Discussion:
[EE] FEMM users here?
Van Horn, David
2018-04-13 17:31:08 UTC
Permalink
I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".


--
David VanHorn
Lead Hardware Engineer

Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
email: ***@backcountryaccess.com<mailto:***@backcountryaccess.com>
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Van Horn, David
2018-04-18 15:24:56 UTC
Permalink
Solved the problem, it turns out that if you have multiple circuits and the currents in both are set to zero, the solver loops.

Nobody else here into FEMM? Surprising. It's pretty capable, and once you get your head into it, the LUA scripting makes it much easier to create your geometry parametrically. PrimaryTurns = 3, PrimarySpacing = 3*mm, that sort of thing.

-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Van Horn, David
Sent: Friday, April 13, 2018 11:31 AM
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: [EE] FEMM users here?

I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".


--
David VanHorn
Lead Hardware Engineer

Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
email: ***@backcountryaccess.com<mailto:***@backcountryaccess.com>

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Jean-Paul Louis
2018-04-18 15:46:59 UTC
Permalink
I quit using FEMM when I moved away from Windows. If there was an OS X or
Linux version, I would reconsider.

I was using it to model the magnetics for high power DC/DC switchers.

Jean Paul

On Wed, Apr 18, 2018, 11:25 AM Van Horn, David <
Post by Van Horn, David
Solved the problem, it turns out that if you have multiple circuits and
the currents in both are set to zero, the solver loops.
Nobody else here into FEMM? Surprising. It's pretty capable, and once
you get your head into it, the LUA scripting makes it much easier to create
your geometry parametrically. PrimaryTurns = 3, PrimarySpacing = 3*mm,
that sort of thing.
-----Original Message-----
Sent: Friday, April 13, 2018 11:31 AM
Subject: [EE] FEMM users here?
I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".
--
David VanHorn
Lead Hardware Engineer
Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Van Horn, David
2018-04-18 19:46:36 UTC
Permalink
http://www.femm.info/wiki/linuxsupport

-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Jean-Paul Louis
Sent: Wednesday, April 18, 2018 9:47 AM
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: Re: [EE] FEMM users here?

I quit using FEMM when I moved away from Windows. If there was an OS X or Linux version, I would reconsider.

I was using it to model the magnetics for high power DC/DC switchers.

Jean Paul
Post by Van Horn, David
Solved the problem, it turns out that if you have multiple circuits
and the currents in both are set to zero, the solver loops.
Nobody else here into FEMM? Surprising. It's pretty capable, and
once you get your head into it, the LUA scripting makes it much easier
to create your geometry parametrically. PrimaryTurns = 3,
PrimarySpacing = 3*mm, that sort of thing.
-----Original Message-----
Sent: Friday, April 13, 2018 11:31 AM
Subject: [EE] FEMM users here?
I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".
--
David VanHorn
Lead Hardware Engineer
Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Jean-Paul Louis
2018-04-18 19:56:02 UTC
Permalink
Thanks David, I will look more closely. XFEMM might do the job for me in a
Linux VM as my native system is OS X.

On Wed, Apr 18, 2018, 3:47 PM Van Horn, David <
Post by Van Horn, David
http://www.femm.info/wiki/linuxsupport
-----Original Message-----
Sent: Wednesday, April 18, 2018 9:47 AM
Subject: Re: [EE] FEMM users here?
I quit using FEMM when I moved away from Windows. If there was an OS X or
Linux version, I would reconsider.
I was using it to model the magnetics for high power DC/DC switchers.
Jean Paul
On Wed, Apr 18, 2018, 11:25 AM Van Horn, David <
Post by Van Horn, David
Solved the problem, it turns out that if you have multiple circuits
and the currents in both are set to zero, the solver loops.
Nobody else here into FEMM? Surprising. It's pretty capable, and
once you get your head into it, the LUA scripting makes it much easier
to create your geometry parametrically. PrimaryTurns = 3,
PrimarySpacing = 3*mm, that sort of thing.
-----Original Message-----
Sent: Friday, April 13, 2018 11:31 AM
Subject: [EE] FEMM users here?
I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".
--
David VanHorn
Lead Hardware Engineer
Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Andries Tip
2018-04-19 06:27:26 UTC
Permalink
Using FEMM here, but was not able to answer your question...

