Toroid core

[Pixelx]

Hello,

How to calculate the estimated power of a toroidal ferrite core?

For cores such as EE, ETD, the winding window dimension is needed. However, how to calculate the estimated power for a toroidal core taking into account the current density.

 

[KlausST]

Hi,

most core manufacturers provide tutorials, design notes, selection guides.
I´d visit their internet site.

Klaus

 

[Pixelx]

Could you send such tutorials from manufacturers?

Best regards

 

[BradtheRad]

StanCor and CoilCraft are reputable brands for inductors and transformers.

In 'olden' times designers consulted charts listing numerous core sizes and shapes, with comparisons as to current-carrying ability and other parameters.

--- Updated Sep 28, 2024 ---

Note: An inductor develops a flux field which is governed more by current than power (V x A).

Volt level is not a chief parameter. Energy in a flux is calculated as:

webers =A x Henry value.

 

[D.A.(Tony)Stewart]

You compute power in a toroid as you would for any non-linear inductor, and the value assigned is when L reduces by 10% at some temp and rms current level.

This current level guarantees non-saturation if the temperature is allowed to stabilize.

The choice of magnetic material and core size/gap will have frequency-dependent losses that determine the heat dissipation.
In gapless cores the energy is stored between magnetic particles in the ceramic dielectric while gapped cores store energy in the air gap with a reduced L value.

https://tinyurl.com/23aa58c4 examples with constant sine 100Vpp Vin variable f, L but constant f*L with the ~same Pd as 50 Ohms.
Repeat with square waves https://tinyurl.com/25fsgtz5

Heat spreaders and forced air cooling enhance power magnetics somewhat in pwr supply enclosures. I have used thermistor (/w epoxy) to sense the main power transformer to control the unit's fan voltage by digital or analog (40~100%) methods from 50 to 60'C.

 

[cupoftea]

https://www.mag-inc.com/Media/Magnetics/File-Library/Product%20Literature/Powder%20Core%20Literature/Magnetics-Powder-Core-Catalog-2024.pdf?ext=.pdf

..THE ABOVE DOC TELLS HOW TO CALCULATE (sorry about capitals) the power for an inductor wound on their torroids.
Give us a shout if you want me to send you the excel calc which is derived from this doc.
Also, they have a calculator on their website which lets you design a inductor on their torroids and it will then calc the losses for you.

 

[Pixelx]

You can send the calculator, I'd love to see it. But I would also like to be able to calculate it manually so that I can understand how the calculator works.

For EE ETD cores, I calculate it using the following formula where I take into account the window size and current density, as well as the coefficient for the converter. Have you had a chance to use this pattern and how reliable is it?

https://www.mag-inc.com/Media/Magnetics/File-Library/Product Literature/Powder Core Literature/Magnetics-Powder-Core-Catalog-2024.pdf?ext=.pdf

I didn't see the formula for maximum power in the document you sent, but I saw the core sleection chapter, but I didn't see a pattern like the one I showed below.
The question is how to understand the window dimension as just space or as a window filled with the entire winding. This is needed to determine WaAc and how to do it in toroidize? Is this the internal surface area in a toroid?
I also have a core from a welder and out of curiosity I want to check how powerful it is. The core is actively cooled, so the power will be higher, but for the passive one, at least counting and estimating the power will be good.

Could we, all of us interested in the topic, manually calculate the sample power for a toroid?
I can take the characteristics of any core and determine what exactly I have. But for this I would need a template such as the one shown below in the attachments


How can I check in practice whether I can actually obtain such power? Wind two windings and provide a variable signal in time, and on the secondary side dissipate power on, for example, a resistor with a specific power? After exceeding the core power, the voltage will probably drop?

 

[cupoftea]

Please see Post #2 of this...

Magnetic permeability

Hi, In my free time I research magnetic materials. I can determine magnetic hysteresis loops and the core magnetization curve. The loop I designate is practically the same as in the core documentation. I have a problem with various formulas and possibilities of determining magnetic...

www.edaboard.com

...so the above shows you how to calc the core loss...and the winding loss is i^2.R but take into account any skin effect.

--- Updated Sep 28, 2024 ---

In order to help you more it is needed to know the following information...
Qu 1...do you understand that to calculate the core loss in an inductor you need to calc the peak magnetic flux density.......then get the graph of power_loss/volume against peak flux density (for your switching frequency) , then read off the core loss from that?

...EDIT...ok, just checked the doc that i linked above.....they actually give the equation relating Core loss in mW/cm^3 to B in page 107....but its on different pages for different types of core material.......it looks like they have changed the equation from the excel sheet that i did.

....Must admit i think the working they show in their doc is a bot overcomplicated...its more intuitive to calc peak B by simply summing up the accumulated B's that you get from the disrete integration....ie step the current, calc the new H, then calc the increase in B from the previous calculation point......then repeat ...then get that B, then sum up all the B's to get B total.

