Magnetics is the Key to Power Supply Design

A frequent discussion we have regarding power converters is about magnetics.  Usually it goes something like this — they inherited the design of an existing power supply and may or may not be having issues with the design.  They may want to update the design with new components, or they were told they were responsible for a legacy product and want to understand how it operates. It could be the design was in process and the previous engineer left or a brand-new design and so forth.  Very few understand how magnetics work let alone how to design magnetics properly.  The biggest key to power converter success is the design, control, and manufacturing of the magnetics. Yet, magnetics gets the least attention of the entire system. 

Here is 200-year-old technology which is not taught in the universities or well-understood, and it’s the core issue with most power converter designs.  I will be chairing a RAP session on magnetics at APEC 2020 in New Orleans, www.apec-conf.org. The title is  “Where does the expertise for the next generation of magnetics come from? The Magnetics companies, or the engineer designing the power supply?" Participants on the panel include engineers from Analog Devices/Linear Technology, Vicor, Payton Magnetics, Rompower, Maxim, and Ridley Engineering (whose representative has taught over 3000 engineers magnetics design and the only known course on magnetics and magnetics design that I know of). It has been difficult to get magnetics companies to participate in this rap session – many either did not respond or politely declined for whatever reason.

Part of the discussion is that some magnetics companies do have competent magnetics experts on staff, and some do not.  Some view the magnetics as a key electronic component, and yet others consider it a materials science and commodity mechanical component.  When discussing magnetics with the supplier, the proper documentation, rigor and oversite are often not provided if they do the design. Lets face it — the largest problem a magnetics company has is doing the design for the customer, and the buyer at the customer farms out their design to every company on earth who ever made an electronic component to see if they can reverse-engineer it and drop the price. This cuts out the original magnetics company.  Conversely, if the design engineer at the customer has the knowledge and expertise with the proper design files, then the magnetics company gives up all the leverage, as well. In this case, knowledge is power and he who holds the information and documentation is in charge.

Depending on the magnetics supplier, we noticed that when measurements are asked for, some magnetics suppliers do not even own the appropriate test equipment to provide frequency response plots of the impedance vs frequency, find the resonance point, and other fundamental requirements for a successful design. They see their role as materials science and mechanical assembly providers, and besides, the only test equipment they can afford is probably a handheld DMM they had to beg their finance people for. Speaking of testing, who is responsible for the design testing? The production testing? What if the design was done either by or for the semiconductor company that did a reference design which was copied?  This should be an enlightening RAP session, indeed.

PSD