Electronics > Projects, Designs, and Technical Stuff
high current DIY inductors
<< < (3/6) > >>
NiHaoMike:

--- Quote from: T3sl4co1l on January 05, 2020, 04:26:17 pm ---Nah, it's fringing through the material itself -- where it's fringing in plane is fine, but through plane is where all your losses happen.

--- End quote ---
Would it help to add some powdered iron pieces between the amorphous cores and the gap?
T3sl4co1l:
Uhhh.... of what permeability?

Replacing air gap with low mu means needing much more gap, which, eh, I'm not sure, it might help.  Effectively it's a wider gap so should fringe even more widely, but the mu will keep some flux in it, less fringes from the base core.  But maybe the increased distance, the wider fringing, turns out worse anyway.

High mu isn't any good, obviously.  Air gap is needed to generate magnetizing inductance.

Or if you mean replacing the core faces with an isotropic material and still using an air gap... I wonder if you're saving much at all by using nanocrystalline.  It's a hell of a lot more expensive than just using powder in the first place.

You're welcome to run it in FEMM, if you can replicate the materials correctly.

Alternately you could wrap the air gap with heavy copper, which forces the flux to flow within the shield rather than fringe out around it (you leave a slit, so it's not a shorted turn).  This is going to cook -- the question is, will it cook as hot as the core alone?  I'm not sure how to calculate losses in either case, so FEMM might be the better method again.

Tim
NiHaoMike:

--- Quote from: T3sl4co1l on January 05, 2020, 06:53:50 pm ---Or if you mean replacing the core faces with an isotropic material and still using an air gap... I wonder if you're saving much at all by using nanocrystalline.  It's a hell of a lot more expensive than just using powder in the first place.

--- End quote ---
So what's the advantage of using amorphous iron over powdered iron in the first place? I was getting the impression that it can run at higher flux levels compared to powdered iron before the losses become excessive, thereby reducing the core size for a lower overall cost. Or do you mean that if I use powdered iron as well as amorphous iron, the powdered iron becomes the bottleneck and the advantages are lost?
MagicSmoker:

--- Quote from: NiHaoMike on January 03, 2020, 04:54:56 am ---...
My requirements, however, are quite a bit more reasonable - on the order of 500uH-1mH at up to 60A or so. (2 of them, one per phase, along with a bunch of much smaller inductors elsewhere in the unit.) The switching frequency is 15-20kHz and size and weight are not that important, low losses and low cost are.
...

--- End quote ---

Magnetics, Inc. Kool-Mu is likely going to have the best combination of cost vs. energy storage vs. losses. Amorphous/Nanocrystalline (Metglas, FineMet, etc.) will give slightly lower losses and require less volume for a given energy rating, but costs a lot more for said energy rating, needs to be discretely gapped for chokes (so, lots of fringing effects and relatively hard saturation), requires a bobbin to hold the windings and banding to hold the cores together. Basically, a lot more cost and labor for a slight reduction in losses/volume.

Powdered iron will be shockingly bad here - high losses and not really all that less expensive than Kool-Mu. Gapped ferrite will have minuscule losses but cost more and require a lot more volume for a given energy rating. Magnetics, Inc. XFlux might be a good alternative, though. I'd double check the losses for the flux swing, but it should still be okay at 15-20kHz and <100mT delta-B.

That said, 1mH * 60A = 1800mJ* which is going to require a massive core area regardless of the material. Magnetics, Inc. makes large E cores out of Kool-Mu which are relatively cost-effective though I have mostly used stacks of toroids. My most recent design was a PFC choke that needed to be 250uH and handle a peak current of 33A at 50kHz. I used a stack of (3) 77736 40u cores with 25t of #8 Litz; each core cost about $10.50, IIRC. Not bad, all things considered, but you need at least 2x the inductance 2x and the current... To paraphrase Jaws, "you're gonna need a bigger boat."


* - or 3600mJ in magnetics-speak.
T3sl4co1l:
I wouldn't say shockingly bad.  The lossy powdered irons (#26, 52) are fine in the low 10s of kHz and low to modest ripple fractions.  The lower loss ones (#8, etc.) are okay at higher frequencies / ripple fractions.  They're cheap, so if you don't mind using a little more space, it's an easy way to store energy.

Kool-mu handles even higher ripple fraction in turn, though not all /that/ much, I've cooked one before...

Tim
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod