Does a switched power supply need a sheilded inductor

Thread Starter

Travm

Joined Aug 16, 2016
363
For EMI reasons?
I've made a couple low frequency switched power supplies, about 1a. Considering EMI now. Lots of black magic talk about EMI on the web.
75-100 khz.

They work great, awesome learning, but if I'm thinking EMI do I need to shield the inductor?
 

Thread Starter

Travm

Joined Aug 16, 2016
363
Well yes, I'm thinking about FCC testing. Not that i have a need, or reason to do this at the moment.
The intended purpose of what i'm doing is exempt, and i'm certainly not paying for it, but I have an obsession with doing things as properly and thoroughly as possible. To a fault even.
 

Papabravo

Joined Feb 24, 2006
21,159
One way to instrument the actual testing on the cheap is to use a simple portable AM radio. Place the radio near the power supply and tune over the range of 540 kHz -1640 kHz. If there is any significant EMI you will hear it. This is not a perfect solution, but it is an inexpensive first step. A better solution would be to use a spectrum analyzer and actually measure the radiated energy across a much wider bandwidth than is available with a portable radio.
 

Marley

Joined Apr 4, 2016
502
Your main inductor should be toroidal if possible. This will reduce the external magnetic field. Then you need to take care with the layout. Especially grounding. Have a PCB ground plane and keep connections short and direct.

Then, have filters on the input and outputs. Often the inductors used here are not toroidal but probably better if they are.

So no magnetic shielding needed. An AM radio will always pick up a lot of EMC close up - this is not very good test.

By the way: I find that most switchmode power supplies (even CE marked ones) fail our EMC emissions tests. Usually extra filtering is required.
 

Thread Starter

Travm

Joined Aug 16, 2016
363
One way to instrument the actual testing on the cheap is to use a simple portable AM radio. Place the radio near the power supply and tune over the range of 540 kHz -1640 kHz. If there is any significant EMI you will hear it. This is not a perfect solution, but it is an inexpensive first step. A better solution would be to use a spectrum analyzer and actually measure the radiated energy across a much wider bandwidth than is available with a portable radio.
Awesome. I will play with that.
 

Thread Starter

Travm

Joined Aug 16, 2016
363
Your main inductor should be toroidal if possible. This will reduce the external magnetic field. Then you need to take care with the layout. Especially grounding. Have a PCB ground plane and keep connections short and direct.

Then, have filters on the input and outputs. Often the inductors used here are not toroidal but probably better if they are.

So no magnetic shielding needed. An AM radio will always pick up a lot of EMC close up - this is not very good test.

By the way: I find that most switchmode power supplies (even CE marked ones) fail our EMC emissions tests. Usually extra filtering is required.
Definately not using a toroidal inductor.
The datasheet for my regulator has an optional post filter, would using this further reduce emissions?
Where and how to filter is still a dark art to me. Although I did find an excellent app now from I think STM micro that I have been reading through when I can.
 

BR-549

Joined Sep 22, 2013
4,928
If you are going to use a micro for control.........you should check out keeping a little inductance......but throwing the inductor away.

Check out joey's post on this sub-forum, about a week ago.......on the CUK switcher.

A quasi-resonance circuit.
 

Thread Starter

Travm

Joined Aug 16, 2016
363
If you are going to use a micro for control.........you should check out keeping a little inductance......but throwing the inductor away.

Check out joey's post on this sub-forum, about a week ago.......on the CUK switcher.

A quasi-resonance circuit.
That isn't what I'm doing right now, but exactly the kind of stuff I like hearing about. Thanks
 
Top