 

[Pixelx]

need to calc the peak magnetic flux density

Is it about determining H at which Bmax occurs? If so, I can easily obtain this information for any material

then get the graph of power_loss/volume

Yes, I can calculate kW/m3 and watt losses for the loop.


If you can, send an excel file with the power calculations for the toroidal core.

So how do you estimate the power for a toroidal core taking into account the current density? Is the formula I sent above correct and how do you interpret WaAc for such a core?

And would we be able to estimate the power of such a core from the welder I sent in the photo?

 

[cupoftea]

Magnetics - Inductor Design

Magnetics offers a number of inductor and transformer design tools and literature to assist engineers in optimizing their Magnetics components. Contact our Applications Engineers with your design questions.

www.mag-inc.com

...this is the torroid inductor design tool...you have to email them for it....it always used to be available on the site, but now it appears you have to send off for it.

The torroid you show in your picture looks like a transformer with pri and sec coils. To know what max power we can get from it, we need to know what is the topology and output power and vin and vout and switching frequency, and what is the torroid part number, so we can see what type of torroid material it is made from. Is it ferrite or powder core.?

 

[Pixelx]

Sure, I'm checking this information, I understand that you want to use this calculator from the Magnetic manufacturer's website?
My question was also whether I can estimate it using the formula I provided above from the Magnetic website and what to do with the WaAc parameter, which is related to the central column and the window size.


I would like to know what your idea is for determining the power of a toroidal core - it can be a general idea first or an example core.
The point is that I have only a toroidal core in front of me, it may even be the large one from the welding machine, I don't know any parameters of the welding machine, I only have the core and I want to determine the power based on this core, assuming some assumptions about frequency, current density, etc. How to approach this problem with a toroid ?

The core appears to be powdery.
In general, how can you tell what type of core it is - ferrite or powder? I can recognize ferrite MnZn or NiZn with a multimeter, but how can I check whether it is a powder or ferrite core?
If I download a hysteresis loop from a powder core, is there any way to recognize that it is powder core?

 

[cupoftea]

Powder core will saturate more gradually than ferrite core......
A ferrite core would generally need to be gapped to handle any decent power.
But these days you get "Integrated gap" ferrite cores, and they are hard to spot.
Sorry for your other qus i have to get back later.

There are Delta B related losses (ie core losses), and there are the winding losses.

Obviously you need a certain inductance.....and a powder cores permeability varies with current....(sometimes the variability only starts above a certain threshold current)...and so you need to check you get your wanted inductance at the peak current, and that you are happy with the possible variation of inductance you may get over your min to max current range.

It may be a non documented Chinese powder core, in which case you will have no way of working out the core loss.
I mean you can blindly assume a certain core loss, then make the inductor and put it through thermal testing, to see if it holds up...but that coudld waste a lot of time.

You could get a high current supply of say 50V, and then make a transistor bridge, with a series blocking cap, then switch the current in a coil round the torroid...and see what the current looks like......reduce the frequency until you see the current slope off into high saturation....you need the blocking cap or else you may get non equal volt.seconds and your flux will walk off in one direction and saturate very quickly..

 

[Pixelx]

Powder core will saturate more gradually than ferrite core......

Does this mean that the magnetization characteristics will be more leaning?

and so you need to check you get your wanted inductance at the peak current, and that you are happy with the possible variation of inductance you may get over your min to max current range.

This is how it can be read from the BH characteristics which we obtain from any core (I can do it)

It may be a non documented Chinese powder core, in which case you will have no way of working out the core loss.

I can also determine and calculate core losses based on the hysteresis loop. This is not a problem. My problem is how to approach the issue on my own and estimate the core power for certain conditions I assume.

I have a core in front of me, let's assume it is ferrite. Dimensions below .
We will calculate the length of the magnetic path and the cross-section.
FullBridge converter type
Frequency 50kHz
Bmax 500mT for 25 degrees Celsius
Wire diameter 2mm
Current density 10mm2

how to calculate core power?

 

[BradtheRad]

I'm thinking of the flux field which a core can produce. Its maximum intensity and what you can extract from it as it collapses. It has something to do with inducing a magnetic field around or inside it so its magnetic particles align in a certain direction. Align all the magnetic particles (or as many as possible), then you have maximum intensity of the flux field. Let the magnetic field collapse and observe how much energy it generates. The only method I can picture is to wrap a wire many turns in the normal fashion.

This device is several steps removed from the air-core inductor. Standard formulas describe textbook air-core behavior. Then there's the ferrite rod or bar core. It attracts and focuses the flux so that wire turns can be lesser ohms and less lengthy, yet not reduce Henry value. Formulas don't easily apply to such inductors. Perhaps we can observe its behavior by activating a strong magnetic field close by (building the flux field), then shutting off the magnet, and measuring current in the wire turns as the flux field collapses.

