Buck converter inductor wire gauge

[sunil21]

when winding a buck converter inductor , is it important to keep an eye on skin effect and use wire gauge like in a flyback.

 

[Easy peasy]

If you can keep the net AC current ( therefore flux ) down to < 5% you can use solid wire quite happily ... as the small amount of AC flows in the skin - and the rest in the bulk ...

 

[sunil21]

This is ridley works demo. could anyone please explain the difference between the layers and parallel wires above.

 

[Easy peasy]

it appears to say 4 wires used in parallel, 5 layers to be used

--- Updated Jan 17, 2022 ---

2 turns per layer if the 10 turns total is correct ...

 

[sunil21]

it appears to say 4 wires used in parallel, 5 layers to be used

--- Updated Jan 17, 2022 ---

2 turns per layer if the 10 turns total is correct ...

It is somewhat confusing for me, that is why i asked the question.
kindly correct me if i am wrong
above said eqn: awg 19 wire, 4 in parallel , 5 layers.
so i use 19 guage wire around this core wind 10 turns (or 12 - to get needed length per wire)
then i cut 4x5 = 20nos 19awg wire as measured above .
then take 4 nos together and wind on this core.
then another 4 and wind 10 turns so on ... (one over the other - total 5 layers.)
join both ends.
.so.. It is like = 19awg , 20 in parallel , wind 10 turns. ?

 

[Easy peasy]

work out the MLT ( mean length of turn ) and thence the total length for the wdg

then work out the R needed for DC loss of 3.9 watts ( I^2 R )

then work out the basic wire size needed for this R = 17E-9 ( copper ) L / Area

from this area, divide by the area of 19AWG, this gives you the total number of wires in parallel to do the 10 turns with

check this will fit in the winding area available ...!

 

[sunil21]

To keep things simple ,, how many circular mills needed for 30amp. if i use 25awg wire how many to use in parallel (I have limited options) .
If use a single torroid , what size is recommended from your experience.

 

[BradtheRad]

To keep things simple ,, how many circular mills needed for 30amp. if i use 25awg wire how many to use in parallel (I have limited options) .
If use a single torroid , what size is recommended from your experience.

Websites with tables of wire gauges and Amp-carrying capability:

amasci.com/tesla/wire1.html

powerstream.com/Wire_Size.htm

 

[Warpspeed]

Your best bet might be to use some of the free online design software such as Micrometals.
https://micrometals.com/design-and-applications/design-tools/

Feed in all your buck regulator parameters, and you get to choose one of the solutions offered, depending on your preference for temperature rise, size, cost, etc...

 

[sunil21]

My requirement is related to solar. - IN - 45v or 90v (series or parallel - at present 45v - 2 in parallel).. output 12-14
say 15-20 amps only. But thought about going with a 30amp design

I went through micro metals site and saw options. They all use a minimum OD=26mm cores and next one OD=38mm torroids.
what I have with me is a single OD=26mm, and couple of T36 (OD=32-33mm) - both iron powder cores but with no datasheets

I went through the Ridley buck design software. what I can use according to the software(non torroidal) are E,PQ and ETD cores - with gaps.
I have few = EE42, ETD39, ETD49, PQ2625, PQ3230 (all epcos with datasheets - n87).
other than that few torroids
R25.3x14.8x10 - T35
R31.5x19x12.5 - N41
R16.9x9.6x6.3 - N87

Is it possible to use any of the last three torroids for this.?

 

[Warpspeed]

T36 is very high permeability ferrite, its not powdered iron. So forget that.
https://www.tdk-electronics.tdk.com/download/528842/227d492866d3a7fe37a1d2329cd0beba/pdf-t36.pdf

The gapped ferrite cores you mention should work, if they are large enough for your needs. You will have to do a trial design either worked out from the published data, or by actual winding and testing.

The process is, work out what wire gauge will carry the required current without bursting into flames.

Estimate how many turns of that wire will physically fit onto the core (totally full).
In the core data, somewhere there should be an I^2L curve of LI squared plotted against air gap for various sized cores.
Its either going to come up with a sensible air gap, or you will be way off the sheet altogether if the core is way too large or small.

Here is some data for various Philips cores (scroll down to pages 22 to 29).
https://www.thierry-lequeu.fr/data/SFAPPL.pdf
Once you gap a core, the gap becomes far more important than the material.
So for a given core geometry, the physical shape and size and gap are going to give fairly similar results regardless of manufacturer or material.
Material choice is more important for estimating core loss, which will be minimal in a choke anyway.
But for rough initial sizing, the above I^2L curves should get you quite close.

 

[BradtheRad]

output 12-14 say 15-20 amps only. But thought about going with a 30amp design

Whatever amount you select as average current, the inductor needs to have a saturation rating greater than that. Ampere waveforms going through your inductor
have peaks and troughs ranging from 10 to 100 percent ripple.

The input filter should be designed to draw continual maximum available power from the PV panel. Then the filter feeds current as a waveform of peaks and troughs, to the buck converter.

 

[sunil21]

I will go with trial and error method. starting from base and go up to find the limit.

the torroids i mentioned above, especially:
R25.3x14.8x10 - T35
R31.5x19x12.5 - N41

I bought this from a shop (this was only available). I have 10 nos with me.
how do we gap this torroids.
if we stack multiple cores and leave a gap in-between will be of any use?

 

[FvM]

You can gap it with a diamond or boron carbide cutting disc. Also Ferroxcube has a few off-the-shelf gapped toroids.

Beyond this option, don't use ferrite toroid core for storage inductors or flyback transformers. Powder core torroids have built-in distributed airgap and can be used, but useful frequency range is limited due to relative large core losses.

 

[Warpspeed]

I went through micro metals site and saw options. They all use a minimum OD=26mm cores and next one OD=38mm torroids.

So what is the problem ?
Just buy one of the specific toroids listed in their solution, and job done.
These often come up on e-bay, or do a google search.

 

[sunil21]

So what is the problem ?
Just buy one of the specific toroids listed in their solution, and job done.
These often come up on e-bay, or do a google search.

I will try to get this ..
The ferrite torroids i mentioned above was bought couple of years back when I was buying E cores for a flyback. I wanted to use it in a buck (I didnt know much about Iron powder cores etc.. then). I just wanted to know if can use it for something , i dont want to throw it away and I cant use it for flyback, as gapping it seems not that easy- from the last reply.

 

[Warpspeed]

Ferrite is extremely hard and brittle, very difficult to cut, and there is no way of knowing how wide the gap needs to be cut anyway.
Far better to get the exact powdered iron toroid that Micrometals recommend in their software design package. They are not that expensive, and it will do exactly what Micrometals say it will do.