Does anyone have experience using FEMM for calculating Eddy Current losses? I'm trying to determine those losses in a simple linear motor, much like a pancake type loudspeaker.

Can anyone point me to some examples?

--
Andries

-----Oorspronkelijk bericht-----
Van: piclist-***@mit.edu <piclist-***@mit.edu> Namens Van Horn, David
Verzonden: woensdag 18 april 2018 17:25
Aan: Microcontroller discussion list - Public. <***@mit.edu>
Onderwerp: RE: [EE] FEMM users here?

Solved the problem, it turns out that if you have multiple circuits and the currents in both are set to zero, the solver loops.

Nobody else here into FEMM? Surprising. It's pretty capable, and once you get your head into it, the LUA scripting makes it much easier to create your geometry parametrically. PrimaryTurns = 3, PrimarySpacing = 3*mm, that sort of thing.

-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Van Horn, David
Sent: Friday, April 13, 2018 11:31 AM
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: [EE] FEMM users here?

I'm having trouble with a LUA script that gets hung in analyze forever.
"Matrix construction" seems to run forever.
When I stop, I get a message "problem loading mesh".


--
David VanHorn
Lead Hardware Engineer

Backcountry Access, Inc.
2820 Wilderness Pl, Unit H
Boulder, CO 80301 USA
phone: 303-417-1345 x110
email: ***@backcountryaccess.com<mailto:***@backcountryaccess.com>

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
RussellMc
2018-04-19 08:07:27 UTC
Permalink
Post by Andries Tip
Using FEMM here, but was not able to answer your question...
Does anyone have experience using FEMM for calculating Eddy Current
losses? I'm trying to determine those losses in a simple linear motor, much
like a pancake type loudspeaker.
Can anyone point me to some examples?
​​
There are many what look to the inexperienced eye like apposite examples
using Garglabet search for

fmm eddy current linear motor *here
<https://www.google.co.nz/search?num=50&ei=IDjYWs39EsGI8wWm1a6gDA&q=femm+eddy+current+linear+motor&oq=femm+eddy+current+linear+motor&gs_l=psy-ab.3..33i22i29i30k1.28120.31697.0.31910.13.12.0.0.0.0.291.1561.2-6.6.0....0...1c.1.64.psy-ab..7.6.1556...0j0i22i30k1.0.H7jf4yeiJZE>*

or just

femm eddy current *here
<https://www.google.co.nz/search?q=femm+eddy+current&oq=femm+eddy+current&aqs=chrome..69i57j0l2.6470j1j4&sourceid=chrome&ie=UTF-8>*

_________________

Many of these run in small (or other) circles but may be useful.

eg

http://www.femm.info/wiki/EddyCurrentBrake

http://www.femm.info/wiki/TubeExample

http://www.femm.info/examples/eddycurrentbrake/eddycurrentbrake.pdf

*Rotating Losses in a Outrunner Doubly Salient Permanent Magnet Generator*

http://www.femm.info/wiki/corelosscalculation


*​​Rotating Losses in a Surface Mount Permanent Magnet Motor*

http://www.femm.info/wiki/SPMLoss

FEMM magnetics examples:

http://www.femm.info/wiki/examples

________________________

This is a CV, but he knows your answers :-)
His papers list looks worth a look from a general EE perspective.

http://www.femm.info/wiki/davidmeeker


eg http://www.femm.info/dmeeker/pdf/PID4189713.pdf
Series-Wound Heteropolar Inductor Motor for Automotive Applications

____________________________________

Related only but interesting -

Rotary-linear motors!!!!

