Hi , I am currently designing a boost converter but I am having trouble finding info on toroid core selection. It has to a be a toroid so it cand be another shape core. If anyone has any useful sources about this subject your help would be appreciated.

 

Have you tried the coil manuf. web sites?
Most of them will answer tech questions and send samples, FREE!
Coilcraft etc.

 

If you know the cross-sectional area, the peak flux density, and the frequency, then you can calculate the number of turns.
The core spec will give you the core loss at the frequency and flux density you are using.
You can calculate the winding resistance and from that, the copper loss.
Where you get stuck is working out the temperature rise.

 

It depends on what your end-goals are.
If You are trying to cram ~100-Watts of Power through a 0.250" diameter Core,
forget about it.
It probably can be done, but it would take several years of very expensive R&D research,
and would probably last no more than about ~10-seconds before going up in smoke.

If miniaturization is not a concern for You, then it's "relatively" easy.
Bigger is virtually always better/easier/more efficient.

Higher-Frequency of operation allows for a reduction in Core size,
but creates other problems which can introduce ridiculous levels of complexity,
depending on just how far You want to push the envelope.

There is no "cookie-cutter" formula for miniaturization of magnetics,
it's heavily loaded with very significant compromises and limitations.

If size is not a concern to You, please say so.
I've collected a lot of general guidelines that may be helpful for DIY Hobbyist type projects.
.
.
.

 

It depends on what your end-goals are.
If You are trying to cram ~100-Watts of Power through a 0.250" diameter Core,
forget about it.
It probably can be done, but it would take several years of very expensive R&D research,
and would probably last no more than about ~10-seconds before going up in smoke.

If miniaturization is not a concern for You, then it's "relatively" easy.
Bigger is virtually always better/easier/more efficient.

Higher-Frequency of operation allows for a reduction in Core size,
but creates other problems which can introduce ridiculous levels of complexity,
depending on just how far You want to push the envelope.

There is no "cookie-cutter" formula for miniaturization of magnetics,
it's heavily loaded with very significant compromises and limitations.

If size is not a concern to You, please say so.
I've collected a lot of general guidelines that may be helpful for DIY Hobbyist type projects.
.
.
.

Thank you for your reply. To be precise my circuit is a PFC circuit for an eV. The circuit takes 230vac and outputs 400VDC and provides 1200 W of power. I need to design the boost inductor in accordance with these values. The circuit uses a UCC28019A pfc controller and the circuit is very similar to the example circuit provided in said controller's datasheet. What I am having trouble with is the selection of the core seeing as how I am required to select a toroidal core. I dont know how to go about selecting the core. I know that I preferably need a powdered metal core or a ferrite core. But I cant figure out the permeability that I need. Also the size of the toroid isnt a problem and can have an inner diameter of 0.95" and above. Figuring out the dimensions of the core is also a bit of a mystery to me. I know its a lot to ask for but you help is greatly appreciated.

 

Thank you for your reply. To be precise my circuit is a PFC circuit for an eV. The circuit takes 230vac and outputs 400VDC and provides 1200 W of power. I need to design the boost inductor in accordance with these values. The circuit uses a UCC28019A pfc controller and the circuit is very similar to the example circuit provided in said controller's datasheet. What I am having trouble with is the selection of the core seeing as how I am required to select a toroidal core. I dont know how to go about selecting the core. I know that I preferably need a powdered metal core or a ferrite core. But I cant figure out the permeability that I need. Also the size of the toroid isnt a problem and can have an inner diameter of 0.95" and above. Figuring out the dimensions of the core is also a bit of a mystery to me. I know its a lot to ask for but you help is greatly appreciated.

In that case, go to Micrometals.com and let their website design it for you.

 

In that case, go to Micrometals.com and let their website design it for you.

If I was designing the circuit for my own purposes thats what I would do but I need to do the calculations myself.

 

If I was designing the circuit for my own purposes thats what I would do but I need to do the calculations myself.

Do you have the figures for the core loss vs.frequency and core loss vs. flux density for the cores you intend to use?
Do you have the thermal conductivity to ambient?

 

Do you have the figures for the core loss vs.frequency and core loss vs. flux density for the cores you intend to use?
Do you have the thermal conductivity to ambient?

