Transformer Problem Homework: Building a 1:1 Isolation Transformer

[LdMorgan]

Homework Statement

I'm working with a stripped-out microwave oven transformer core. I've split off a stack of laminations 1.25" thick, and cut out the middle leg of the "E" lams.

I want to build a hollow-core 1:1 isolation transformer, and input a slow (about 3 Hz) pulsed DC current through it that will cyclically induct the core to 16,000 gauss.

I am aware that this will be an inherently inefficient transformer. The current and power required do not particularly matter--it's the flux in the core that I want to tinker with.

The sectional area of the core is 0.744 sq. inches, and the Mean Path Length is 12.375 inches. The relative permeability of the core is (probably) 40,000. There is no air gap in the core.

I've been banging on the numbers for almost two weeks and I'm not making much headway, other than learning that transformers are a lot more complicated than they look.

I really need some help figuring out the coils for this critter.

So far I've gotten answers ranging from 12 to 223 turns of 30-ga magnet wire running at 0.13 amps, and I could be off by an order of magnitude, or more. And, somehow, I don't believe that a half a watt is the right amount of power...for a core that weighs about 4 lbs.

Homework Equations

If there are any around, they are definitely in hiding.

The Attempt at a Solution

Tried a lot of different ways, including working up an air core coil and correcting it for an iron core, but none of the answers I've gotten look even remotely reasonable.

 

[LdMorgan]

I think I've made some progress on this.

I have calculated the magnetic reluctance (R) using Hopkin's Law where R = the length of the magnetic circuit / the permeability of the material X the core sectional area.

I found the permeability by multiplying the relative permeability (40,000) by the magnetic constant (4 pi X EE10-7).

Accordingly, R worked out to 19.8943678.

Then I went back to the basic statement of Hopkin's Law that F = the flux X the magnetic reluctance.

For 1.6 tesla, F (the mmf required) worked out to 31.734 ampere-turns.

So...if I wind the coils from 30 ga magnet wire, and limit the working voltage to 0.13 amps, I'll need 245 turns.

Not too bad--I'll have plenty of room on the core.

This problem doesn't seem so horribly difficult now (given that I have things right here). I think much of my previous difficulty was keeping track of the units (CGS vs. SI) and not quite knowing where to start.

If this looks like a bad solve, someone please give me a yell.

Thanks, all!

 

[berkeman]

Looks like you are doing some good work. My only comment would be to check the saturation flux density for the core. 1.6T is mighty high for a ferrous core. Do you know what material the core is made out of? Is it laminated iron (for 60Hz power conversion)? I couldn't tell if it was the power transformer core that you got out of the oven...

Welcome to the PF, BTW.

 

[LdMorgan]

Yes, 1.6 tesla is about the most the core could possibly handle. It is/was the main power transformer from the microwave oven. Laminated and made for 60 Hz. The core was 2 1/2" thick before I split it, and probably handled 12000-15000 watts of power.

Even with low voltage input and half the core mass, I'm a little leery of the amount of power it may load up when almost fully saturated.

From what I've been reading, microwave oven transformers are usually cheap & inefficient because the manufacturer pays for the iron and copper, but the customer pays for the power it uses. So they're made to run over-saturated as a matter of course, and the excess heat (wasted energy) just gets fanned away.

Shoddy & cheap means the iron probably isn't M6 (which saturates at 2.3 tesla) but something somewhere between recycled tuna cans and re-rolled manhole covers.

Whatever the cheapest transformer iron is, that's probably what I'm working with.

But it's still a little early for precision because everything is make-do at this stage. I'm just happy to just have a starting point to work from.

I'm using a little 2D magnetic simulation program called Visimag which is a) free, b) simple, and c) free.

Did I mention the program is free? (Yay free!)

It's a great little program for total nubes like me. I've never done anything with magnetics before. (This is officially a Learning Experience!)

Vizimag has a full-version 30-day free trial, then you can buy it for $39.75. A very reasonable price. in my opinion.

It's available at http://www.vizimag.com/.

I modeled the core, and can get some good hints about it, but the program isn't 3D so I can't really simulate it accurately.

I tried out the numbers I came up with for the coils, and Vizimag tells me I'm off by an order of magnitude. That means I either lost a decimal when I converted the scientific notation, or the program is insufficiently 3D.

(Hmmm--probably both.)

It says I'll need 2450 turns to create the desired 1.6 tesla of flux in the circuit, so I'll probably wind up buying a fluxmeter.

In the meantime, I won't be shorting the coils out across my tongue...

 

[LdMorgan]

Thanks, by the way, It's good to be here. I've been reading some really interesting stuff...

 

[The Electrician]

Shoddy & cheap means the iron probably isn't M6 (which saturates at 2.3 tesla) but something somewhere between recycled tuna cans and re-rolled manhole covers.

M6 doesn't saturate at 2.3 tesla. Only alloys containing cobalt (such as permendur and supermendur) saturate at flux densities over 2 tesla.

 

[LdMorgan]

Ah! Thanks--I'll have to recheck the spec sheet I was looking at.