9 Simple Tricks to Improve EMC / EMI on Your Boards - Practical examples (with Min Zhang)

[@zhitailiu3876]

Every hardware engineer should watch this video! Totally life saver!

[@RobertFeranec]

Thank you Zhitai

[@simonbaxter8001]

Absolutely fantastic video Robert and Min. More like this please!

[@RobertFeranec]

Thank you very much Simon

[@weeeeehhhhh]

@@RobertFeranec I really like the interview techniques here too. It can be hard for someone as knowledgeable as Min Zhang not to skip over detail which he takes for granted, you did a great job of presenting all the detail

[@DarianCabot]

Can we just take a second to appreciate how much trouble Min went to when researching the power supply example and creating these slides. Robert is asking some very good questions, and Min shows an exquisite understand of the DUT from his in-house "pre-compliance" configuration. I'm part way through Min's paid online EMC training course, and I can say his attention to detail and ability to explain things in practical terms is a hallmark.

[@VijayKumar-lk9ir]

Uh808h

[@VijayKumar-lk9ir]

Uh808hu

[@AlbertRei3424]

How is his paid online course?

[@dominiqueguelordingala1935]

A very good demonstration. And I'm most happy with you Robert as you asked almost all the questions on the behalf of many. Excellent video.

[@erikmjelde4428]

I am so glad I found your channel. You have been filling my brain with all kinds of PRACTICAL knowledge after recently completing my BSEE.

[@majdithouabtia5582]

Thank you for bringing experts who can show us experiments and not only talking about theory. This guy just goes straight to the point and that's usefull! Big like and respect for such people!

[@bimmerboard]

I would really like to see a high voltage (120-240V) version of this discussion.

[@____________________________.x]

I've often asked other channels to measure the noise of these convertors, so it's really useful to see it done here at last. This should help my problem with using one in a car, as it's interfering with the radio. Thanks

[@eljeffo2871]

Please bring Min back for the inductor cutout discussion! Information is really lacking in this area :)

[@sguwenkaVlog]

Yes, I also have a question about cutouts under inductors and transformers. For example, many pcb designguides from ic manufactures recommend gnd cutouts under ethernet/manchester transformers in all layers.

[@tomasbergh]

I think cutouts in the groundplane is interesting when having onboard cm-chokes, isnt that the case?

[@cadd4276]

Thanks so much to Min for the great slides, and for taking the time to apply the improvements incrementally and reveal the various noise contributors. So insightful. Thanks for organizing this, Robert. I really hope to see more videos featuring Min. The intro, theory, process, simulation, measurements, setup, slides were great. Thank you!

[@Christe4N]

Thanks Robert, it's so good to see it demonstrates like that. Thank you both for the effort put on this call and video

[@RobertFeranec]

Thank you very much Christe4N

[@arunpant1723]

A gem from you Robert, amazing explanation in such a simple way step by step by Min. Would love to have more such practical case studies. Also tekbox LISN for measuring common mode & differential mode performance is very helpful. I learned that for the first time, didn't knew before about it. Thanks for the amazing content.

[@edouardmalot51]

For this test, can sur PSU be replace with a battery ? (no EMI)

[@RobertFeranec]

I do not know direct answer to this, however, in reality, during the test we usually use the power supply which will be supplied together with the board (so we are sure everything together will pass).

[@ProtonOne11]

You can replace the PSU with a battery, if the equipment will only be battery operated too. It acctually makes a difference if you use an isolated supply or a grounded (GND is earth) supply too, so that needs to be considered in your test setup.

[@MachOneDesignEMC]

@@ProtonOne11 thanks for the answer. indeed.

[@rewq7041]

Excellent video. I have spent two weeks to collect and watch all these information and techniques about EMI/EMC and PCB layout of a buck converter before watching this. The video is definitely a great short and detail summary of showing how to complete your EMI debug and filter design.

[@Packer1290]

What blew me away is we are often focused on filtering the switcher output. I'm shocked by how important it was to filter the front end.