Modeling PM Rotary-Linear Motors with TwinStator Using 3D FEMM
https://digitalcommons.lsu.edu/cgi/viewcontent.cgi?article=3518&context=
gradschool_theses

LSU 2010 masters thesis.



--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mi
Andries Tip
2018-04-20 11:33:20 UTC
Permalink
Thanks Russell, for the links. The examples are a bit more complicated than I had hoped for, but they put me on the right track.

-Andries


-----Oorspronkelijk bericht-----
Van: piclist-***@mit.edu <piclist-***@mit.edu> Namens RussellMc
Verzonden: donderdag 19 april 2018 10:07
Aan: Microcontroller discussion list - Public. <***@mit.edu>
Onderwerp: Re: [EE] FEMM users here?
Post by Andries Tip
Using FEMM here, but was not able to answer your question...
Does anyone have experience using FEMM for calculating Eddy Current
losses? I'm trying to determine those losses in a simple linear motor,
much like a pancake type loudspeaker.
Can anyone point me to some examples?
​​
There are many what look to the inexperienced eye like apposite examples using Garglabet search for

fmm eddy current linear motor *here
<https://www.google.co.nz/search?num=50&ei=IDjYWs39EsGI8wWm1a6gDA&q=femm+eddy+current+linear+motor&oq=femm+eddy+current+linear+motor&gs_l=psy-ab.3..33i22i29i30k1.28120.31697.0.31910.13.12.0.0.0.0.291.1561.2-6.6.0....0...1c.1.64.psy-ab..7.6.1556...0j0i22i30k1.0.H7jf4yeiJZE>*

or just

femm eddy current *here
<https://www.google.co.nz/search?q=femm+eddy+current&oq=femm+eddy+current&aqs=chrome..69i57j0l2.6470j1j4&sourceid=chrome&ie=UTF-8>*

_________________

Many of these run in small (or other) circles but may be useful.

eg

http://www.femm.info/wiki/EddyCurrentBrake

http://www.femm.info/wiki/TubeExample

http://www.femm.info/examples/eddycurrentbrake/eddycurrentbrake.pdf

*Rotating Losses in a Outrunner Doubly Salient Permanent Magnet Generator*

http://www.femm.info/wiki/corelosscalculation


*​​Rotating Losses in a Surface Mount Permanent Magnet Motor*

http://www.femm.info/wiki/SPMLoss

FEMM magnetics examples:

http://www.femm.info/wiki/examples

________________________

This is a CV, but he knows your answers :-) His papers list looks worth a look from a general EE perspective.

http://www.femm.info/wiki/davidmeeker


eg http://www.femm.info/dmeeker/pdf/PID4189713.pdf
Series-Wound Heteropolar Inductor Motor for Automotive Applications

____________________________________

Related only but interesting -

Rotary-linear motors!!!!

Modeling PM Rotary-Linear Motors with TwinStator Using 3D FEMM https://digitalcommons.lsu.edu/cgi/viewcontent.cgi?article=3518&context=
gradschool_theses

LSU 2010 masters thesis.



--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo
Van Horn, David
2018-04-20 16:24:01 UTC
Permalink
It's a bit interesting to get into.
I've been waiting two weeks now to get access to the yahoo group, but not holding my breath.
To me, the LUA scripting is the most useful. I'm looking at modeling individual turns on a ferrite inductor with 30 to 60 turns in the primary, six in the secondary, and looking at core loss, eddy current loss etc.

The scripting examples I have seen are not very clear and I don't think they are good teaching examples.

Here's how I create an object in LUA. I set my sim to millimeters and then define:



mm = 1
Inch = 25.4 -- so that I can dimension in native units like "CoreLen = (6 * Inch)" etc.

FRN and FRP are "FerriteRadiusNegative" (0 in the special case of the ferrite)
FLN and FLP are "FerriteLengthNegative" and "FerriteLengthPositive"
These describe the corners of the rectangle which will become the ferrite.