Even less easily with a toroid. The toroidal inductor takes a rod-shape and bends it in a circle and joins the poles so it has no recognizable N & S. I wouldn't know how to induce a magnetic field around it or in it. As for formulas, a different set come into play, with graphs and parameters several steps removed from air-core inductors. Yet I think we'd say a toroidal inductor is cousin to the air-core type.

 

[Pixelx]

I could use inductance and calculate the energy, but it doesn't really make sense because the core itself doesn't store energy, there's no gap. Energy is transferred from one side to the other automatically, that's physics. So how to approach such a problem. My idea was with the formula above and the WaAc parameter for EE ETD cores, etc. it makes sense, but I don't know about the toroid.

Having a small core, e.g. a toroid made of the same elements, and a much larger toroid also having the same composition, will the magnetic induction B reach higher values? e.g. for a small one 300mT and for a large one it would be 800mT with the same H?

Any suggestions on how to estimate the power, energy of the toroid?

 

[KlausST]

Hi,

because the core itself doesn't store energy

The core surely does store energy.
it´s about 0.5 * I * I * L
(this is physics, too)

With flyback converters exactly THIS enrgy is used.

***
But yes, there are those push-pull converters ... where the energy is "transferred" as you say. Still they store energy in the core.

Klaus

 

[Easy peasy]

How to calculate the estimated power of a toroidal ferrite core?

This is very straight forward - first look at the losses in the core for the freq and flux peak you wish to run - usually kept to 100mW / cm^3 or less,

knowing this and the input voltage ( or output voltage ) you can determine the turns for the pri and sec,

it is then an easy matter to determine the max size of the wire that will fit on the toroid given the turns - obviously too large a diameter wire for the wdgs will not fit,

remember - if the pri is 100 turns say, and the sec is 50 - then the sec wire with be twice the area of the pri ( but not twice the diameter, this is more than twice the area )

Now, knowing the turns and the max wire sizes that will fit - you can work out the Rdc of each winding - ( you can - if you wish - take account of the longer length of the turns as you add more wdg layers to the toroid - and up the wire sizes accordingly as the layers increase ),

knowing the current required - you can calc the I^2R of each winding and add all these up and the core losses - if this is too high - you need to start again with a bigger core.

This is where experience and the type of cooling come into play - a typical power type 2" ( 50mm ) OD toroid optimally wound would dissipate about 3 watts in still air = more with a fan on it.

Happy designing Pixelx

 

[Pixelx]

This is very straight forward - first look at the losses in the core for the freq and flux peak you wish to run - usually kept to 100mW / cm^3 or less,

Losses for the 3F3 core for approximately 400mT are approximately 3000mW/cm^3. On the X axis there is B and above it ^ does this mean the average value of the induction? Because I don't know if I'm reading this chart correctly?

I can measure the losses in mW/cm^3 for the entire hysteresis loop with a meter of my own construction and determine watts from that. Is this a more accurate result? In my opinion, yes, because we see real losses resulting from the remagnetization of the core.

knowing this and the input voltage ( or output voltage ) you can determine the turns for the pri and sec,

Yes, that's right

remember - if the pri is 100 turns say, and the sec is 50 - then the sec wire with be twice the area of the pri ( but not twice the diameter, this is more than twice the area )

Will the secondary with 50 turns take up more space because it will have a larger wire diameter? Is that what you had in mind?

knowing the current required - you can calc the I^2R of each winding and add all these up and the core losses - if this is too high - you need to start again with a bigger core.

Yes, it makes sense knowing the losses on the hysteresis loop + eddy current losses which I am unable to determine. Is it possible to determine eddy losses knowing the wire and hysteresis losses? Is there any pattern to this?



Another question is what are the formulas I provided above that allow you to estimate the core power for different topologies? Should each core shape be approached this way? This approach makes sense in my opinion for different cores


What is it with this energy in the core? In FlyBack, energy is stored in Halfbridge, this energy is transferred from one side to the other, but this field is also variable in time and increases and decreases in the core, and the power will depend on this field and density? The energy can be calculated E = ½ L i^2 and I found such formulas which I added in the attachment. We can calculate the energy W based on magnetomotive force and flux, only these two parameters change over time, so which value will correspond to the energy from the formula 1/2Li^2? Would you need to take the effective value of flux and magnetomotive force?

 

[BradtheRad]

In the flyback...
There is no particular voltage coming out of the secondary as the charging cycle ends. However there is a certain Amperes (or Webers=A x L) which has to go somewhere. This is the power of the inductor. And it's practically invincible to the point of causing high voltage spikes and possibly arcing.

Your secondary has a resistance, so the formula V=A x R applies. Notice the load R plays a part in determining the voltage.

That's in the flyback. Current is supposed to flow in primary at the opposite time it flows in secondary.

In a transformer, current flows in both primary and secondary at the same time.

 

[Easy peasy]

I know it's tricky to read some of the above - but 100 to 200 mW/ cc is the preferred upper limit for losses - you have to pick a Bmax and frequency to fit inside this - as mentioned above.