Thats the thing I dont know what core to use! I am sorry if I am being a bit annoying but I just need to figure out what kind of core I have to use. The switching frequency is a=60khz

 

Thats the thing I dont know what core to use! I am sorry if I am being a bit annoying but I just need to figure out what kind of core I have to use. The switching frequency is a=60khz

Either a gapped ferrite (toroids are rather tricky to gap) or a high-frequency iron-powder (52 grade, for instance) or sendust or molypermalloy (which might be pricey).
With iron powder, you have the added complication that the inductance reduces gradually as the dc current increases - it happens on all cores, but on ferrite it remains reasonably steady then drops abruptly as you approach saturation,

 

Either a gapped ferrite (toroids are rather tricky to gap) or a high-frequency iron-powder (52 grade, for instance) or sendust or molypermalloy (which might be pricey).
With iron powder, you have the added complication that the inductance reduces gradually as the dc current increases - it happens on all cores, but on ferrite it remains reasonably steady then drops abruptly as you approach saturation,

Yes I have chosen to use a sendust core for their price and availability in my country. How do I go on about selecting the size and of the core though? I would expect there to be formulas to calculate the required minimum permeabilty. Again I am talking about formulas because I need to calculate these values myself. Thanks again.

 

1 step. Decide the input voltage. For example 220V 50 Hz what gives 310 V after rectifier. Apply the formula turns per Volt =50/A (cm2). Make correction divide to 1000 because of 60 kHz and *1.1 due to meandric signal form and multiply to 4 because ferrite have 0.25 Teslas instead of 1.0 T. Thus, the 1x1 inch will make a 2.54x2.54 cm2=7.75 turns per Volt at normal network or 0.034 turns per Volt at 60 kHz and ferrite, demanding for primary 11 turns.

Apply the 3 A/mm2 to estimate the demanded cross section. For example, want 10 A (3.1 kW). Then cross section have to be slight over the 3 mm2. But 60 kHz cannot go inside more than 0.3 mm at 60 kHz, thus the wire cannot be thicker than 0.60 mm having 78.5%*0.6^2=0.283 mm2. Consequently, the wire count in the litz wire must be 0.6x11 strands.

Now, guess the size of this litz wire 3x4 strands or 1.8x2.4 mm=4.4 mm2 and 11 turns will occupy the 50 mm2. Add the carcass, insulation, airgaps and get 100 mm2. That must pass with the half of window area, other half is the secondary bobbin. Thus You need the 1x1 inch cross section ferrite with 200 mm2 window area and get the 3 kW.

PS some very advanced ferrites have B(max) 0.36 T instead of suggested 0.25 T. Then may use a slight smaller ferrite however better to not be gready, let it be better too large than too small.

PPS - if the task is not 3 kW but 30 kW 300 kW or 3000 kW, may use the I-shaped ferrite bricks from US-India producer Cosmoferrites (0.36 T, size 1``x4``, price about 1 Eur per piece, nearest whole-seller at Czech). It may be stacked in every immaginable core form and size and be glued with epoxy or cyanoacrillate.

PPPS - If any DC co-current is existant, then formula for air gap calculus. Flux fringing is caused by the fact that the reluctance of the concentrated air gap is much greater than that of the core. McLyman suggested formula F=1+l(gap)/sqrt(A)*ln(2*l(window)/l(gap)) [51]. Kazimierczuk [52] offered formula F=(1+l(gap)⋅(a+b+2⋅l(gap)))/a⋅b. Hurley and Wolfie [53] said better is F=(a+lgap)⋅(b+lgap)/a/b.

53) Colonel William T. McLyman, Transformer and Inductor Design Handbook, CRC Press, 2004, ISBN 9780824751159, section "Fringing flux" in chapter 8

51) Marian K. Kazimierczuk, High-Frequency Magnetic Components, John Wiley & Sons, 2011, ISBN 9781119964919, p. 38

52) W.G. Hurley, W.H. Wölfle, Transformers and Inductors for Power Electronics: Theory, Design and Applications, John Wiley & Sons, 2013, ISBN 9781118544679

 

Yes I have chosen to use a sendust core for their price and availability in my country. How do I go on about selecting the size and of the core though? I would expect there to be formulas to calculate the required minimum permeabilty. Again I am talking about formulas because I need to calculate these values myself. Thanks again.

Unfortunately, you don’t calculate which core is suitable, you guess which might be suitable then calculate to see whether it is.
If it isn’t, you guess again and recalculate, until you find a suitable candidate.
Sendust cores are available in several permeability ratings.
You calculate the number of turns to give the required inductance
L=P.n^2 where P is the permeance of the core ( the figure usually given in nH/turn^2)
then you calculate the magnetomotive force (amps x turns)
then you calculate the flux φ=F/R
where F= magnetomotive force and R= reluctance
(reluctance is the reciprocal of permeance)
then divide the flux by the cross sectional area to give the flux density.
Is it more than the maximum? If so try again with a bigger core.

Eventually, you find a core which might work.
Calculate the length of wire, and work out its resistance, and the heat it will dissipate (The copper loss)
Then if you know the core loss parameters you can add the core loss to the copper loss.
Then if you know the thermal resistance of the core, you can work out how hot it will get.
if it’s too hot, try another core.

 

Thanks for all your help everyone. I've figured it out now.