 As the MC said, I'm here to try and turn EMC, what is traditionally a quite esoteric and boring and black magic-esque subject into something that's perhaps a little bit more understandable and a little bit more interesting and with some audience participation at the end, a little bit more fun. So let's see. You can if you want. So what we're going to talk about is electromagnetic compatibility testing, radio testing, other compliance, generally all of those regulatory hurdles that you need to overcome if you actually want to sell anything because we're all in the business of making things. At some point, you want to set it free in the wide world and that involves exchanging it for real money. So how do we go about this? We're just going to try and keep it narrow to the scope of Europe. If we start talking about worldwide compliance, that's like an all day course. So we're going to go there. Just so I can get a flavour for who I'm speaking to because I haven't had the pleasure of meeting many of you before. Who here has heard of EMC before? Can you have a few hands? Okay, good. You put your hand up if you've done any EMC testing or you've been to an EMC test lab before. Okay, quite a few. How many of those people have had EMC problems at the test lab? Good, good. It's not just me then. So this is me. I've been electronics engineer for far too long and an EMC specialist for far too long as well. About back end of last year, I got made redundant from my previous company and they didn't want their EMC chamber. So with my hard earned redundancy money, I bought it and all the EMC test lab assets and now I run my own business. It's big, it's scary, but it's also very rewarding and fun. There's no way I could have got a pass out to do this from my previous job. So now that I'm the boss, I make the decisions and it's much more interesting. I get involved in really wide range of stuff. My speciality is EMC problem solving. So usually people have radiated emissions problems and they need some help to fix it. I also get involved in radio testing and a bit of safety testing for the low voltage directive, sea marking, environmental and vibration testing, really wide range of stuff. So I guess the question, a good question to ask is why. It's what my children say to me, but why dad? Why do we have to do this? Why do we have to do EMC? It's a very good question. The problem is that the wireless infrastructure, we're so dependent on it. It is a critical infrastructure. Let's all imagine that our phones, Wi-Fi, nothing else works. Just imagine that for a second. It would be fantastic. You could get so much work done, have really valuable human interactions. Yeah, okay. So it's better that we don't actually try and break the wireless infrastructure. And radio interference is one to many. So one device, one faulty laptop could wipe out all of the wireless infrastructure in a small area because it's one to many. So really, us as hardware engineers, we're on the front line of protecting the RF spectrum. It's like we need some kind of superpower under our shirts. So inevitably, without some kind of regulation, you can't have spectrum coexistence. If everyone was left to do their own thing, it would be chaos. It would be anarchy. Some people like anarchy. The European Union do not like anarchy. So we have a thing. Let's say, for example, our thing has a radio in it. We would like to exchange our thing for some cash money, and we would like to do this in Europe. Okay? What do we need to do? We need to have a CE mark on it. The CE mark is that magical mark that gets you in. It says that you've conducted this assessment against all this huge list of directives that the European Union has. The assessment is mandatory. It doesn't say how you have to conduct it. You just have to do one. Okay? That's an important point. The directives that we're most interested in. If you're making something with a radio in, it's fairly simple. This is either a radio receiver, a radio transmitter, a combination of the two. You have to comply with the radio equipment directive. This was new in 2014. It replaced some really old clunky directives that weren't very nice. If you don't have a radio, that's fine. It's just that EMC directive applies. So you can see how we've got a fairly common framework between the two. The radio equipment directive, as we'll see, does call up the EMC test. So you have this idea of the radio equipment directive has essential requirements. These are the short paragraphs that say you must not interfere with radio services. You must play nicely with the spectrum. The EMC part says that you can't interfere with things and interference can't hurt you. And the safety is you're not supposed to kill, hurt, damage may many people or livestock or animals or whatever. So it's a fairly straightforward thing to think about. How do we demonstrate that we meet these essential requirements? The essential requirements is just like a paragraph saying thou shalt not. So there are what are called harmonised standards. Huge lists, lists and lists and lists on the European Union website of these harmonised standards. These give you the meat and veg of how to comply with a standard. So you'll have ones that say, this is an example for Laura. So you'll have a radio standard that says you've got to transmit this kind of chirp, this kind of timing, this kind of duty cycle. You'll have EMC standards that say your radio should be doing this when you make the test. And you'll have a safety standard that's 324 pages of fun. And if you apply the harmonised standards, you get an auto pass, you get a presumption of conformity. That's effectively an auto pass. You don't have to apply harmonised standards. You can dig up some crusty old standard from the 1990s, if you want to. And argue, if you sufficiently argue your case then, yeah, you can apply it, but it's not the auto pass and you'll get questioned about it should anyone decides to go through your documentation, which we'll come to later. Everyone likes documentation, right? So the harmonised standards then call up basic standards. So you can see we've got a hierarchical model. We've