Passing Conducted Emissions With a Buck Regulator : EMC For Everyone #3
HTML-код
- Опубликовано: 28 янв 2022
- Passing Conducted Emissions With a Buck Regulator : EMC For Everyone #3
In the third video of the EMC series I take a filter board and place it before the buck convertor. This filter board allows me to test and finally get the board to easily pass conducted emissions testing.
-- Supplemental Video --
/ 61806521
-- Blog Post --
TBD
-- Evaluation Board Series --
• Custom Eval PCBs
-- EMC For Everyone Series --
• EMC For Everyone
-- Support me on Patreon! --
/ microtypeengineering
-- Discord Server --
/ discord
-- About MicroType Engineering --
MicroType Engineering is a full service electronics and mechanical product design firm. We offer turnkey support no matter how far along in the design process you are. We have full schematic capture, PCB design, firmware, mechanical design, and prototype/small-batch assembly services. Please reach out to learn more! microtype.io/ 
Наука
Its a great video. Your content has more information as compared to others.
I saw lot of videos on EMI/EMC from well know youtubers.
Everyone is talking about , How to reduce EMI/EMC from PCB , so that it can pass a TEST.
But NO ONE is discussing about : How to resolve PCB , which is not working properly due to
1. EMI/EMC from PCB.
2. due to power supply
3. EMI/EMC due to inductive load.
Thanks for taking the time to make such a nicely produced video. May I add some observations?
1) To expand on what you said, with only one LISN you are only able to measure differential mode emissions (+ve to -ve) and not common mode emissions (+ve and -ve together relative to a nearby metallic structure). All of the measurements you are making are for differential mode noise only so it isn't a surprise that only the differential mode pi filter components worked at reducing the emissions.
2) Putting a common mode choke in the filter board _will_ reduce the common mode noise, you just were unable to measure it with your single LISN setup.
3) Tekbox LISNs are confusing in that they provide a black terminal which is for the DC negative line, right? Not so, this is to connect to the metallic reference plane underneath your Device Under Test in a two LISN test setup. It is not meant to connect to the power supply line. The bond to the nearby metal plane should be made with as short a wire connection as possible.
4) The limits you've chosen from EN 55022 standard is for an AC mains port. There are no limits in EN 55022 for DC power port emissions, instead we would have to look at the generic standards such as EN 61000-6-3 instead. Limits for DC power ports are higher than AC mains. Class B DC power port levels are more like Class A mains conducted emissions limits.
5) The Tekbox LISN you are using is a 5uH part designed for automotive type measurements with a 5cm gap between metal plate and DUT (refer to standard EN 55025 / CISPR 25). The EN 55022 measurements relate to a 50uH two line LISN and a 80cm distance between DUT and floor (40cm between DUT and walls). Again, this doesn't mean that your test is of no use, it just means that you can't compare the results to the limit
6) There is no regulatory requirement to measure conducted emissions on a DC power port for the majority of equipment. The term "DC Power Port" has a very specific meaning in the context of EMC. Google search for "unit 3 compliance what is a dc power port" to read more.
7) This is exactly the kind of measurement you must make for automotive components, where there is a lot of concern with interference to on-vehicle radio receivers - see the CISPR 25 standard mentioned earlier. The kind of filter you are describing is essential to reduce these emissions to below the limits (again, different from EN 55022).
8) There is a small amount of leakage inductance with every common mode choke - the air gap between wires and core, the inductance of the loop area of the connecting wires - all this gives a small amount of inductance that appears as two uncoupled inductances, contributing to the differential mode attenuation of the filter. Typically this Lleak is around 0.5% to 1% of the value of the Lcm inductance.
I would be interested to see your follow up video when you get your second LISN. Keep up the good work.
All the best in EMC
James Pawson @ Unit 3 Compliance
This is amazing feedback, good work! The video is also great btw, very few videos about this. I was wondering about the use of a DC LISN myself. If the product is meant to be used in a product that is connected to AC mains, should it then not be tested against an AC mains LISN? EN55032 would be that.
@@EliasBakken I think that using a DC power LISN like this is fine if you are only making *comparative* A-to-B measurements (e.g. we've changed a component, what was the effect). You are right, for compliance measurements for AC mains (and widely distributed DC power buses like those around a server rack) the 50uH CISPR 16 LISN called up by EN 55032 would be required.
@@unit3compliance481 That makes sense, thanks for the clarification.
That was quite an eye-opener. Not all differential chokes help, Thank you for the amazing video!
In my opinion you should release the extended version to the public. You can have it on Patreon as exclusive for a couple of days, but then make it available everywhere. Thanks!
Thanks for the feedback, hope you enjoyed 😊
Will wait for the supplemental videos... Its really helpful
Please continue the EMC series...
Very good video!!!!! I'm a pro EE and loved this.
I love this video!! Thanks a lot for share this knowledge, gold for a engineer
Glad you enjoyed it!
Damn, you're so helpful! Love these videos!
Thank you alot 🧡
Very much enjoying watching the custom evaluation board series.
I'm a hobbyist, and wanted to work with GaNFETs first hand while the chip-shortage struck. Sourcing parts in my country is difficult at best of times, so it's not looking feasible in near future.
