Sloman
Guest

Sat Apr 21, 2012 9:51 am   



On Apr 21, 2:36 am, Jamie
<jamie_ka1lpa_not_valid_after_ka1l...@charter.net> wrote:
Quote:
BillSlomanwrote:
On Apr 20, 1:22 pm,BillSloman<eac...@gmail.com> wrote:

I've used this model to simulate a real - gapped - inductor in a test
circuit, and it seems to work.

In a more complete simulation, it seems to inject 1mA current steps
from time to time - my simulated data is littered with 1nC charge
pulses from a nominally 1.44mH simulated hysteretic inductor.

Has anybody got an alternative - better behaved - formulation of the
model or a work-around. Embedding the simulated inductor in a filter
network doesn't seem to help - or not at least for a filter network
that won't impair the fidelity of the simulation.

The inductor I'm simulating is an EPCOS RM14 cores with a 1.9mm air
gap and an Al of 160nH per root turn.

The Spice data is "HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019
N=95"

In the Spice netlist

"L20 N034 N039 HC=20. Bs=.49 Br=.18 A=0.00017 Lm=0.07 Lg=0.0019 N=95"

Sorry. False alarm. The current spikes were coming from another part
of the circuit - the OPA27 has input protection diodes, and the
current was coming from there.

It's embarrassing that it took as long as it did for the penny to
drop.

--
BillSloman, Nijmegen

Why should it be embarrassing ? You should be used to that by now!

I've got delusions of competence.

--
Bill Sloman, Nijmegen

elektroda.net NewsGroups Forum Index - EDA CAD Electronics - Artefact in the John Chad model of a hysteretic inductor in

Ask a question - edaboard.com

Arabic version Bulgarian version Catalan version Czech version Danish version German version Greek version English version Spanish version Finnish version French version Hindi version Croatian version Indonesian version Italian version Hebrew version Japanese version Korean version Lithuanian version Latvian version Dutch version Norwegian version Polish version Portuguese version Romanian version Russian version Slovak version Slovenian version Serbian version Swedish version Tagalog version Ukrainian version Vietnamese version Chinese version Turkish version
EDAboard.com map