Here is the rewritten article:

As I was testing my tiny design board in the EMC lab, I noticed that it didn't pass the radiated EMC test due to its long connector cables preventing ringing. But anyway, let's start by saying that without a load, all boards passed the EMC test. However, as soon as we went up to the max output current, things started to look pretty terrible and no board passed the test.

So does that mean that none of these boards are legal to sell? Well, there is actually a loophole with such circuits that says that as long as the board passes EMC tests in its normal powered on state, it is perfectly fine to sell. Because if you want to draw more current and push it over the EMC max limits, then this is your problem as the consumer and you are responsible to not influence others negatively.

And with that mystery solved, let's ask the question whether my tiny design here, which of course performed the worst from all 3 boards, could be improved in a way that it passes the EMC test? I of course asked Lorandt this question and he said the first solution is all about optimizing components meaning you add more or better capacitors to basically lower the peak current. The reason is that this current basically uses the copper traces as antennas to broadcast its interferences, so it makes sense to keep that as low as possible.

If you want to try that out yourself then feel free to visit Red Expert from Würth Elektronik which is a calculation tool where you can exactly see what components make what difference. In my case though we only got problems at the switching frequency harmonics meaning we need input and output filters to deal with those. And yes, we need them at the input and output because we got a buck/boost converter here which in buck mode makes trouble at the input and in boost mode makes trouble at the output side.

Truth be told, finding a suitable filter is a piece of cake for my EMC expert, but time was working against him because we only had 30 minutes left in the EMC lab. So he probably went a bit overboard with this filter setup here which uses a common mode choke on the input and a pi filter on the output that only lets low frequencies pass. But as you can see, these filters did the trick and the board passed the radiated EMC test, lovely.

But let's say both solutions didn't work, what else can we do? Well, the next step would be to improve the PCB design. For example you want your traces as short as possible especially where you got high current ripples, so that they don't act like big antennas. Or we can also switch to a 4 Layer PCB instead of a 2 layer design.

The idea is that you have the upper and lower layer to do the routing and use layer 2 for GND and layer 3 for the supply voltage. This way you got one big GND copper layer that you can easily reach through vias to reduce antenna designs. I actually wanted to give this solution a try and came up with my own 4 layer design that is my smallest one yet.

But the sad news is that it simply didn't want to work and I had no time to fix it before visiting the EMC lab. So yeah, that sadly didn't work out but at least you are now familiar with the 3 big solutions when it comes to EMC problems.

And I know that this topic is not super easy to digest which is why I placed some links in the video description where you can learn more about it. And with that being said I hope you enjoyed this video as much as I enjoyed having fun in an EMC lab for a whole day. As always don't forget to like, share, subscribe and hit the notification bell. Stay creative and I will see you next time.

