A Practical Look At Chokes For EMI Control

Radio frequency electronics can seem like a black art even to those who intentionally delve into the field. But woe betide the poor soul who only incidentally has to deal with it, such as when seeking to minimize electromagnetic interference. This primer on how RF chokes work to reduce EMI is a great way to get explain the theory from a practical, results-oriented standpoint.

As a hobby machinist and builder of machine tools, [James Clough] has come across plenty of cases where EMI has reared its ugly head. Variable frequency drives are one place where EMI can cause problems, and chokes on the motor phase outputs are generally prescribed. He used an expensive choke marketed as specific for VFD applications on one of his machines, but wondered if a cheap ferrite core would do the job just as well, and set to find out.

A sweep of some ferrite cores with a borrowed vector network analyzer proved unsatisfying, so [James] set up a simple experiment with a function generator and an oscilloscope. His demo shows how the impedance of a choke increases with the frequency of the test signal, which is exactly the behavior that you’d want in a VFD – pass the relatively low-frequency phase signals while blocking the high-frequency EMI. For good measure, he throws a capacitor in parallel to the choke and shows how much better a low-pass filter that makes.

We love demos like this that don’t just scratch an intellectual itch but also have a practical goal. [James] not only showed that (at least in some cases) a $13 ferrite can do the same job as a $130 VFD choke, but he showed how they work. It’s basic stuff, but it’s what you need to know to move on to more advanced RF filter designs.

6 thoughts on “A Practical Look At Chokes For EMI Control

  1. No comments? Seemed like a great practical use situation and lab test method to me. Always is interesting observing the simple ways to test and then the rack mount version methods… even if with bench equipment since the devices are getting smaller and smaller. Maybe I’m deep into studying filters for the time being and my first SMD 0603/0805 soldering was practicing for the those. The one in the back was my first attempt: https://www.facebook.com/photo.php?fbid=3700447600017015&set=p.3700447600017015&type=3&theater

    1. Yeah interesting video. I actually watched it to find out how putting a cap in parallel with the ferrite would improve its performance (ok I’ll admit I was just going to watch it and post a critical comment because clearly that’s incorrect) but watching the video I realized the misunderstanding. Cap goes in parallel with the source not the ferrite.

      Good practical information for those looking to learn about ferrites and filters. Phasor-level circuit analysis will get you very far in noise land.

  2. Bigger is not always better, try a smaller ID ferrite. Less ID is much more effective for the same cross sectional area and a lot cheaper and easier to find from a reputable source like Laird products from Allied or Digikey. Relative effectiveness from one ferrite to another of the same composition: = Turns*(Cross sectional area)/ID, IIRC.

    I have chased EMI issues quite a few times and learned a lot in the process. Including the characterization of the knowledge/noise ratio of select colleagues who always talked about EMI like it was a mysterious beast that would come out of no where and cause problems in your crappy designs. It isn’t. No forethought went into minimizing EMI on several of the machines I had to redesign a lot of. Some of it was fixed in 10 minutes by adding ground terminal blocks nearby cable ends in the wireway and cutting litterally 25ft of shield extension wires shorter to the new TBs. I had a pile of GRN/YLW at my feet. That pissed of the resident know-it-all who was “investigating” the problem. Nothing to investigate, this is basic stuff I told him. Too late now I cut all the wires, bye! Walked away. Didn’t want to hear another story about how water has memory or how many Italian Villas he owned. I am not joking…

    Besides the ferrites, any phase cables must be shielded and the shield terminated to ground with a shield clamp like an Icotek 37620. I really like those clamps. DON’T solder on a wire extension to a braided shield and run to a terminal block or other ground, no matter how short. You need surface area. I have best results with just one common mode ferrite, with the wires wrapped together instead of separately. As close as possible to the VFD/Drive. A cylinder is better than a donut, you want that extra cross sectional area and minimal ID. Always do a temp test after. I have had to reduce turns through or upsize after finding the ferrite was getting too hot. You can get rid of a lot of noise if you concentrate on the Drives and servos but don’t overlook good shielding practices for other cabling. I went from about 300v of noise to 40mV in one instance. Not sure I fully believe 300V, but that was what was on the scope with a diff probe.

    *Line side EMI filter for drives like he shows
    *Proper shielding of all cables
    *Ferrites on phase cables
    *Segregation of noisy bits
    *grounding all sheet metal and ensure incoming PE is good.
    *Ground loop awareness

    1. Thank you for the long post. I have a question about “A cylinder is better than a donut”: do you have experience with other shapes and their relative effectiveness at a given size? (Like a double E, or pot type)

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.