got requirements and then sub-requirements and the basic standards actually tell you how to do the test. Connect this lead to here, turn it on, set the dial to 10. Does it blow up? Yes or no? So that's how the structure of the standards works. It's actually quite simple. The hard bit is following it through the chain and all the hops and sometimes it goes back up into another directive and then back down again. It can get quite complicated. So if we take each of the, those three stacks, the radio, EMC and the safety and just take a quick look at each one and have a look at what sort of tests, what sort of things we might need to do in each of those. If you're wanting to put a radio into something, by far and away, the easiest way to do it is a pre-approved radio module. These are commoner garden. You can get these for anything from a couple of dollars up to, you know, a few tens of dollars, depending on how esoteric the radio is. They're fantastic. All the tests are done. Some nice Chinese laboratory has performed all these tests, hopefully to a high quality. It minimises your risks. You don't then have to spend lots of time developing it. It just makes your life a lot easier. There's a large marketplace. There are so many. You go on Alibaba or something like that. Look for a Wi-Fi module. The hard part is choosing which one you want to buy. They come with downsides. So they're more expensive because someone has taken it to a test lab. The test lab has charged them many thousands of pounds. So they have to pass that cost on. It still doesn't exempt you from responsibility. By integrating this module into your product, you're responsible for the module. So you have to source a module that's of a good quality. You can't just pick the cheapest one that you find and expect it to pass. The chances are it won't, the label of cooked corners. There's a maxim in EMC test world that CE plus CE does not equal CE. You can take two C approved components. Let's say we have a motherboard and a graphics card. You can put them together. That does not automatically mean that that system then complies with all of the requirements for CE mark. You might have, by doing that unintentionally, you might be radiating some noise at some clock frequency that you didn't expect. So you have to test at the sort of the module level. That's the key thing there. You still need the radio test report. Just because the manufacturer says it's CE marked, do you have the radio test report? What does it say in it? Does it actually say that it's passed? That radio report forms part of the proof that you need to provide, should anyone come asking, that your product is actually okay. The manufacturer should also, if they're a conscientious manufacturer, they should also give you installation and use recommendations. So they should say, put it here, don't run these traces near it. Your antenna should be like this and mount your antenna like this, especially if they have an internal antenna on the board you would expect to have some kind of installation guidelines. And if indeed they are FCC approved, they have to supply that with the module. Of course, you don't need to buy one. Just design your own. You can if you want to, but I don't recommend it. This is a list of tests for Laura that's the 300 220-2 standard calls up. There's quite a few there. And a lot of those tests take a while to set up, take a while to run. You'll need to run them on all the different RF channels and different transmit modes and different chirp timings and duty cycles and min max temperature and supply voltage. And all of this, a lab that charges you a thousand pounds a day. Wow, so that's going to take a while. To give you some context, the company that I used to work for, we developed a really nice little video streaming box. It was a really good product. It had a Wi-Fi module in there. So it was a four by four, five giga hertz MIMO module. It was a pretty high end. We also had another Wi-Fi module and a Bluetooth module in there. And the quotations we got for testing all of those to the relevant standards for America and Canada, 56,000 pounds. And that was just the lab time. Not counting the time spent for our engineer shuffling backwards and forward to the lab when the lab actually turned out to be kind of what's the word I'm looking for? Incompetent, that's the word. So realistically it came out more like a hundred thousand pounds for two wireless modules. Okay, so if I were you, pre-approved modules, it's definitely the way forward. Yeah, that's pretty much what I said. And you can't underestimate how much resource and time and equipment, something like that, will take as well. Because it will suck your resource into a black hole and you'll never get it back again. So think on. EMC, does anyone know how to spell EMC? No one, okay. You missed a trick there. So we're interested in how our thing exists within the electromagnetic spectrum. We're concerned about what it might emit. How much noise does it give out? Is it going to interfere with things? We're also concerned about how immune it is to things. What happens if I walk across a nice nylon carpet and touch it and give it a static shock? Does it reset the unit? Does the unit continue to work fine? What happens if Derek is running one of his very fine Tesla coils nearby, game over for the unit? The typical example is a mobile phone and a radio. I take my mobile phone, I put it next to a radio. What happens? Anyone? That's in part a failure of the radio. It's not sufficiently immune to the out-of-band emissions coming from the mobile phone. It's arguably also a fault of the environment. The mobile phone is too close to the radio. The radio hasn't been designed to withstand the 10, 15 volts per metre that the mobile phone is giving to the radio. It's fairly easy to fix. I just move the phone away from the radio, or I can turn the radio off. But what if you can't do that? What if your next-door neighbour has, I don't know, what if he has a GSM repeater and his badly-ginstalled GSM repeater is degrading all of the cellular capability in the area? That's not so easy to fix. We also got to think about