High-power powdered iron core

[n85]

Hi,

Does anyone know where one can obtain a relatively large powdered iron core (preferably toroidal for EMI reasons and preferrably pretty cheap)? Everything I find on ebay is too small according to my limited knowledge/calculations and as far as I know buying single pieces in rather uncommon sizes is very difficult from manufacturers.

Also, does anyone know how dependable the calculation programs provided by manufacturers are? I have tried a program from micrometals and it suggests 5-inch or bigger toroids!!!

Thanks,
N

 

[IRstuff]

To do what, exactly?

TTFN

FAQ731-376

 

[n85]

A (possibly foolish and overly ambitious) attempt to build a one-off high-current constant current source using the buck topology.

By high-current I mean electroplating/anodizing/TIG-welding currents (hence the above adjectives). I know these applications tolerate very high ripple currents but reducing ripple should help with EMI, hence the wish for a core with good energy storage. This automatically implies larger volume since there is a limit to how much energy you store per unit volume before saturation.

As I'm inexperienced with magnetics I'd rather overkill than flop.

N

 

[zappedagain]

try Amidon (

https://www.amidoncorp.com/)

 

[Comcokid]

For powered iron cores. the Micrometals software is easy to use and they have plenty of cores at good prices. You also need some knowledge of what the design goal is of many of these software tools in order to manipulate them to your best design advantage - this requires knowledge of the calculation steps involved in inductor design.

Given what you've indicated, you may be arriving at 5" toroids because the window area of most toroids is not adequate for the number of turns required for the inductance you want. Observe what your fill factor result is - for toroids a fill factor of 0.4 is the maximum allowed. Once the fill factor approaches that number the software will select a larger core. This 0.4 limit is typical for toroids as a automated winder is assumed (and clearance necessary for a winding shuttle). You can manually wind beyond this some, but the software doesn't know that.

Also, don overlook stacking cores, or using more than one inductor in series if it works for your design.

You can try to get Micrometals to sample you a core, or you can contact one of their distributors. Many core distributors inventories can be checked online to see if they have the core in stock - always the best option. Distributors have a minimum quantity buy and don't want to acquire something in their inventory unless a customer is going to use it in volume.

Also check surplus places like Nebraska surplus sales if you only need a few large cores - but you will have to design manually with what they have.

 

[Warpspeed]

One thing you may be overlooking, is that the buck topology can sometimes be designed for quite low ripple current at a low switching frequency, and a laminated iron core may be entirely feasible. But not always !

I have done this successfully at the multi kilowatt level with standard iron transformer laminations.
Just remember that predominantly dc current produces a predominantly dc field, and iron works just fine at dc.

If both the ac volts per turn and switching frequency are conservative, eddy current heating in the iron will not become a significant problem.
Likewise skin effect in the wire will not be a problem either.

If your requirement is low switched input voltage, and very high dc current, you may be able to get away with it.

 

[n85]

Warpspeed, you have caught my interest!

Yes, I'd like to design with low input voltage, very high output current. Inductance would have to be huge though to keep ripple small! (also, how small is small and how big is big???)

What I will try to build is a "box" to take input from a common MIG welding supply (rectified AC, peaks in the 60-80V range) and give me a controlled DC current source for TIG/plating. Once the arc is started TIG should need output between 12 and 35V (at least that's what the display on the welders at work says). Plating should take less voltage.

My initial thought was to work in the many tens of kHz range, above audible frequency, to keep the physical inductor size manageable, hence the intention to use iron powder.

Do you have any good data/reading material about this kind of work?

N

 

[Warpspeed]

Design your choke in the conventional way, just as if it were to operate directly from a rectified mains source, at say 100/120Hz ripple frequency at the required operating voltage and current.

Now if for example, you run it at say 10/12Khz the ac ripple component will be one hundred times lower, and the eddy current losses in the core, and skin effect losses in the wire will be low as well.

This obviously does not work when designing a transformer!
But it is very effective with chokes for buck regulators running at low voltages and frequencies.

I once built a buck converter to reduce 48 volts to 12 volts for an obscure solar/wind project, where the main battery was 48 volts, and there were some very high current 12 volt loads. I experimented with switching frequencies between 1Khz and 20Khz with the same choke, and to my surprise it worked just fine anywhere in that frequency range.

This is rather off the wall engineering, and will probably horrify the purists, but it is a practical solution. That is, if the idea of a huge heavy ugly lump of iron laminations does not horrify you.

Another trick with buck regulators is that running two in parallel (180 degrees out of phase) greatly reduces the high current pulsing at the input.
It costs you nothing in extra transistors, heat sinking, laminations, or copper. But it adds some redundancy and ripple reduction.

 

[sreid]

http://www.micrometals.com/

 

[DHambley]

You may also want to look at the Magnetics Inc. MPP catalog. The "how to" sections clearly lead you through inductor designs. I have bought several large MPP toroids on ebay for experimental work. Once in a while, I see one which is close to what I need.

 

[ScottyUK]

Ferroxcube's handbook is worth a read. They have a good range of U and I sections for building up really big cores.

http://www.ferroxcube.com/

http://www.ferroxcube.com/appl/info/HB2009.pdf

----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[n85]

Thanks to everybody for the suggestions.

Unfortunately this project is currently on hold, but if there are any developments I will post feedback.

N