Hi there,
Yes it is true that winding your own inductor for power electronics
is a bit more involved than doing so for an air core inductor for RF
work. That is mostly because of the saturation characteristic of
the core which has to be observed as well as the value of the
inductance itself. That means there are at least two things which
have to be considered when designing the inductor rather than
just the 'usual' one. Add to that the possibility of the need for an
air gap, and you end up with a design that requires three parameters
to get right. Since the core itself is many times not changeable,
you have to predict all this before you buy one or you will find that
it is not suitable for the intended application.
What this means is that we have to pre calculate the design BEFORE
we order the core to make sure that it will work for what we want to
use it for. In the old days, this would be done by calculating L with
a prospective core, then calculating H, then adding an air gap, then
recalculating L, then recalculating H, etc., until we reach a design
that works. If the air gap is too large we can move to a different core
material and start again.
A problem that comes up with toroids though is that it is difficult to
make an air gap of exact specifications without some pretty expensive
machinery. I have experimented with 'cracked' cores in the past, where
we break the core up into four sections in a vise and then glue it back
together with the resulting cracks constituting the air gap, but it's not
easy to do this and the core has to be reconstructed carefully and so
in general it's rather hard to do unless you are stuck with a given core
and can not get another one. It's better to buy a core with a lower
permeability and start from there, so you dont need an air gap.
Here are some forumulas for L and H for constructions without air gaps:
H = 0.4*pi*N*I/ml
L = 0.4*pi*N^2*u*A*10^(-8) /ml
L is in Henries and H is in Oersteds.
u is the relative permeability, ml is magnetic path in cm, A is area in sq cm,
I is current in amps, N is number of turns.
Basically what you could do is iterate these forumulas until you come
up with a workable design. This will help to show you why it is so
much easier to purchase a design that already meets the specs that
you need rather than wind one.
If you still want to do this anyway, you can solve for ml using a value of H
that is maybe 25 percent lower than the max H for the core in question and
then inserting that into the equation for L and choose u and A to match
and look for a core of this size and material. The idea is to find a ready made
core that will fit these equations even if it ends up with a little bigger
inductance than you needed in the first place. H must be lower than the
max for the core because of other variations that are encountered in a real
application.
For constructions with air gaps the formulas are a little different, but as
i said it is difficult to make an air gap in toroid cores.
BTW it is best to test the inductance with a little dc excitation current as
some of the small inductance testers wont always show the correct inductance.
This is because of the slope of the BH curve near zero H, which is much less
than once we move up on the curve a bit for inductors with cores. One way to
do this is to use the inductor in the circuit it is intended for, and check the current
and the di/dt with a scope for different load currents. If the current shoots up suddenly
that means the core is saturating. If the di/dt is not correct that means the
inductance value is not right.
I've built a number of these kinds of converters in the past for currents of 3 amps, and
the inductors required come out to less than 1 inch in diameter. That's not too big
unless you need a very micro miniature design.