Help determine mu permeability e core w/ simple method l/c sine generator + scope

[obiwan]

Hello friends ! I have many different E , EI , ER, R cores from different manufacturers I bought sometime.
These cores are kept grouped in the same bags and boxes the day I bought them, so no mix up possible.
But they don't have any inscription, code, part no. , model, etc.
I need to determine mu permeability by using a simple method.
Earlier today I took one E 30/15/7 no gap , and calculated an inductor using the Excellent 7300 program from Starichok51

Proposed L = 703 uH
Frequency f = 60KHz
N = 21 turns

The program input/output as follows:

I connected a 0.01 uF known value ceramic capacitor in parallel with the coil, applied a 60 KHz low level sine wave to this circuit and monitored on my scope.
With these values of L and C , the circuit showed perfect resonance at 60 KHz with maximum amplitude of the sine wave.

Then used the formula:

u = L / Lo

... where L is the value of the inductance using the core , Lo is the value of the inductance without the core (air core)
The value of the air core resonance occurred at 954.55 KHz that means L = 2.78 uH , so

u = 703 uH / 2.78 uH = 252.8

I find that this must be wrong. When I bought these cores the material was Epcos N27 and u should be arround 1500.
So I am doing something wrong here, or the formula or method is not correct.

Could you please enlighten me !
Thank you.

 

[cableiso]

I think that something is not correct, although I cannot quite explain what.

I have here a known core with ur=2400, I get (for 10T) 776nH with the coil in free space and 370uH with core. This is just slightly lower Al than listed on the datasheet (3800) so I believe things are working as they should.

However, my own calculation of ur using your method gives ur=477 instead of ur=2400. I then went through the equations and I wonder: is it that the area and path length are quite different in the case of core vs no core? This seems like the likely place for a discrepancy, since the N^2 factor is definitely identical.

The only time I did a permeability estimation was using the "two-winding" BH measurement with a triangle current and integrator.

 

[obiwan]

Thank you cableiso, last week I had to do the same test with also weird results.
A friend brought an U shaped large ferrite core of unknown origin.
We tried to figure out mu, and did the same procedure.
I am using my old TRIO AG-203 oscillator and my Tektronix TDS 210 scope.
Using a good quality polyester 0.1 uF capacitor and 10 turns of some magnet wire.
Again the mu from L1/L2 = (u0*ur) / (u0*1), or L1/L2 = ur and the mu throwed 3 or 4 !!!
We should find out about this issue, I am totally lost.

 

[obiwan]

OK, I think I know what the problem is. I called a friend and he corrected an issue in the procedure.
1. find the resonance of the inductor at a selected frequency using a certain number of turns and a known capacitor, with the core, calculate inductance
2. remove de core and find the resonance FOR THE SAME FREQUENCY by changing the known capacitor value or adding turns to the inductor, calculate the inductance.
3. at all times keep the amplitude of the generator low checking that the resonance does not vary with amplitude.
4. use these values of inductance for calculation mu = ro / r

I will try this tomorrow.

Links of interest:

http://elnaferrite.com/wp-content/u...al/Measuring_Soft_Ferrite_Core_Properties.pdf
http://www.johnbreslin.org/files/publications/19960900_vpec1996.pdf


Regards

 

[Triode]

Does it fit the equation?

L = (N^2 * u0 * ur * A) / l

L = inductance
N = turns
u0 - perm. of free space
ur - relative perm (what you want)
A - Core cross sectional area - get from core D.S.
l - core magnetic length - get from core D. S.

 

[obiwan]

Hi Triode, I really don't know ...
I am looking for a simple method ... just wind 10 or 15 turns of any wire available, a generator and an oscilloscope.
As far as I have consulted with a couple of experienced friends, it can be done by simple dividing the inductance with core and without it (air).
I have to recheck because I think that the problem was that I didn't pay attention to frequency.
Frequency should be the same at both measurements of inductance, so the capacitor value or the number of turns should be adjusted.
In my case, I made the first measurement adjusting for maximum resonance , but in the second stage (air winding) I looked for resonance with the same capacitor, hence the frequency climbed a lot, and the resulting mu was bad.
I will try again with the test and be back with results.
Any ideas ?

 

[Triode]

I think you can just wind 10 turns, plug L, N, A, and l into the equation and solve for permeability.

For a E 30/15/7, the l is 67mm. The A is 60mm2. And u0 is 1.257e-9 H/mm.

So for your case:

ur = (L * l) / (A * N^2 * u0)
ur = (720e-6 * 67) / (60 * 441 * 1.257e-9)
ur = 1450

I estimate your core has permeability of 1450.

But perhaps you can just look at AL value and make a guess
If you got 720u from 21 turns, use
L = AL * T^2
Your AL = 1594.

Since it's nearly impossible to get perfect coupling unless clamped really well it's not unusual to get a value slightly lower than a specification. And plus, the material varies!
Your material looks like 3C90 (1900) or 3C94 (1600) if it's a ferroxcube. It's too low to be 3C81 or 3C91 (both AL=2500). Using the AL method for your guess is probably easier than permeability..

 

[obiwan]

valuable data ! ...but you are starting from the point you know u0 and this is not the case.
We could measure de core, but we have many of them in bags as I explained before, so it is almost impossible to know what kind of material we are talking about.
I am trying to find a method to know mu from scratch.
Do you think it is possible ?

 

[Triode]

We do know u0! u0 is the magnetic permeability of free space. It is a constant, valid anywhere in the universe in any material. The relative permeability (ur) is what you want from your core, and the inductance of the core is a product of u0 * ur, how big the piece of material is (A) and how long the flux path is (l), times turns^2. Since u0 (4*pi*1e-10 H/mm) is a fixed value, we just use the measurement of L to tell us what the core's permeability ratio is with respect to u0. Don't worry, everyone calls it permeability when they mean relative permeability. Actual permeability is a number much smaller than 1. Relative permeability is always greater than 1000 for an ungapped core of ferrite material.

But really, don't you think it's easier to guess the core material based on AL value? Even if you determine permeability, the AL value for your core is still a function of core area and path length. This is why the ferrite mfg make a separate datasheet for their core material - this is the only place you will find a true permeability graph of the raw ferrite material. Once the ferrite is molded into a core, the better way to describe it is AL because now, A and l are fixed for that size molding.