another inverter project

Author Message
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3338
 Posted: 09:13pm 21 Jan 2020

Nick, that link to the inductor calculator is pure gold.

The data books are a bit limited, because they have no idea what you plan to use their cores for, and they have limited space to include data for gapped cores, or huge extremes in operating frequency. There is also the problem that different manufacturers make similar cores from their own ferrite mix recipe which may be end up being very slightly different. Ferroxcube are the big guns in Europe, but TDK have a huge market in Asia and the east.  Their products are pretty much interchangeable but not identical.

Published data is certainly a great help initially, but it can never replace the results of actual testing.

What you need for choke design are called Hanna curves. This plots total air gap along the X axis and L I squared on the vertical Y axis. The scales are both logarithmic, and the "curve" is actually a straight line which is appropriate for a specific ferrite size and geometry.

L I squared is actually the stored energy in the choke, and becomes larger with a wider air gap. Scroll down to page 39

On the right hand side you will see the Hanna curve for U100-57-25. That curve only extends out to about 3.5mm total gap, but right at the end of the line the quoted energy is 200 mJ.

What that means is, that core set with that air gap will always end up with Amps squared times inductance value of 0.2 Joules.
So, amps x amps x inductance (in henries) = joules (just below expected saturation).
If we put on just sufficient turns to produce 50uH, we get:
Amps squared = 0.2 joules divided by 50uH
Amps squared = 0.2 / .00005 = 4,000
Amps = 63.24  with the specified 3.5mm total air gap (2 x 1.75mm spacer)

Now a 1.75mm spacer is not a lot, and we can do better with a wider air gap but there can be issues with that.  The problem is fringing around the air gap.  If the air gap is in the middle of the winding, as it will always be with two identical core halves, that can create some odd effects that can make a nonsense of any calculations.

It depends on the application.  For chokes with a lot of dc and very little ac current ripple component, a central air gap should work pretty well. If its for something like a flyback switching power supply,a big central air gap can create some really nasty effects. There will be horrible ringing and the waveforms will look truly awful. That is all caused by fringing around the airgap influencing the voltages developed in the turns immediately above the air gap.

So the ferrite manufacturers do not like to recommend using really huge air gaps for general applications, although you should be able to get away with big gaps in a large over designed dc choke that has a very low ac ripple component.

You can get slightly more by extending the line on the Hanna curve, but will quickly run into diminishing returns. Much better to go to a larger core cross section.
Cheers, Tony.