Here at CET, we do custom transformer and inductor solutions every day. One such part was a high current custom choke desired by a customer of ours. The design of this part actually went through a couple of iterations. Initially, we designed for performance keeping cost in mind. Our solution was a performance success, but the customer thought we could do better on our price.
We went back to the drawing board and came up with a solution that both worked electrically as well as worked in terms of the customer’s budget. In most cases when we are asked for a choke it is required to have some inductance @ 0A and then some lower inductance at some particular load current. The difference is usually something like a 5% to 10% drop in inductance.
The solution we came up with, after determining the customer only cared about the inductance at a load current (25A in this case) was to wind the inductor on a core that did not saturate at 25A, but this core did have a “not so ideal” B-H curve. This is what made the core inexpensive. As the currents neared 25A, the change in flux density did not match in step with changes in applied Amp-turn/meter. This meant the permeability dropped considerably which means so does the inductance.
Below is the inductance of a toroid equation. We knew the initial relative permeability of our core, so we took an “educated guess” as to what decreased value it would be at 25A. We set the required inductance at 25A equal to this equation and solved for the turns. This was then used in the magnetic field intensity value (Amp-turns/meter) and this value was then used to find the % initial permeability due to an applied magnetic field intensity off the core manufacturer’s datasheet. This % drop was applied to the initial permeability to get a new relative permeability for the turns determined.
This new permeability was then used to get the inductance using the above formula with the same turns as previously. This was compared with what was needed. The turns were then adjusted in an iterative manner, each time calculating a new magnetic field intensity, and from this finding a new % initial permeability that was used to get a new inductance to compare with what was desired. For every new permeability value found, inductance was calculated and then turns adjusted to again meet the desired inductance.
The iterative process resulted in a toroid wound with 68T’s with a no load/0A current inductance of 885uH and an inductance at 25A of 270uH, exactly what was desired. This inductance drop was nearly 70%! This is much more than we would usually ever design for but again the customer did not care about the no load/0A inductance and so this allowed us to engineer a cheaper solution by taking advantage of a core whose B-H curve was only somewhat linear over a smaller range and yet could handle the flux density without saturating. The core was also decent in terms of core loss.
Given the frequency of use Litz wire was the preferable choice but to cut down on cost, the skin depth was found at the operating frequency and a wire diameter is chosen such that the skin depth was a bit more than the wire radius. This size wire was not enough for the 25A of current but we found winding 4 of these QUADFILIAR had an equivalent area that would suffice in keeping the DCR low enough so as to not create excessive power loss due to resistance.
With a design in mind and approved by the customer, samples were made and performance measured. All electrical values recorded were within just a couple percent of that predicted and the part worked just as the customer had hoped. Production orders came soon after at a price the customer was happy with.
If there’s way, we here at CET will find it. Whether your driving concern is performance, cost, or both we always do our best to satisfy your needs.
