2
\$\begingroup\$

I am planning to design a 65W power supply which will use a PWM frequency of around 100kHz. I am specifically concerned about conducted EMI (not as much about radiated for now). I am currently considering using the following to lower my EMI:

  1. LC ladder filter
  2. Common-Mode choke (using 2 inductors)

Knowing this, if I chose to ignore efficiency, wouldn't is be better to you a higher frequency? If I choose a higher PWM frequency (say 1MHz), wouldn't have lower conducted emissions?

\$\endgroup\$
1
  • 1
    \$\begingroup\$ The actual risetime/falltime of the switch is a harder contributor to EMC. 100kHz and all its harmonics is alot easier to remove than a massive spike at say... 20MHz \$\endgroup\$
    – user16222
    Mar 19, 2014 at 20:30

2 Answers 2

2
\$\begingroup\$

There is a third option and that is to modulate the PWM frequency across a range around the centre frequency of 100kHz. There are a few chips that do this and this "spreads" the emissions around and can make passing the EMC testing easier. This is because the average time at one spot frequency is reduced and, in effect you are distributing emitted energy across a wider part of the spectrum. Some say it's a bit of a cheat but plenty of folk do it.

Any filtering you apply (sensible filtering of course) is going to help. I also found that a copper shield around the switching transformer also helped but there will be anecdotal evidence from several folk. I used a common mode choke and I believe this to be very important.

Choosing a higher frequency of 1MHz may not improve things - the switching rise times and fall times have to be so much sharper to get the same efficiency from the supply and this just pushes emissions to a different part of the spectrum.

Here is an article from MAXIM that explains how spread-spectrum modulation improves emissions in a class D audio amp. It's a rather approachable document and that is why I've linked it.

This document from TI is also very useful in that it takes you through how the labs measure emissions and tells you how the spread spectrum modulation helps.

The circuit below is a good example of using spread spectrum and generally I've found that Power Integrations (the company) provide really useful products and design help: -

enter image description here

\$\endgroup\$
1
  • \$\begingroup\$ I see... yes, I will definitely be using spread-spectrum \$\endgroup\$
    – hassan789
    Mar 19, 2014 at 23:20
1
\$\begingroup\$

There isn't really an optimal frequency. The filter has to be designed to attenuate properly to meet the requirements.

Common mode chokes (CMCs) are a must. But they have to be coupled, two inductors isn't going to cut it. Here is an example of a CMC that can be useful in the tens of megahertz, especially when used with shunt capacitors.

Higher frequencies can allow use of smaller inductors and capacitors which may have better parasitic characteristics. Or higher frequencies may allow you to get farther down the attenuation curve on the filter you do have. But you still have to be aware of the spectrum that needs to be attenuated. One other thing about higher frequencies is that many of the multi-megahertz converters are soft or resonant edge switched, but then you have to deal with restricted operating ranges.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.