-- ASSIGN THE GROUPS -----------------------------------------------------------------------------------------------------
AirGroup = 0 -- Group 0 is air
FerriteGroup = 1 -- Group 1 is ferrite core
-- and so on.. for all your objects. Nothing special about the group numbers, could be 33 or 7.. It's just a tag.

-- Materials ---------------------------------------------------------------------------------------------------------------
FerriteMaterial = "Soft magnetic ferrite (Fe-Ni-Zn-V)" -- adjust as needed.
--FerriteMaterial = "Pure Iron" -- adjust as needed.
-- These names are pulled from the materials library
- This section just assigns "FerriteMaterial" the proper string value for use later.
----------------------------------------------------------------------------------------------------------------------------

-- Library materials used in this simulation -------------------------------------------------------------------------------
mi_getmaterial('Air')
mi_getmaterial(FerriteMaterial) -- Assigned to a library material or custom material above
--mi_modifymaterial can be used to change the BH curve etc.
----------------------------------------------------------------------------------------------------------------------------------

-- Create the ferrite core -------------------------------------------------------------------------------------------------
-- Define this first because the air size and label position is dependent on this
FerriteLen = (66 * mm)
FerriteDia = 12
FerriteRad = (FerriteDia/2)
FRN = 0 -- Ferrite Radius Negative (because the ferrite is on the axis of symmetry!)
FRP = FerriteRad -- Ferrite Radius Positive
FLN = (-1 * (FerriteLen/2)) -- Ferrite Length Negative
FLP = ( 1 * (FerriteLen/2)) -- Ferrite Length Positive
---------------------------------------------------------------------------------------------------------------------------

-- Define the ferrite core (Group 1) ---------------------------------------------------------------------------------
if (EnableFerriteCore > 0) then
print("Creating Ferrite Core") -- output to console for debug

-- Define where the label will go, insuring it's always inside the geometry.
FLAX =(FRP/2) -- Special case, because the ferrite is on the axis of symmetry
FLAY =(FLP/2) --
-- The label has to be inside the block, wherever you position the block.

-- Create the nodes for the corners
mi_addnode(FRN,FLP)
mi_addnode(FRP,FLP)
mi_addnode(FRP,FLN)
mi_addnode(FRN,FLN)

-- Select those nodes
mi_clearselected()
mi_selectnode(FRN,FLP)
mi_selectnode(FRP,FLP)
mi_selectnode(FRP,FLN)
mi_selectnode(FRN,FLN)
mi_setnodeprop('',FerriteGroup)

-- Everything needs a group number so you can manipulate as a group

-- Add the segments between the nodes
mi_addsegment (FRN,FLP,FRP,FLP) -- top
mi_addsegment (FRP,FLP,FRP,FLN) -- right
mi_addsegment (FRP,FLN,FRN,FLN) -- bottom
mi_addsegment (FRN,FLN,FRN,FLP) -- left

-- Now instead of just corner points we have corners and lines.

-- Set the segment properties
mi_clearselected()
mi_selectsegment (FRN,FLP,FRP,FLP)
mi_selectsegment (FRP,FLP,FRP,FLN)
mi_selectsegment (FRP,FLN,FRN,FLN)
mi_selectsegment (FRN,FLN,FRN,FLP)
mi_setsegmentprop('',0,0,0,FerriteGroup)

-- Now the lines are part of the group and FEMM will see this as an object

-- Create a label
mi_addblocklabel(FLAX,FLAY)
mi_clearselected()
mi_selectlabel(FLAX,FLAY)

mi_setblockprop(FerriteMaterial,AutoMeshCore,MeshSizeCore,0,0,FerriteGroup,0)
-- FerriteMaterial is a variable and could be "Air" to turn off the ferrite, or "Pure Iron" or any other material
-- in the library or that you care to define.
-- AutoMeshCore and MeshSizeCore are variables I added to control the meshing.

end -- EnableFerriteCore

mi_clearselected()
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
Loading...