So, I've got a request... can you please have a look at LMG5200 and use it as a power stage for a simple buck converter? My original plan was to use an ESP32 as the multiphase controller (PWM at 4-8MHz), drive 4x LMG5200 as 4 phase power stage to form a high-power lab-bench supply (actually a precision load tester), which can be controlled & monitored wirelessly from other systems (including mobile app or even stream decks). Also wanted to make the solution as small as possible (by using smaller passives at high freq), as a stretch goal.
Feel free to take it up, if it seems interesting to you. All the best, if you do!
Good video, with good explanations. Just wondering if that is actually the correct test setup to have the DUT this close to the reference plate on the table. At least as far as i can remember, we always put the DUT at 10cm (~4in) above the reference plate for all conducted emission and conducted interference tests. I don't have the EN55022 at hand to check if this is actually the recommended setup for this test.
Great video. You have such an engineers personality haha (its a compliment) hahahahaha
Could that 200 KHz peak be coming from the powersupply itself? Or the electronic load? Running a "clean scan" without the step down should give a baseline for troubleshooting. Great video! Very interesting to see the effects of the filter.
Removing downstream caps would be interesting! Great video
Thanks! Noted.
Nice video...i have a issue with 210khz failing CE...but that's on AC power line which carries like 80A AC current and has power line communication on it...
interesting, removing caps from the main board would be interesting , also comparing the point when it gets bad with the datasheet recommendations would also be interesting
Great video! Could you cover any books you recommend and testing equipment?
Thanks a lot,
Could you please address the issue of EFT, and Surge
I searched a lot but can't find anyone who addressed this issue
For boards with multiple buck regulators do we need to use a pi filter at the input of each buck or can we have one pi filter for all the buck regulators connected to a single power source.
What software are you using for the scans?
What is as blue box connected to spectrum analyzer ? Amplifier ? Attenuator ?
Why the y axis is in dbuv?
Hint: On your input ferrite heating up - What do you suppose it's DC resistance is from the graph you have shown ? IE it's not heating up from V/Z- it's just DC resistance
Uh, not sure what you're getting at exactly? I think I said that it was heating up from the impedance. Impedance includes DC resistance.
@@MicroTypeEngineering heat is from dissipation of dc energy not from the ac related energy
I have a question I have a commercially built switch mode power supply that causes interference in the 1Mhz to 50 MHz range. It powers a Ham radio transceiver. Is there any way to quieten down the power supply either by adding filtering on the input or output or both without causing other problems? I have enjoyed the series so far on EMC compliance standards.
Yeah that's certainly an issue with switching regulators. Normally you add a multi stage LC filter on the output that then goes to the audio equipment.
I'd say the most important thing to look into is to make sure your filter design can handle the worst case power output of the supply with a margin of error. Sometimes adding too much capacitive load with a filter can cause even more issues (in rush current, ringing, ...). But thats hard to estimate or pretty much impossible to predict, especially with and off the shelf PSU, without doing some trial and error.
@@ProtonOne11 I think it will be down to some trial and error. I will try a pie filter on the output rated at least 30 amps giving me margin of error of 10amps. I will let you know how I got on at the moment I am in the middle of a workshop clean up and rearrangement.
Which one of the successful designs is more efficient?
Eh, neither one likely effects efficiency significantly. I'd guess the PI filter is a little bit more so, but not much.
@@MicroTypeEngineering Good to know! My mantra is no diodes, no resistors > 0.01 ohm in series and minimize inductors whenever possible so this was of particular interest for optimizing efficiency while hitting regulations.
Ok so you understand a lot about circuit. Great video I need help it's just a statement . Ok I learn how to make a hall effect circuit to drive a fan simple it get too hot and the irf9540n burn's up it's rate for 23 amps 100 volt. I need something higher on all points. If you can help i will pay you for the information and i'll have to test it first thanks I hope you Will response It has a greater purpose than a fan Thanks
I'd appreciate it if you got through some EMI content at some point. There's a lot of stuff out there on EMC, but it always seems to stop short of discussing EMI. (And while the FCC rules for the US focus mostly on EMC, the CE rules for Europe seem to have a big EMI component.)
Uh, can you define your definition of EMC/EMI. I think yours differs from the primary one!
@@MicroTypeEngineering Yeah, I might have my acronyms mixed up. I'm thinking about immunity testing, which is something that I think is part of CE compliance for Europe.
Common-Mode Chokes are almost always a bad idea for reducing Conducted Emissions on a classic +/- DC design where the "negative" connects to your GND plane. Many EMC Experts have debunked this time and time again. Mode Conversion is what causes Common-Mode Noise to radiate from a differential power or data line like this, and a Choke will not help Mode Conversion. It's companies like Wurth who sell these chokes who make it seem necessary in design guides and application notes. I am a huge fan of Wurth, but unfortunately, this is just misinformation they perpetuate.
Yeah, honestly, once I can figure out a good few methods to force CM currents on a test board, I'm definitely going to start testing chokes more on this series. Honestly, I normally would just throw one on most boards. Definitely going to double check that from now on. Thanks for your thoughts!