[@millomaker]

Thank you for this video, I learn a lot :)

[@KeithRisebroughUK]

Great video, nicely worked through. Q) Where the text on the results graph shows 10db attenuation (which I understand is due to the transient protection device) does this mean that the measurements are actually 10db higher across the sweep? or is this compensated for?

[@MachOneDesignEMC]

Hi Keith, the EMCView software has compensation for transient protection, attenuators, amplifiers, etc. So all results shown in the video are compensated values. The reason we put the info on the video is because when putting things like transient protector, attenuators, it will affect the noise floor. So just want to clarify to avoid confusions.

[@ricardonunes6724]

Very interesting video! There are a few things that could be prevented by doing a better layout. Input capacitors of buck converters should be as close as possible to the FETs to reduce loop inductance. Also, I would never use an electrolytic capacitor alone. ESR and ESL are terrible. Unless it is a low current converter, the input ripple should be horrible.

[@MachOneDesignEMC]

Absolutely agree!

[@ВасилийДанилов-е7щ]

Thanks Robert for such a valuable content. This material helps a lot to understand EMI/emc concept , which is sometimes confusing. I think it would be very nice to see practical examples of how important PCB (for example 2 vs multilayers) is in terms of EMI. We always trying to follow lots of recomendations of how to do pcb properly, but what specific steps make impact in practice? If we would see 10 pcb with different layout and see, which is best/worse...

[@lordcape]

Min is an EMI crack! Great video!

[@rjrodrig]

Ming are how did you calibrate the measurement? Also how do you mitigate in rush current on power up with larger capacitance values potentially exceeding 1v/ohm derated values?

[@MachOneDesignEMC]

The test set-up was calibrated in our lab since we do EMC service for our clients all over the world. Calibration is almost a per week job. Your second question was a great one. We didn't. Because this is a demonstration video, we wanted to focus on the main topic. Often for inrush protection, design engineers would put things like inductor, or resistor with switch etc.

[@ray-charc3131]

A very good and valuable explanation. BTW, why EMC regulations apply to electrical appliance products that go to Europe but not to US ? While US needs the FCC regulation compliance for electronic products, what is the differences between the two regulations?

[@MachOneDesignEMC]

Most of the time, FCC and CE regulations are the same. It is really a trade issue rather than technical barrier. Each country will have its own rule, but often you can trace it to the EN harmonised standards.

[@akhil3rockstar]

It's very good work. It will be more interesting if he explained like this The method of designing the filter Identify impedance then find filter configuration then design the filter.. Stability .. How the common mode noise is generated.. Etc he is explaining only in the presective of conducted emission. The output capacitor has very much importance in my view . If you measure the current through the capacitor then it's will be high di/dt loop.

[@MachOneDesignEMC]

Hi, that's great feedback. We will take your points. The output capacitor is important, but as not as important as the input cap for buck converter, for other topologies such as boost or SEPIC, yes, one would also need to improve the output capacitor

[@shaon6923]

Hi Robert, it will be great addition for us (your viewers) if you make some video particularly on Common mode noise in smps and how it take care of it at the initial stage while designing psb and what are the effective way of mitigate the CM noise. Thank you a lot for your great contributions to power electronics world. Min and Robards, you guys are awesome. Some interaction with Ken Wyatt or Dr. Bogatin or with Rick Hartley will be great on this CM noise 😀

[@mohammedabdeltawab9882]

I think hot loop definition is not clear in this video. in my opinion it's the loop with high PULSATING current. so frequency contents includes high frequency band due to step edges compared with ramping current and the high amplitude implies high H field.

[@MachOneDesignEMC]

Very good and constructive feedback Mohammed, you are absolutely right by the definition. in this case, the input side of the DC-DC does have high pulse of current due to the switching events. I will remember your clear definition and use it from now.

[@rakeshpanda2116]

Do we really put differential mode and common mode filter before LISN? Is it even allowable? I think this will give incorrect measurement.

[@MachOneDesignEMC]

No we don't, you are absolutely right. The way we put the LISN mate is not 'before', but rather, using coupling and decoupling circuit to obtain signal. treat it simply as an ideal current transformer. Of course, it won't be ideal in reality, so it will have some impact, but not much. Hope this helps.