WEBVTTKind: captionsLanguage: enNot Again!It is not hard to  realize that this voltage converter here that powers my audio amp has some serious problems  with EMC aka electromagnetic compatibility.EMC is basically the ability of a device to  interfere with the functionality of other devices which this converter definitely does with the  audio amp through its wired connection.Another example would be the low power  WiFi jammer I accidentally created in a previous video; but there the interference  was mainly done through radiation.Summarized we can say that EMC is  super important and that is why there are guidelines in place you have to follow if  you want to sell your circuit boards.But I often wonder if these EMC  guidelines were truly checked when it comes to cheap boards from Asia. For example I got this super useful battery powered buck/boost converter here that is  so useful for your portable projects, that I even redesigned it and sized down in a previous video. And now it is finally time to put them all to the test and see which one is actually compliant  with EMC guidelines and if they are not; can we fix them in hindsight to make them compliant? Let's get started!This video is sponsored  by Würth Elektronik who not only make some of the best power electronics on the market;  but are also experts when it comes to EMC problems. But you will see that very soon. Now first off, when it comes to conductive aka wired EMC measurements, then you need quite a  bit of specialized equipment to do that at home.And when it comes to radiated ones, then there  is now way you can do that at home because you need a room that shields and absorbs all other  radiated waves along with a very unique probe.That is why I visited Würth Elektroniks innovation  center in Munich which apparently not only comes with awesome office spaces; but also with pretty  unique labs, tons of specialized equipment and most importantly an underground EMC lab. There they got one room for conductive EMC measurements  and one for radiated EMC measurements.And luckily for me, I also got help from  Lorandt who is an EMC specialist that even got his own video series; if you want to check that out. And with that being said, we started off with the conductive EMC measurements and since we  had the lab for one day, we only focused on the Asia Board, my initial smaller DIY  design and my smallest DIY design.Now for the measurements we for one  pushed the boards to the limit by drawing the max current on the output and then did  the measurements again with the load removed.This is important because when the max current  is flowing, we are dealing with the maximum magnetic fields and when no current is flowing,  we deal with the maximum electric fields.And before I get to the results of this  conductive EMC test, let me clarify that even though the boards usually get powered  by a battery and not the power grid, these results are still super important.For example if you got one big battery pack like in an electric vehicle, the conductive  EMC interferences created by the buck/boost converter would basically spread to all attached  components and that can cause lots of problems.And with that out of the way, we got closer to  finishing the measurements by using a LISN on the boards input which is basically a model of our  power grids impedance across which we can measure and observe the current spectrum that gets drawn. This spectrum tells us what current harmonics get created and according to EMC guidelines, they are  simply not allowed to exceed a certain max value.And surprisingly none of the boards did that;  no matter if they boosted the input voltage up to 5V or reduced it down to 3.3V in buck mode. But if we want to go into detail; then my tiny version performed the worst,  followed by the Asia Board and the best was my bigger DIY design;  probably due to the fact that it comes with more capacitors. By the way these current harmonics come with a base frequency of around 2.4MHz which is of course  the switching frequency of my converter IC and thus obviously the main culprit for all possible  EMC problems, that we sadly can not get rid of because that is how switching converters work. But so far everything looked great and thus we moved on to the radiation tests  which were a bit more spectacular to look at.I mean you got this big antenna here that even  turns in the middle of testing to catch the horizontal and vertical radiations. This antenna is by the way exactly 3m away from the tested circuit  which of course also gets hooked up with standardized wires that are 80cm long. Yeah everything here has a purpose, even the big capacitor we added to the input  of the circuits that basically prevents ringing due to the long connector cables. But anyway; next we did all radiation measurements and let's start off by saying that without a load,  all boards passed the EMC test.But as soon as we went up to the  max output current, things started to look pretty terrible and no board passed the test. So does that mean that none of these boards are legal to sell? Well, there is actually a loophole with such circuits that says that as long as the board  passes EMC tests in its normal powered on state, it is perfectly fine to sell. Because if you want to draw more current and push it over the EMC max, limits then  this is your problem as the consumer and you are responsible to not influence others negatively. And with that mystery solved; let's ask the question whether my tiny design here,  which of course performed the worst from all 3 boards,  could be improved in a way that it passes the EMC test? I of course asked Lorandt this question and he said the first solution is all about optimizing  components meaning you add more or better capacitors to basically lower the peak current. The reason is that this current basically uses the copper traces as antennas to  broadcast its interferences, so it makes sense to keep that as low as possible. If you want to try that out yourself then feel free to visit Red Expert from Würth Elektronik  which is a calculation tool where you can exactly see what components make what difference.In my case though we only got problems at the switching frequency harmonics meaning we need  input and output filters to deal with those.And yes, we need them at the input and output  because we got a buck/boost converter here which in buck mode makes trouble at the input and in  boost mode makes trouble at the output side.And truth be told, finding a suitable filter  is a piece of cake for my EMC expert, but time was working against him because we  only had 30 minutes left in the EMC lab.So he probably went a bit overboard with  this filter setup here which uses a common mode choke on the input and a pi filter on the output  that only lets low frequencies pass.But as you can see, these filters did  the trick and the board passed the radiated EMC test, lovely. But let's say both solutions didn't work, what else can we do? Well, the next step would be to improve the PCB design. For example you want your traces as short as possible especially where you got high current  ripples, so that they don't act like big antennas.Or we can also switch to a 4 Layer  PCB instead of a 2 layer design.Here the idea is that you have the upper  and lower layer to do the routing and use layer 2 for GND and layer 3 for the supply voltage. This way you got one big GND copper layer that you can easily reach  through vias to reduce antenna designs.I actually wanted to give this solution  a try and came up with my own 4 layer design that is my smallest one yet. But the sad news is that it simply didn't wanted to work and I had no  time to fix it before visiting the EMC lab.So yeah, that sadly didn't work out but at  least you are now familiar with the 3 big solutions when it comes to EMC problems. And I know that this topic is not super easy to digest which is why I placed  some links in the video description where you can learn more about it. And with that being said I hope you enjoyed this video as much as I enjoyed  having fun in an EMC lab for a whole day.As always don't forget to like, share,  subscribe and hit the notification bell.Stay creative and I will see you next time.