Are my Circuits ILLEGAL to use?!

In this video we will be having a look at three buck/boost converter boards built around the same IC, the TPS6302. One of these boards you can get for cheap from Asia. The other two ones are DIY designs I made myself. And the main question for this video is whether they pass EMC guidelines which are super important when you want to sell your board. We will be testing conductive and radiated EMC. And at the end I will show you how you can fix EMC problems of such boards. Let's get started!

Websites that were shown/used during the video:

Thanks to the Würth Elektronik for supporting this video.

0:00 EMC Problems?
1:32 Intro
1:54 EMC Measurements at Home?
2:56 Conductive EMC Tests
4:38 Conductive EMC Results
5:29 Radiated EMC Tests & Results
6:32 Legal to Sell?
7:02 Fixing EMC Problems
10:03 Verdict

[askLorandt]

Was a pleasure for me to assist you! 🤩

[wuerthelektronik]

It was a great pleasure to support and accompany you at our location in Munich!
The video turned out really cool! We are already looking forward to your next visit! :)

[Peter_S_]

I have been designing products for 35 years and I have always had excellent experiences with Würth Elektronik components. They are excellent.
In such a design as this DC-DC converter I would immediately switch to a 4 layer PCB, but I would use both layers 2 and 3 for ground with plenty of via stitching. I would also flood any unused space on layers 1 and 4 and stitch those to ground too. It's a crazy rabbit hole. Robert Feranec is an excellent resource for expert lectures on EMC and PCB design.

[reverse_engineered]

Here's a funny story. The office where I work was using several consumer-grade Wi-Fi routers as access points. We are in a research campus near a company that works with satellites. One day the CRTC (Canada's telecom regulator) showed up with a rather expensive RF spectrometer looking all over our building like they were witching for water. Turns out one of those AP's radios was putting out too much energy into adjacent frequencies and those frequencies happened to be what they were using to communicate with the satellites. Every time they had to turn their dish towards our building, they got a bunch of noise. Just goes to show how important it is to meet EMC requirements!

[mikeselectricstuff]

These boards are components, and as such are outside the scope of CE marking, so meaningless to say they are not legal.

[reedreamer]

Reducing EMI from power supplies is an art. Every switcher in my house has some kind of noise suppression. Over the years I've become adept at finding the cheapest and most effective solution. I do this, of course, to minimize noise to my ham radio receivers. The real question is, how to I get my crazy neighbor to let me fix his noisy military grade radio jammer power supply.

[kennener]

A couple of comments from an old PSU & EMI guy... Sometimes the worst source of EMI is the gate drive current spike inside the chip (assuming the MOSFETs are integrated), which typically needs to be dealt with caps at the chip supply pins. Also, a single uF-range cap isn't enough in most applications. I'll use tiny caps in the 330pF-3300pF range as close to the device pins as possible, with larger bulk caps a bit further away where there is more room. Remember, it is the high-frequency edges that matter most when it comes to suppressing EMI, and larger bulk caps don't do a great job at high frequencies.

[Nik]

Full time PCB designer for 10 years here. For EMC reduction, i would start at better layout. Its usually the cheapest solution when it comes to volume production. Component changes should be done if you really need it, since you can save money and space. Nowadays virtually every board is a multilayer board for this reason as well as complexity.
For 4 layer boards the "traditional" stackup of Signal - GND - Power - Signal is better than a 2 layer board, but its till quite bad. The main reason is that when you switch the signals from top to bottom, you also have to switch the reference layer, and you cant connect GND and Power for obvious reasons. There are some kludge ways to improve this, but its a lot better to go to a better stackup.
Now days what is recommended is a stackup is Signal + Power - GND - GND - Signal + Power. This is an amazing way to minimise radiated emissions and it really doesn't take all that much space.
As others in the comment section have pointed out Robert Feranec has great videos on this topic. My personal favorite are with Eric Bogatin and Rick Hartley.
Great video i loved it.

[RichUK]

I went to a Würth EMC expo (by invitation) at Duxford some years ago. Half a day listening to some very interesting theory and sales (genuinely....Würth really are good at this) and half a day looking at the aircraft in the Museum. Thanks Würth...a great day.

[thlchmst]

This kind of research and experimentation can solve a lot of problems in more complex or sensitive electronic devices. More videos like this, please.

[brylozketrzyn]

EMC sometimes feels like voodoo and certainly is very close to magic. A few things I learned over time: 1. Use fully shielded inductors whenever possible. 2. Inductor orientation matters: you want the start of a coil to be closest to switching element 3. Avoid long loops 4. C0G caps can really help with higher harmonics 5. Feed through caps are better than shunt ones

[parakart]

So good to see how much the industry has changed. I got into EMC testing over 30 years ago when the radiated testing was all done outside on an open area test site, normally with a manually tuned receiver system. So much easier these days with a screened room blocking out all of the extraneous noise.

[Sulentar]

Würth Elektronik München is just 1km away from me, so it was you the day i thought i saw you, now i feel bad for not stoping the car and meet you

[inventorkr]

Excellent explanation of the electromagnetic effect of electronic circuits.

[tommihommi]

at JLCPCB and similar manufacturers, there's almost no price difference between 4 and 2 layer boards, especially compared to shipping and assembly

so just always start with 4 layers for DIY projects unless it's a super duper simple design.
You make your life layouting easier, it'll work better for EMC as well as general performance, and you can make it fit into the size you want more easily.

The stackups of
PWR/DATA
GND
GND
PWR/DATA

or

PWR/DATA
GND
PWR/DATA
GND
are typical recommendations that work for many applications

[FamilyOfEyles]

I'm glad you've brought this subject to our attention; it's often ignored and causes headaches trying to work out why circuits are misbehaving.
A few years ago I was fortunate to be involved in setting up both types of EMC lab at my workplace - it was nice to see the equipment again in this video, it brought back some memories!

[TRONMAGNUM]

What a lucky guy! Glad you were able to share your visit with us.

[adfjasjhf]

I would love to see more collabs like this! Thanks for this!

[Rob_]

Würth not only deliver great components, the also share their knowledge about it. I love their books and "the Trilogy of magnetics" is one of the books I own and use a lot. We had this book at a previous job so as soon as I got a new job, I had to buy this book myself. I refused the gesture from my current boss to buy this with company money. The book is that good that I want to have it on my own bookshelf in my own workshop 😁
And yes: Würth components are not always the cheapest but the capacitor and inductor kits saved me that much development time that they are worth their money

[charlesurrea]

You also have to keep in mind the shape of the traces on the board themselves too.
I didn't realize it until I had gone into radio that 90° angles emit RF like hoses.