[@burkantorun]

I learn so much from your videos and interviews. Thank you

[@edmilsonsantos2295]

Great video, Robert! Make more videos with Min, please.

[@Violator420]

I recently had an issue passing CE102 from 300KHz to 1MHz. I decided to come to youtube to find help and came across this wonderful video. I would like to thank you and Min for taking the time to share your experience and knowledge. Turns out, simply increasing the total input capacitance was enough to be below the limit!! THANK YOU!!!😃

[@Violator420]

This is one of the most important videos in my career

[@TheWhatnever]

Great Video! Im wondering how different separate LC and Common Mode Chokes are to using the Inductance of the CMCs for the LC filter.

[@rjrodrig]

How did you plot the mask limits on the spectrum analyzer? Did you add that in excel?

[@MachOneDesignEMC]

Hi, we used EMCView software, which is really handy for tasks like this, for more info, see kzbin.info/www/bejne/foDJqmB-g9icr7M

[@michaelk.1108]

Thank you, Robert. A lot of very useful information. Very important topic. I always simulated the input filter from the other side. I think the result should be the same, because it is a linear system. Another reason to add ceramic capacitors: ESR of aluminium capacitors increase with low temperatures. I guess the results of the original board would be even worse below 0°C.

[@Swenser]

What would be your go to method for stopping digitalTV interference caused by LED ceiling light? Would it be more likely to be radiated rather than conducted? Shouldn't shielding around DC board grounded to earth help with that?

[@MachOneDesignEMC]

It depends. shielding around DC board ground, but you will need a good solid continuous ground plane, other wise the shield won't work effectively

[@Swenser]

@@MachOneDesignEMC I passed cables close to light through two small ferrites clip ons (one not enough) and that solved the problem

[@thanosprionas6919]

Excellent video!! Very useful! Could Min provide us more details about making emc test board in our lab?

[@MachOneDesignEMC]

Thanks for the feedback. Please follow us on the latest publications on emccompliance.co.uk/emc-testing-troubleshooting

[@zaikindenis1775]

Thank you for the video. But why do you test with that ground plane directly under the DUT? Normally, the device should be placed 0.8m above the conducting plane (floor) by the standard. The close ground plane worsens the results... Thanks!

[@RobertFeranec]

I will ask Min to have a look at the questions here.

[@MachOneDesignEMC]

Good question, yes, if you put the unit direct on the ground, the result will be even worse. We had some insulation material between the DUT and the ground plane. The important thing here is to demonstrate the techniques, and it is a pre-compliance set-up, so we are doing comparison study rather than quantity study.

[@zaikindenis1775]

@@MachOneDesignEMC Thank you!

[@kestralzoe9310]

It is so nice that we can watch this kinda talk video for free, thanks to Robert. I am sure it's gonna cost a good amount of money to have a workshop outside with all the juicy content like this.

[@MachOneDesignEMC]

Many thanks!

[@khalidmahmood8239]

Really it is super useful to understand the filter simulation and how to deal with EMI. After watching this video hardware engineers can easily understand the EMI and its filtering. It is really a very good effort made by both of you. Thanks.

[@TYGAMatt]

Wow. What an incredibly informative and interesting video. I am just a hobbyist but this is super useful. I'm trying to develop a capacitive discharge ignition system and I am sure that I can implement many points from this video to improve the performance. Many thanks to both of you for sharing your extensive knowledge.

[@yunuscay8870]

Realizing how the parallel MLCCs change the EMC performance in practice is very satisfactory. Thank you :) Keep up. The common and differential mode may be explained more with vary examples.

[@Necrocidal]

Always great when recommendations come with empirical data to back them! Thanks!

[@MrYamashici]

Wow such quality content for free. What a time to be alive!

[@parttymalien]

thanks for these tips. i've been trying to pass a 3kw totem pole pfc and full bridge llc. after a week of debugging conducted emi i can merely see an improvement especially at high frequency 4Mhz😂😂

[@jacovanzyl6424]

Don’t go for EMC/EMI testing before watching this video…… extremely well explained!

