Ferromagnetic core modeling.

Hello everybody,

I have the need now to simulate ferromagnetic cores. I thought about it and don't recall info in the help file so, I asked google's Gemini AI "Proper method to model a ferromagnetic core with LTspice?". Gemini was kind enough to include a code example.

* Define the material properties
.model my_ferrite LAMINATION(
+ LOSSES=1
+ DC_RES=1e-6
+ THICKNESS=0.5m
+ STACKING=0.95
+ H_C=100
+ B_SAT=1
+ MU_I=1000
)

* Create the inductor
L1 1 0 1uH
.model L1 INDUCTOR (
+ L=1uH
+ IC=0
+ A_L=1e-6
+ R_SER=0.1
+ Rg=1e6
+ PARAM: Lr=1uH
)

* Define the core material for the inductor
.model L1_core my_ferrite

* Excite the circuit (example using a voltage source)
V1 1 0 SIN(0 1 60 0 0)

* Simulate and analyze the results
.tran 0 0.1 0 1u
.end

Would this be a good start to modeling ferromagnetic cores?

#156874

Only if you understand what is going on. Given that most LTspice simulations involve creatin a graphical schematic rather than a text based netlist I'm surprised it came up with an approach that is so "last century". You can do whatever you want, but I would not waste much time with writing netlists.

#156875
Змінено

On Tue, Dec 17, 2024 at 08:03 PM, Richard Andrews wrote:

... I asked google's Gemini AI "Proper method to model a ferromagnetic core with LTspice?". Gemini was kind enough to include a code example.

...

Would this be a good start to modeling ferromagnetic cores?

I rather doubt it. Gemini seems to have gotten it wrong again (no surprise!) because none of that is documented syntax for LTspice. Although there is some syntax that LTspice accepts that is undocumented - usually in an attempt to work with models that were unique to other SPICE-like simulators (PSpice, HSPICE) - I suspect what Gemini came up with is too far off-base and is likely to be rejected. Did you try it? What did LTspice say about it?

I think it's another case of AI er, I mean, AU (Artificial unIntelligence).

To be fair, it did get the voltage source right. ;-)

I recall that HSPICE had magnetic models of this sort which could be described in mechanical terms. Presumably HSPICE used some sort of built-in pre-processor to convert the mechanical description into inductances, much the same as you can get from stand-alone tools or websites that can do the same thing for you.

Andy

#156877

Thanks guys. So, how can I start to simulate cores in LTspice? Is there a way or should I not bother?

#156882

The help DOES tell. Look for Chan Model; that will get you close.

Jim Wagner

Показати цитований текст

On 12/17/2024 11:34 PM PST Richard Andrews via groups.io <richardandrews.ma@...> wrote:

Thanks guys. So, how can I start to simulate cores in LTspice? Is there a way or should I not bother?

#156884

Thanks Jim, that's what I need I believe.

#156885

On 18/12/2024 08:35, Richard Andrews via groups.io wrote:

Thanks guys. So, how can I start to simulate cores in LTspice? Is there a way or should I not bother?

I wouldn't bother asking the AIs, just yet. They have a habit of completely misunderstanding SPICE syntax. I've never found them helpful.

You might consider looking the LTwiki. There is a chapter there on The Chan model, but there is plenty of meat in another chapter on Transformers, much of which is still useful for non-coupled inductors.

Please also don't forget that the group has a huge amount of experience from users that has been uploaded to files. If you just put "Chan model" in the search box, you get 34 hits, not all of which are directly pertinent, but many are. Check them out.

--
Regards,
Tony

#156888

Yes, I found the wiki Thanks to Jim.

#156890

You might also be interested in a book on Inductors by Marcos Alonso. He uses a combination of Python and LTspice to do the modeling.

Amazon.com: Inductors: Variable and Controllable: 9798500248169: Alonso, J. Marcos: Books

#156892

I have repeatedly shown how it is quite simple to model transformers with cores with hysteresis loops. I posted the models on LTwiki. I have many examples of circuit calculation. It surprises me that my achievements are not noticed. I used cores with Chan and Gilius-Atherton models. Also powder cores.

Type in Google: bordodynov LTspice and you will immediately receive the desired link.

