Help decipher Magnetics Inc. Specs

[gearheadgene]

I need help with the method of chosing a core from Magnetics, Inc. This is supposed to be a toroidal design. Following the data sheet, I get stuck on item 4B, 'calculate the bias in Amp-turns from : H = N * I / Ie'

What is Ie? They include an example, yet it's not clear what Ie is. Here's the data sheet:
Magnetics Inc. Core Selection

[gearheadgene]

I think I can answer my own question, it's the path length.

[gearheadgene]

I really don't quite understand something. Let's say a DCM flyback design calls for:

a) primary inductance = 10 uH
b) primary current min = 0
c) primary current time-averaged = 1A
d) primary current mean = 2A
e) primary current peak = 4A

According to Magnetics, Inc. data-sheet, "Core Selection Procedure", the very first item says to calculate L*I^2 where I=DC current. From above, which is the DC current? Is it the mean, peak or average? I'm trying to get the answer from Magnetics directly, and they currently say it's the peak number - but I'm not convinced since the data sheet clearly says "DC".

Any help?

[sawreyrw]

Don't believe everything you read. It is the peak current that determines the size (energy capability; Energy =1/2 * L*I^2 ) of the core.

[winny]

Yes, sawreyrw is right.

Just think of saturation. If you design it for the average current, it will loose inductance or even saturate the core at your maximum current.

[dtproff]

Power Integrations has a pretty good Excel spreadsheet that might help you.

The best PSU book I've read was McLymans Magnetics book. Greak information there that isn't readily avaiable elswhere.

Tony

[gearheadgene]

OK.

Next they (Mag Inc) have a method to derate the inductance/turn (Al) by the DC current. The charts are called 'Permeability versus DC Bias Curves' where the X-axis is 'DC Magnetizing Force (NI / cm)'

So again, use the peak current numbers when calculating N*I ?

I can understand using the peak current when selecting the core because you don't want to saturate the core under any condition, including peaks. But, if I understand this permeability derating, it changes with DC magnetizing force (N * I) and I is constantly changing. Seems to me that the approprate current to use is the average value, so that the resulting inductance per turn (Al) represents an overall figure, not just the inductance at the peak.

If I use the peak value, the derating changes the Al figure lower, so it takes more turns and therefore the inductance is going to be higher than expected.

Is this the what really is happening?

[zilog]

The more oersteds (ampere-turns per meter) you give the core, the lower will the permeability and thus the inductance for a given current be. I perfer to use a material that can take lots of energy storage before it starts to saturate just because of this reason. Check the difference between -26 and -2 materials, I have chosen -2 for the output inductor for my push-pull current mode smps to be able to consider my system more or less the same for all load levels.

A question for the others: Is there any "common" way of compensating say a boost or buck-converter when you know that your inductor is walking into gradual saturation for the higher load spectra?

OK. Next they (Mag Inc) have a method to derate the inductance/turn (Al) by the DC current. The charts are called 'Permeability versus DC Bias Curves' where the X-axis is 'DC Magnetizing Force (NI / cm)' So again, use the peak current numbers when calculating N*I ? I can understand using the peak current when selecting the core because you don't want to saturate the core under any condition, including peaks. But, if I understand this permeability derating, it changes with DC magnetizing force (N * I) and I is constantly changing. Seems to me that the approprate current to use is the average value, so that the resulting inductance per turn (Al) represents an overall figure, not just the inductance at the peak. If I use the peak value, the derating changes the Al figure lower, so it takes more turns and therefore the inductance is going to be higher than expected. Is this the what really is happening?

[dtproff]

Maybe this will help.. It's my flyback calculator worksheet.

Dor DCM/TM/QR mode use Delta-I of 200% and set the dutycycle to 50%. For CCM I use a dutycycle of 60% and delta-I of 35-50%.

Part 1 of 2

[dtproff]

part 2 of 2