The Top Questions About EMI

February 22, 2017

David Bourner’s recent webinar “Simple Ideas to Make EMI Issues a Thing of the Past” generated some great questions about approaches to mitigate EMI. Perhaps this isn’t surprising as EMI can present real challenges to engineers.

If you missed the webinar, you can watch it on demand, and check out the best questions and Dave Bourner’s insightful answers below.


Is it possible to avoid using input common mode ferrite filters with the proper PCB layout only, particularly for a hard switching SMPS with external discrete MOSFETs?

The art of developing great noise control in an electronic assembly is the correct combination of the right discrete components, layout and placement of all the electronic functions required. There is a relationship between the frequency of the noise and which particular measures are most effective within that frequency range. With the higher frequencies associated with optimal ferrite performance these components are best placed close to the noise sources to be suppressed, in the same way as we consider the use of X and Y caps.]

For best effect you should think about using ferrites, with appropriate damping added as necessary, in conjunction with X and Y capacitors. They can with care, be made to work together very well for HF noise suppression.  They do the job of really holding down most of the HF noise energy.

If you want to use ferrites, I would advise you to be very cautious about them. A ferrite, from one manufacturer with the same part number as another manufacturer’s – may perform quite differently. As far as ferrites are concerned, the main concern there is that they are providing suppression at the current levels that are passing through them, in the way that the data sheet explains.

I’d recommend spending time studying the really simple components like the ferrite or the capacitor referring to datasheets and having the applications engineers deal with any issues you perceive in applying them correctly. Ask yourself, is this the right ferrite to use for this application?


What simple equipment can be used in a lab to take your own measurements before going to an expensive test and certification house?

That’s a good question. I would say that if you’re working on conducted emissions, you should be using an oscilloscope with a sample rate of about 2 GSa/s, which translates into a bandwidth of roughly 500 MHz.

You should be using probes that are properly designed for high frequency noise measurements. In other words, don’t use elements such as pigtails attached to the ground lead of the probe!  You should think about a coaxial presentation of that probe to the circuit on the test, which ensures that the shield and the signal conductor are very close together.


Is it good practice to connect a protective earth (PE) with zero volts?

At some point, if you want your shield enclosure or your chassis to be at a safe voltage, you have to eventually connect things that are metallic to some sort of zero voltage reference. Generally speaking, when it comes to noise control, star-connected shielding, basically connecting to a single point, which is close to ground, is the best approach.


How important is the location of the power supply?

Normally you are going to locate the line filter very close to the input side receptacle where power goes into the enclosure, and the power supply needs to be really close to that. It’s no good having long cables connecting the input filter to the power switcher, even within the chassis: those cables are going to radiate! Placement orientation, particularly of inductors, is extremely important.


What typical value do you recommend for the input and output capacitors? And what selection criteria would you use?

It will depend on the nature of the product that you’re putting those X and Y caps around. There are also safety ratings associated with the X and Y caps.

I would always go to the data sheet of the power module I was using and check what the manufacturer recommended. Then I would look at the leakage current, because leakage current at line frequencies is constrained; it has a threshold or maximum permitted limit which will depend upon your applications.  I would make my assessments of leakage currents and make sure that I don’t exceed those.

Whatever happens, I think you need to be planning for this fairly early on in the design, not at the end of it.

Tags: , ,

Comments are closed.

Find out more about our Cool-Power Buck Regulators subscribe to vicor newsletter Contact Us

Get Connected