# Thread: Fantastic paper: On Size and Magnetics

1. ## Fantastic paper: On Size and Magnetics

I'm often frustrated by armchair engineers that seem to think the entire industry is wrong and that they're right because they have a basic fundamental understanding of switch mode power supplies (as did I when I started this site) and "on the surface" things should be different.

On such argument is that with higher switching frequencies, there should automagically be smaller magnetics. But we haven't seen that much, have we?

So I found this great paper called "Why Small Efficient Power Inductors Are Rare" by some folks at Dartmouth College. I printed it out (old school, I know) and read myself to sleep (because it's.. well... engineering stuff), but when I woke up, I finished it and really appreciated the work that went into it and now I'm sharing it here:

https://cpb-us-e1.wpmucdn.com/sites....ling_Paper.pdf

2. ### The Following 3 Users Say Thank You to Jon Gerow For This Useful Post:

ashiekh (1 Week Ago), Microflop (1 Week Ago), rafal_iB_PL (1 Week Ago)

3. Is it possible that the future moves to capacitive based voltage transformation rather than inductive? by this I mean capacitors as part of the voltage transformation, and not just for smoothing.

Crudely speaking, one charges up two capacitors in series, and then rewires them in parallel, so halving the voltage. I can already see problems with this approach, as to charge a capacitor up to voltage V from a source of voltage V implies that at least half the energy is lost.

I see myself more as a rocking chair idiot than an armchair Engineer.

I was amazed to learn that beyond a certain frequency Litz wire is a liability rather than an asset.
Last edited by ashiekh; 1 Week Ago at 03:41 PM.

4. Very interesting paper.

Originally Posted by Jon Gerow
On such argument is that with higher switching frequencies, there should automagically be smaller magnetics. But we haven't seen that much, have we?
At least concerning the VRMs of graphics cards or CPUs we do see smaller magnetics by ramping up the switching frequency. As long as losses are no concern the inductance and thus the magnetic size behaves inversely proportional to the switching frequency. And of course multi-phase converter design as you can also see with those VRMs is a very easy approach to get the magnetic size smaller. As you half the inductor current by every doubling of the converter count. As magnetic size is equal to the square of the magnetizing current with every doubling of converter count you get half the total inductor size. For isolating converters things are different of course.

@ashiekh
Yeah, switched capacitor converters exist, but as far as I know they only work as they rely on some inductance as well. Because those converters only work by utilizing resonances which of course only exist as you have some inductance in your circuit.

5. Why don't PC power supplies use multi-phase? that would reduce the smoothing burden.

Inductance would reduce loses from using capacitors alone.

6. Originally Posted by Microflop
At least concerning the VRMs of graphics cards or CPUs we do see smaller magnetics by ramping up the switching frequency.
True, but you're also talking about DC to DC buck converters at a much lower voltage and a smaller power range.

7. That's right. For isolating converters things are a bit different. But even for higher voltage buck or boost converters the downsizing of both capacitors and inductors by interleaving can still be applied.

@ashiekh You can't really smooth the output voltage ripple by interleaving a LLC converter for example. The ripple you see at the output voltage rails comes from the 50 / 60 Hz mains grid, so only bigger bulk caps help for that matter. At the Enermax Maxtytan there is a interleaved boost pfc, but indeed you see it relatively rarely in PSUs nowadays.

8. ### The Following User Says Thank You to Microflop For This Useful Post:

ashiekh (1 Week Ago)

9. Thanks, I thought the ripple was at the switching frequency.

Seems to me that mains ripple could be eliminated with better switch mode feedback rather than larger capacitors.
Last edited by ashiekh; 1 Week Ago at 04:42 AM.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•