Magnetic lamination thickness 2

[sfyten]

I am working on the design of an electromagnet that uses silicon iron laminations. I am looking for a formula or chart that will help me determine what the thickness of the laminations should be. The electromagnet will operate at 0-400 hz.

 

[VEBill]

The purpose of laminations is to prevent or reduce unwanted eddy currents. If you Google the following terms, you'll likely find what you need: eddy currents laminations thickness.

http://www.google.com/search?q=eddy+currents+laminations+thickness

 

[CarlPugh]

0.014 inch thick laminations are probably a good place to start. This thickness is frequently used at 60 hertz, so is readily availiable.
Thinner laminations will have lower loss/temperature rise at 400 hertz.
At 400 hertz and say 4 kilogauss there will be audible noise with the noise increasing with higher flux density.
Is audible noise a concern?
In this application, does the flux density decrease with increasing frequency?

 

[sfyten]

This might sound crazy, but we are trying to get a 1.3 Tesla field in the air gap at 400 hz (the core will be "C" shaped, with a .080" air gap).

 

[CarlPugh]

Worked at company where they made a 400 hertz transformer. The major requirement for the transformer was minimum weight.
The transformer weighed about 100 pounds, was forced air cooled, used 0.004 inch thick laminations of silicon steel in C core and operated at 14 kilogauss. When energized, transformer howled like a banchee. (Could hear it from a block away). (If my conversions are correct, 14 kilogauss is 1.4 tesla and one block is about 201 meter)
Transformer was made with three cores, with air ducts between cores for cooling.

 

[IRstuff]

This is cross-posted from thread340-115215

TTFN

 

[sreid]

Take a look at Metglas core materials.

http://www.metglas.com/

 

[UKpete]

sfyten, I'm reposting this from the magnetic engineering forum (I don't know if there are rules against posting the same question simultaneously in two eng-tips forums, but it's best avoided).

I have never seen a specific formula that you can use, however lamination suppliers do supply design tables giving the watts loss per kg (or lb in the US) for a given peak flux density, usually 1.5T (15kg) and frequency.

An example :

http://www.orb.gb.com/

(if you select "brochures" from the products menu, then download the most appropriate brochure, e.g. Non-oriented, fully processed and look on page 25, the table shows the watts/kg at 400Hz for a range of flux densities). If you are in the US, there is a brochure for ASTM grades.

If you know your working flux density (note PEAK values are always used, not RMS) and the total mass of steel, you can calculate the total loss. This assumes the flux density is uniform throughout the core and it is sinusoidal.

In an electromagnet it may be possible to use oriented steel provided the flux is always in the preferred direction in all parts of the core, this is more expensive but has significantly lower loss.

You will need to design for the worst case (i.e. 400Hz), probably needing a 0.35mm grade or equivalent.

 

[CarlPugh]

When the core is assembled, there is insulation and some spacing between the laminations.
Stacking factor depends on different factors, but following is probably correct for most situations.

0.014 inch thick laminations, stacking factor 0.95.
0.004 inch thick laminations, stacking factor 0.90.
Metglas laminations, stacking factor 0.80 (0.80 was give some years ago for metglas laminations)

Core will operate at 1.3 Tesla/Stacking factor

0.014 thick silicon steel laminations probably have excessive loss for this application.
When stacking factor is considered for metglas, the saturation flux density is to low and the core loss is excessive.
?This leaves the best choice being laminations of 0.004 inch thick oriented silicon steel?

 

[EdStainless]

4 mil or 6 mil are probably where you will end up. If you require a lot of operation time at high freq and field then Metglas looks very nice. Th elosses will be significantly lower with the amorphis material. Remember, the noise that you hear is a result of power loss.
When we tested standard lams we only lightly clamped the stacks. When we tested 4 mil and 6 mil at 400Hz we had to bolt them together to a base plate, just for the lab test.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[sfyten]

Would the skin depth at 400 hz be a relevant consideration in determining the proper thickness of laminations?

 

[CarlPugh]

If you use the core loss at 400 hertz, it is not necessary to consider skin depth.

 

[wrong100]

I THOUGHT YOU SAID 0 TO 400 HZ.
THIS IMPLIES DC.
WHILE OTHER PROBLEMS ARISE AT 400 HZ
THAT ARE DEPENDENT ON CURRENT, FC, WIRE SIZE,AND A FEW OTHER THINGS THESE ARE GENERAL MINIMAL AT 400 HZ.
THE UNIT MUST NOT SATURATE AT DC.
YOU SHOULD BE ABLE TO FIND A PROGRAM
TO DO YOUR SPECIFIC NEED ON THE NET.
IF NOT LET ME KNOW. YOU WILL NEED TO KNOW
MANY THINGS, SUCH AS MATERIAL (IRON OR STEEL OR AIR
AND CURRENT AND ACCEPTABLE POWER LOSSES, SIZE OF THE CORE,
ETC).

MAG-INC.COM
METGLAS.COM
MICROMETALS.COM

ARE WEB SITES YOU SHOULD TRY.