Learning Forum

[Sai Venkataramanan]

Hi, 

I am simulating the small signal impedance response of custom fabricated toroid by experimentally measuring the permeability and loss tangent sweep at 1MHz to 1GHz. Since I am not using BP curves, I am not getting core loss as-is in the solution data. But fundamentally, loss tangent should also contain a little bit of eddy and hysteresis information. 

Regardless when I calculate EM losses, my values are high (6W) when they are supposed to be in mW. My question is:

1. What is the formula in fields calculator for getting my eddy and hysteresis losses of the core? Is it possible to separate them?
2. Since these are R&D toroids custom-made at our facility, what other properties (like conductivity/resistivty of core, mass/density of core) is vital for loss calculations?
3. If I increase my peak AC current excitation to a relatively large signal (say from 1mA to 200 mA), is there a way to ask Ansys to proportionally increase the loss tangent so as to account for higher core losses due to higher pk-pk AC excitation strengths? (Excitation by impdance analyzers are less than 10 mT, so unless I have a BP curve, core loss at higher excitation window (+/- Hm) seems impossible). 

Thanks for your time. Please feel free to comment even if you don't have the answer to all the questions. 

[HDLI]

Hello Sai,

    1). Ferrite core

     The loss tangent is only for hystersis loss, not for eddy current. You can enable eddy effect and enter conductivity in solid core materail, when the material is not lamination. After inceasing peak current, you may need to change the loss tangent value. However, you could select hystersis loss model to calculate hysteresis loss, and use eddy effect and conductivity to calculate eddy current loss. 

     2). Lamination

     If the materail is the lamination, you need to use electrical steel in core loss model. With the very high frequency, the coeffiient of the eddy current loss is very important. You have to have measured loss data for the core loss calculation, or estimate skin effect on the eddy current of the lamination.

HDLI  

 

[Sai Venkataramanan]

Hi, 

The material I am using is not a lamination and is a high resistivity material (so steel may not be a good description). Since I do not have BP curves, I cannot use core loss models in Maxwell.  

1. From the loss tangent values I have, how to get the hysteresis losses (as in the formula or setting in ANSYS to get it). Integrating EM losses over the core volume yields very high values.
2. If I set eddy effects to core (disable core losses) and add the experimental material resistivity/conductivity value - then how to calculate/retrieve eddy losses in the core (via fields calculator)?
3. Please note that permeability and loss tangent values obtained from impedance analyzer are experimentally verified with VNA based Inductance and AC resistance values. So I don't think my loss tangent values are wrong, and am fairly certain that my setup in ANSYS or formula for getting the hysteresis losses are wrong.  Please let me know if you need more information

[HDLI]

Hello Sai,

     Please search Maxwell help manual to find loss tangent and hysteresis loss's description and formulation.

     First, I suggest you to enable eddy effect to calculate eddy current loss, and please check if you have skin effect. If having, please add skin depth mesh for it. What is resistivity of the core?

     BTW, your calculation is for loss density, not total loss, so this density value is always high.

HDLI

[Sai Venkataramanan]

Hi, 

Core resistivity is basically 1/conductivity which we found to be around 5*10^-5 S/m for our most resistive core. We measured resistivity based on a custom-built setup verified with literature reported experimental values. I have already enabled eddy effects for the winding (Copper).

So per your suggestion:

1. Integrate ohmic losses over the volume of the copper winding to get eddy losses of the winding - then check for skin effect.
2. Add skin effect based meshing if necessary
3. Input my experimental value for conductivity of the core (highly resistive as mentioned above). 
4. Attached is the geometry of my inductor (Core + winding). If I enable eddy losses in the core, and turn off core losses - then probably integrate EM losses over the volume (not divide by the volume (i.e) not calculating density), will that be a good metric? I will search the maxwell help as you suggested but just making sure I understood what you said. Thanks for your time and patience! 

[HDLI]

Hello Sai,

      It seems that the resistivity is extramely high and can be ignored, so we do not need consider skin effect in the core. However, you may need to consider skin depth in wires if the wire size is not small enough, to get correct copper loss.

      For the core loss, please just integrate EM losses over the volume in the original result. If it is not density, does the result make sense? Thanks.

HDLI

      

 

 

[Sai Venkataramanan]

Now the values are lower. However, the hysteresis loss values are much lower than expected. I expected in milli watts but these are in the order of 10^-7. 

Scl : 6.81405940311547E-07
Scl : Integrate(Volume(Cylinder2), EM-Loss).

Cylinder 2 is the toroid. 

That is probably because I am driving a small signal (1 mA) signal through the inductor. Pls let me know your two cents. 

[HDLI]

Hello Sai,

     Yes, it is low. I agreed with your thought or loss tangent value of material property. Thanks.

HDLI

 

[bidding.estimatings]

To calculate eddy and hysteresis losses, you'll need to consider factors like conductivity, permeability, and density of the core material. As for adjusting the loss tangent in Ansys, you can model these effects with a scaling factor, but it’s best to consult with electrical estimating services for more precise calculations related to your custom core materials and excitation levels.