[@conesillyvalley7182]

U r assuming that all pcbs were designed like the one in discussion, I am designing smps for military applications for many yrs and these basic mistakes are never done

[@thetnaingoo1538]

Thanks so much Robert for choosing this topic and thanks so much Min for your great and clear-like-a-crystal kind of explanation ....

[@MachOneDesignEMC]

Many thanks for the feedback!

[@Lasseu]

Thank you, this was super useful for me. Very interesting. Thank you Min 😊

[@vimallmp]

Its very good

[@akadir786]

Amazing to watch Robert, very helpful, keep it up. How did you get the Frequency response curves of the filter Caps you added the 4.7uF 470nF and the 150nF ? Datasheets dont give them. Does the package size 0603, 0805, 1206 make a big diffrence?

[@MarcinKurczalski]

I shouldn't have waited with watching this... Great example and process. One thing that I did not get is from CMC simulation, how did you come up with capacitor value (150p) and AC magnitide (210m) - a rule of thumb?

[@flyingbirds6794]

@1.12.17 What is the reason of using coupled inductors at the left? I think you call it as common mode choke.

[@ats89117]

Nice video, but it was mentioned that the transistors were incorrectly placed and this was never addressed. The board layout also wasn't obvious. I have seen other videos where a light is shown through the board to show the location of the tracks and the top and a reversed copy of the bottom are shown to allow a view of the layout that can be commented on.

[@pigi_experiment]

There is something that I don't really get. Has the video show, the power supply has huge EMC/EMI problems. How is possible that this device is still now available to buy from Amazon ? Did any authority has never perform any sort of EMC/EMI test on any device available on the market?

[@conesillyvalley7182]

Disingenuous to say "most pcbs purchased from Amazon will fail EMC test" More accurate would be to say "most cheapo chinese pcbs will fail EMC test" Then again Min Zheng won't comment honestly about his own country's atrocious quality of electronics Robert, this is NOT an honest engineering discussion -shame!

[@kennethng8346]

Long but seriously worth it. I like the details and showing changes at each stage and explainations of what is happening.

[@sigfreed11]

This is a great video to demonstrate the basics of Power Distribution Networks 👍🏻

[@chaosopher23]

I work with 600 volts routinely with tubes. I know, I'm trying to perfect sealing-wax and the art of lamp lighting, but there's just something about tubes that say they're still awesome devices that haven't lost their place. Scaling up...

[@李天宇-e2d]

It is quite helpful to understand the EMI issues in PCB. Thanks a lot.

[@weinihao3632]

That was a great & informative video with an awesome, well-prepared guest! Maybe I missed the explanation, but what causes the jump in the noise-level of about 10 dB in all the graphs @ exactly 12 MHz?

[@MachOneDesignEMC]

Hi, it is not 12 MHz, since the scale is logrithmic, it is 30 MHz. the answer can be found in Michael Mayhofer's reply to one of the viewers, just check the comments in the stream and you will find the answer.

[@gabrielemarocco]

This a practical and powerful demonstration, thanks! The sentence "You want to deliver energy immeadiately" should be always remembered when designing a power supply I didn't know the existance of the LISN MATE in order to separate common mode noise from differential noise. A doubt. In the datasheets of some buck controllers it is suggested to use a 10uF or 22uF ceramic input capacitor. Can these high values ceramic input capacitors produce resonances at low frequencies?

[@MachOneDesignEMC]

That is a great question Gabriele, i would vote for 22 uF, the resonance really is caused by the layout of the design, isn't it, in a good design with trace and tracks short and tidy, i don't think the capacitor can easily resonate that much.

[@tomasbergh]

Yes, but it depends also on the surroundings of the capacitors such as routing.

[@Daniel-zw5ey]

nowadays most of the DCDC chip vendors give application notes for their chip for free, which contains already the recommended capacitances values, coil, and layout hints.. which is a good and also a bad thing for HW engineers. people then just know how to use that chip but don't know how to optimize the circuit if a failure occurs. so thanks for the great video ! one question I do have, is why is only the so-called hot loop so critical. in fact, most of the time, the high side and low side FETs cant conducted simultaneously, which means the "hot loop" is not ON. what I have in mind for BUCK design is always keep the "charge" loop (the loop where only high-side FET is on ) and the "discharge" loop (the loop where only low-side FET is on)as small as possible.

