Re: Inductor saturation current
Dr. Slack, Tue Jul 08 2008, 02:14PM

ignore the cross section (not needed, as long as airgap W and H are >> airgap length)
and assume the core has infinite permeability (for a good first approximation)

find the Bsat from the material specification (varies with frequency and acceptable power loss, choose an appropriate value)

compute the Hsat.air needed to get the Bsat across the airgap using u0 = 4piE-7

divide by the length of the gap to get ampereTurns (a proportionally large measurement error can come in here)

divide by the number of turns to get amps



As a minor refinement, you can add the Hsat.core to the Hsat.air needed to get Bsat through the length of the core, but it will be small compared to the airgap, and ignoring it gives an error on the safe side

Re: Inductor saturation current
tesla500, Tue Jul 08 2008, 07:25PM

In equation form:
B = H * u0
B = I * n / l * u0

where
B = flux density (Tesla)
H = magnetizing force (Amperes per meter)
I = current through winding (Amperes)
n = number of turns on winding
l = total length of core gap (meters)
u0 = permeability of free space (Henries per metre) = 4*Pi*10^-7

Solving for I:

I = B * l / (n * u0)

So, knowing the maximum flux density B for your core material, and the winding specs, you can compute the maximum current.

David

Re: Inductor saturation current
Dr. Dark Current, Tue Jul 08 2008, 08:25PM

Thanks


Edit- my core is of unknown material from SMPS, from what I found online I used Bsat=0.3T, does this sound about right?

Re: Inductor saturation current
GeordieBoy, Tue Jul 08 2008, 09:26PM

If you are not sure of Bsat I would err on the cautious side, because gapped ferrite tends to saturate more abruptly than the gradual loss of permeability that you typically see with iron-powder.

Also remember to use the peak current in the calculations. That is the worst case DC bias plus the positive AC ripple contribution if it is for something like a buck choke or DC link choke that has a +/- ripple component.

By-the-way if you are totally unsure of the material properties, dimensions and the air-gap, there is a test that you can use to find the approximate onset of DC saturation for "junk box chokes". I'm sure the power electronics guys on here can describe it in more detail but it basically involves discharging a large capacitor across the choke and monitoring the rising current using a CT, current probe or Rogowski coil. Current initially rises linearly at a rate determined by the applied voltage divided my the inductance, then the current takes off at a much steeper slope as saturation sets in and inductance falls. You can can approximate the position in the current ramp where the gradient changes.

-Richie,

Re: Inductor saturation current
Dr. Slack, Wed Jul 09 2008, 07:00AM

Bsat=0.3T sounds fairly reasonable for low freuqency operation.

I was a bit inaccurate in my first post saying that Bsat varied with operating frequency, it's the manufacturer's recommended Bmax that varies, and it's to do with the power dissipation due to going round the hysteresis loop. For reasonable ferrites, you should be able to work all the way up to Bsat at frequencies to 100kHz without the core getting too warm, derate Bmax at frequencies above that.

Re: Inductor saturation current
Steve Ward, Wed Jul 09 2008, 07:15AM

For reasonable ferrites, you should be able to work all the way up to Bsat at frequencies to 100kHz without the core getting too warm, derate Bmax at frequencies above that.

Most power ferrites ive looked at suggest derating above 20khz, but these were strictly big power ferrites, where you are less likely to run really high frequencies.

Ive tried the ramping input current to watch for gapped ferrite saturation. The current was switched from a big lytic with an IGBT. I personally have done this with special high bandwidth hall-effect type current sensors (so i can do DC and AC up to 300khz BW).

Re: Inductor saturation current
Dr. Dark Current, Wed Jul 09 2008, 07:29AM

Thanks guys. Though i have no way of looking at single "transient" events with my tube scope
I think I'll start a new thread as I'm lost in the dc link inductor design...

Re: Inductor saturation current
Steve Conner, Wed Jul 09 2008, 09:26AM

Just use a square wave instead of a single pulse, then. I've used a square wave from a signal generator boosted by a 100w audio power amp for this purpose before. And a low value resistor in series with the choke works fine for current sensing. You can use the same setup to measure ESR of capacitors.

Re: Inductor saturation current
uzzors2k, Tue Jul 15 2008, 10:21AM

Quick question: The total air gap length in a U or E core set is the summed length of every air gap in the core, not just the one air gap on the arm which the winding is placed on?

Re: Inductor saturation current
GeordieBoy, Tue Jul 15 2008, 07:02PM

Yes, the total air gap is the sum of all the gaps in the magnetic path. But...

It is best to only gap the centre limb of ETD-cores and E-core sets, or the centre post in things like RM-cores and pot-cores as this minimises the external magnetic field. The outer most limbs of the E-core and the shell of an RM or pot core is not normally gapped.

However, for prototyping you divide the total air gap into two. You put half of the air-gap in the centre limb, and half of the air-gap in the outter limbs.

So just to clarify one last time. If you have an E-core set and put 1mm gaps in the centre limb AND each outter limb. The total air gap length is 2 x 1mm = 2mm. Not 3mm because only half of the total flux takes each of the two external limbs as its return path. I hope that is clear.

-Richie,