#156893

Please don't be discouraged by these repeated appeals for help, Alexander. New members do not know where to search for advice, and some members hope to avoid searching by asking for advice here. That does not devalue your very helpful contributions to our group.

On 2024-12-18 17:09, Александр Бордодынов via groups.io wrote:

I have repeatedly shown how it is quite simple to model transformers with cores with hysteresis loops. I posted the models on LTwiki. I have many examples of circuit calculation. It surprises me that my achievements are not noticed. I used cores with Chan and Gilius-Atherton models. Also powder cores.

Type in Google: bordodynov LTspice and you will immediately receive the desired link.

-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying

Virus-free.www.avg.com

#156894

You would probably find the work by Alan Payne interesting. Add "ferrite" and google

Показати цитований текст

On Wed, Dec 18, 2024, 10:19 AM John Woodgate via groups.io <jmw=woodjohn.uk@groups.io> wrote:

Please don't be discouraged by these repeated appeals for help, Alexander. New members do not know where to search for advice, and some members hope to avoid searching by asking for advice here. That does not devalue your very helpful contributions to our group.

On 2024-12-18 17:09, Александр Бордодынов via groups.io wrote:

I have repeatedly shown how it is quite simple to model transformers with cores with hysteresis loops. I posted the models on LTwiki. I have many examples of circuit calculation. It surprises me that my achievements are not noticed. I used cores with Chan and Gilius-Atherton models. Also powder cores.

Type in Google: bordodynov LTspice and you will immediately receive the desired link.
-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying
Virus-free.www.avg.com

#156896

On 18/12/2024 18:09, Александр Бордодынов via groups.io wrote:

I have repeatedly shown how it is quite simple to model transformers with cores with hysteresis loops. I posted the models on LTwiki. I have many examples of circuit calculation. It surprises me that my achievements are not noticed. I used cores with Chan and Gilius-Atherton models. Also powder cores.

Type in Google: bordodynov LTspice and you will immediately receive the desired link.

Please don't be discouraged, Alexander. A good number of us recognise your expertise and willingness to help. Unfortunately, many users often don't know where to look for help. (I'm being a little generous here. I should stop.)

--

Regards,
Tony

#156897

Yes sir, I downloaded your TRANSFORMers.ZIP last night.

#156961

Interestingly, today I am reading "The Chan inductor does not currently directly support Mutual Inductance, so unless only a single-winding inductor is being modeled, multiple windings must be added via additional circuitry. The simplest way to do this is to construct an ideal, unity coupled transformer with as many windings as required and then put the Chan inductor directly in parallel with any one of the windings (taking care to set the turns to match that particular winding). Since the magnetic effects (including non-saturated inductance) are already modeled by the Chan inductor, the inductance of the transformer's parallel winding must be enough larger (>>10) not to significantly load the Chan inductance. As before, the inductances for the other windings should be scaled by the square of their individual turns ratios (with respect to the paralleled winding). ".

#156973

I am reading here. "A is the bead magnetic cross section in square meters, use dimensions C*(A-B)/2, area is in mm2 hence the u".

So, is the area in meters, micrometers, or millimeters? hence the u?

Seasons Greetings.

#156974

In square mm. 1 mm=1e-3 m, 1 sq mm = 1e-6 sq m

Показати цитований текст

From: LTspice@groups.io <LTspice@groups.io> On Behalf Of Richard Andrews via groups.io
Sent: Sunday, December 22, 2024 2:17 PM
To: LTspice@groups.io
Subject: EXTERNAL: Re: [LTspice] Ferromagnetic core modeling.

I am reading here. "A is the bead magnetic cross section in square meters, use dimensions C*(A-B)/2, area is in mm2 hence the u".

So, is the area in meters, micrometers, or millimeters? hence the u?

Seasons Greetings.

#156975

Thanks Dave.

#156980

You can usually write 'µ' in email and elsewhere using ALT-0181 on the numeric keypad.

On 2024-12-22 22:51, Richard Andrews via groups.io wrote:

Thanks Dave.

-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying

Virus-free.www.avg.com

#156986

I am a Linux user John.

Ctrl + shift+u for one second then the Unicode in hexadecimal.