[@sunrise7824]

Hi, Sir I want import flatbat scanner Pcb board image in Pcb Altium 17.1.9. What I do?

[@kaushal.k6763]

How is the board failing the EMI test, If majority of the fixes was applied to the input signal. What I mean, is that the major contribution to the noise was due to the fact that the input signal given to the buck converter was not clean. So, If we want to use a buck converter which doesn't perform a input voltage filtering, will obviously have a poor EMI performance. Also, when performing a EMI test, the input supply is always filtered so that the input doesn't have an effect on the board under test.

[@andrewb386]

When connecting multiple high speed probes to DUT from a single scope, should all the probe wires be bundled together? In the rock band roadies world, they would call this a "snake" bundle. It supposedly to avoid creating ground loops in between the various cable shields.

[@Emanavas01]

Absolutely useful. Please continue bringing more videos like this. I subscribe now. 🤩

[@abhichokshi496]

Good video How to control emi in non isolated power supply like converting 230Vac to +12Vdc. And its layout consideration. I am working on this but its mysterious as it invilve high voltage.

[@kacperbardecki3671]

"I think it's cool for people to see how these small components make a big difference." ... I now understand why there are so many components on a PCB, I think.

[@hasnainmughal3046]

Never ever enjoyed a technical video more then this one, Thanks for sharing Can you share some video about Protecting microcontroller from EMI when switching inductive loads through an on board electromechanical relay, i have implemented all solutions available in books and internet but some how it manages to reset my MCU's some times

[@naveengupta5588]

Amazing video ,i have one doubt so if we place MLCC cap before Electrolytic one will it affect the results??

[@rjordans]

Quite happy to recognize most of these techniques, still, thanks for the great overview Min and Robert!

[@theminertom11551]

I have spent a lot of time in FCC EMI labs. Usually, you are "on your own". The operator is there just to work the machine and give you the results.

[@scottanderson102]

Any ideas on how to reduce common mode above 30MHz without using a common mode choke to save cost. Like improving layout in some way?

[@mariuszgajdel8003]

We are just after 2 days of course with Min. He change our picture of design the PCB by 180 degrees

[@songpandy9590]

In page6, the bottom right picture shows some devices. Are these device used in your own lab before going to a formal EMI testing?

[@sadeghghorbani3055]

Thank you so much for sharing this great video For a low power device can we use a simple RC filter(instead of LC filter) to reduce differential mode noise?

[@mohammedabdeltawab9882]

@Robert Feranec, would you please make a similar video for EFT , ESD and surge

[@andrewschiessl9941]

This was an extremely useful video! The topic is relevant for almost all products as the power supply is a critical piece of the design. Looking forward to more EMC videos as these topics are typically overlooked in college/universities.

[@jiali4123]

Great video. One question. Could you please share more info. about the antenna shown in the right side of slide 6? I wanna reproduce that emission test setup in my lab. Thanks.

[@muhammedmufeed1967]

Awesome video, good questions and explanations.

[@tantratron]

Would you have a similar video or tutorial on how to reduce the radiated disturbance from 30MHz to 1GHz ?

[@a_pullin]

Wow, really great stuff. Good to see it making its way onto youtube. (Robert, your content is great and def keep making it ... but you were really tripping him up in his presentation, just a little note)

[@velinr]

Wow. Very good and practical explanation. I'm eager to see part 2 - dealing with radiated emission (for the same board hopefully).

[@velinr]

I'm curios. There is a sharp step in the spectrogram, exactly at 30MHz. Is this caused by the SA (higher noise floor above 30MHz) or it is actual measurement (caused by DUT)? It is very strange to me.

[@MachOneDesignEMC]

@@velinr Hi Velinr, from 30 MHz, the standard defines different RBW (resolution bandwidth) and scanning time, hence the scan will use a different set-up as compared with measurement below 30 MHz. When I first got into the EMC field, I had question exactly like you. All the best

[@velinr]

@@MachOneDesignEMC , that means that SA switches settings in the middle of the sweep. This I didn't know. I don't know much about EMC measurements, so I find stuff like this very interesting. After all, every engineer, who designs a product, should obey the rules of the EMC. But they are not aways clear. Thank you for the explanation!

[@gustavomoyano9217]

excellent! thanks for sharing. A professional and fun video explaing clearly how the real electronics works in an end product.

[@thenet0120002]

The one big drawback with digital or switching power supplies is a badly designed supply can transmit a lot of noise. Using smd ceramic capacitor sometimes with a choke or resistor can reduce the noise generated by pulsing circuits which feed the main circuit to the secondary circuit. I bought a supply for my computer it had higher amperes but it is noisier than the original it make watching a TV nearby hard . It has no effect on the computer itself but it too is digital not analog . This need to filter out these frequencies flowing thru your supply. You need to know approximately the frequency transmitting so you can filter it out without affecting the frequency running this supply. Very common problem why designers do not just add a filter set up to limit noise is beyond me. Too many companies act like this does not matter even though the FCC for one has rules for omissions.

[@GuyFawkes911]

Great Video I hope you do another design series where you apply the emc theory we learned. My suggestion a flyback powersupply.

[@thenet0120002]

My hope is that engineers will develop a standalone circuit that knocks out the frequency of all these supplies much like reducing noise in noise quieting headphones . This way designers can have a tunable frequency cleaner that separate from the circuit that cleans up or cancels frequency transmission without big changes to a limited quality supply. It would involve pink and white noise filtering as a possible cleaning effect. You battle the curves to smooth this noise curve. A safer way to contain noise.

[@jimjjewett]

I got a lot more out of this the 2nd time through. I think some of the "throw-away" comments from kzbin.info/www/bejne/p4e1doiCeaenoM0 were useful context that I hadn't realized I was missing.

[@gilbertgonzalez7890]

Thank you Min and Robert for this video, so informative and interesting :)

[@michaellove7912]

@about 30:00, hot loop area is mentioned, it is a bit confusing to me, if the board has a solid ground plane and switching frequency is high, the return current on the GND plane should follow right underneath the top layer, so the area should be the length of the trace on the top plane times the PCB thickness. The way the slide is shown implies it is the area encircled by the red lines, is this not technically correct?

[@mehmetdincturk9496]

Amazing video!! Thank you Robert!

[@ricolauersdorf687]

I see, i need return path very close to each wire (depending on current and frequency of course). No i ask myself, what do i do with the wires connected to inductor in a dcdc unit (you call it the noise spot, its the part that changes its voltage very fast) As i understand, it is not good to have parasitic capacitance to these wire, but how then do i give it its return path?

[@eljeffo2871]

Awesome video! Thanks to both of you!

[@tHaH4x0r]

12:30 How about using batteries to supply the device under test, instead of the power supply? Wouldn't that be a very cheap way of getting a very cheap, low noise source? Very very cool video nontheless. Extremely useful, and helps people get good EMC design without access to some (expensive) testing facility. I am curious how mr. Zhang does EMI precompliance testing for radiated emissions, with a 'standard' TEM cell?

[@MachOneDesignEMC]

Hi, I also did an interview with Robert on troubleshooting radiated emissions after this interview, please check. You can also check my own EMC channel, just search engineers' guide to radiated emission tests! please provide feedback!

[@emcDV]

Nice! Very practical video for improving the EMC performance of a device.

[@KamleshGadhvana]

thank you very much very useful video

[@arpanrathore6826]

Thanks a lot sir.......really appreciate both of your work✌

[@sefakaraca2749]

using actually a capacitor array with different values isn't actually that useful? if we used it this way we can't be sure the noise will be filtered out effectively, can we?

[@fernandoi8958]

I would love to see a pcb redesign to show how layout alone could improve the EMC by a lot...

[@RomanAKozak]

Very interesting. Thanks for the simple explanations

[@cmtevega]

Another great video!! Congrats Robert!!

[@edwardsernas5089]

Amazing value shared with this video. Big thanks!

[@modulo-engineering]

Perfect for Beginners to learn more about it. thx Guys!!!