what environment we're going to put in. Most products are going to be domestic products, so they're going to be going into domestic or commercial premises. The EMC requirements are fairly gentle. Industrial, all the levels go up again. Everything becomes harder. You might be allowed to emit a little bit more noise, but the immunity requirements are much higher as well. Add one on to that. You get a medical, industrial, automotive, and military is somewhere on the roof in terms of the requirements. So if you're trying to get a military product through, then well, good luck, you're a better man or woman than me. And the last point is you must assess the product. That's the only real legal part of the standard says you must assess the product. You can't just say, you know, an assessment might be, well, it's just an LED and a battery, so it's clearly okay. An assessment might be a 300-page technical document. This graphic just serves to illustrate emissions and immunity and how we might think about them. We have an EUT, our equipment under test. You'll see that acronym quite a lot when you do EMC testing. It has ports with cables on. You might, on the immunity side, you might be having a, you've got a transmitter in an antenna. It's broadcasting radio waves. Some of it's being picked up on the module, the enclosure. Some of it's being picked up on the cables. You might have conducted noise coming from something noisy coming down the cable. And vice versa, you might have noise being emitted from the unit either in radio wave form or being conducted down the cable. These are the kind of things that we're trying to deal with when we do testing. Some problems that you can get, these are all, apart from the banana skin, these are all ones that I've been doing recently. Top left, we have a, what's left of a power supply after a lightning surge simulation. It didn't take too kindly to it. I've got the video, but it has me saying a naughty word in it and there are small people present so I can't play that. The bottom left hand, which is kind of difficult to see here, is actually some noise breakthrough on a satellite amplifier that was not very well designed with a shielding. As you sweep noise through the spectrum broadcasting noise at it, it actually leaks through. So you can see there's a wanted carrier here, but there's actually, on the right hand side of that spectrum analyzer plot, there's noise starting to break through, which is a failure. On the top right, we see noise, this is a one wire interface, which is actually surprisingly vulnerable to EMC. So, you know, you want to consider if you're going to get a one wire interface on the end of a long cable, my recommendation would be don't or use a very well screened cable because it will be affected. Bottom right, this is just a modification that I've had to do to a customer's product, adding a small copper ground plane under a particularly noisy node on a switch mode power supply. So these are all some of the typical problems that you might see when you're doing EMC testing. The banana skin, there is one of my fellow consultants, a guy called Keith Armstrong, has written chapter verse and more books and articles than I would ever do. He collects EMC stories. Every time there's an EMC problem in the news, he files it away on his website. I think it's emcstandards.co.uk. If you search for EMC banana skins on Google, you'll find it. It's fantastic reading, all of these things that have gone wrong. I think my personal favourite and it's a favourite because the guy actually survived. Sunday league football match somewhere, I think it was somewhere in Norfolk and two Apache helicopters flew over and the radar signal from the Apache helicopters was enough to trigger this guy's pacemaker and he fell over and thankfully there were some first aiders on hand and he survived. So that's the sort of the high end of what can go wrong when you have an EMC problem. So yeah, check out EMC banana skins. That's it makes a very good reading. So some other EMC problems you might come across, especially if you're integrating a radio into something. You can get intra system or what we call platform level interference and this is where the device is the circuit that you actually have designed is interfering with your own radio module. This is a real pig to find. It is really, really tricky going back to the previous company I worked for we spent weeks trying to find out why certain wi-fi bands had pretty much no range and it turned out to be the 24th, 25th harmonic of the HDMI interface was sitting right in band of this wi-fi carrier, the 24th harmonic. That's just it's way up there and there's practically no energy in it but because it was so close on the board it was coupling in. It didn't show up on the EMC test. We had to get a 150,000 pound real-time spectrum analyser and do some sniffing around with wi-fi antenna to find out where it was coming from. Nightmare. You can get around it. You can practice low noise design. You can apply lots of shielding to things. That's probably what we should have done but it was a consumer product so we were being tight and not spending any money on it. You can practice low noise design. You can sniff around. It's tricky to find. Half the time you might not even realise you have a problem. It's only when you're pushing the envelope of data rates that you might come across it. You might have antenna issues so you might have local structures detuning the antenna or coupling noise into the antenna, decentralising your preamplifier and your receiver and there's also the raising of the noise floor. The more electronic devices we have, the higher the noise floor gets. I mean just in this room everyone's got a phone, there are a few laptops, there's a camera. You can feel like the noise floor creeping up. The denser or electro dense we get as a society. This problem is only going to get worse. You know there's the typical my favourite trope of the Christmas lights interfered with my wi-fi which you often see in some of the lower budget newspapers every Christmas. It does happen. I tested a product from a customer that had those little sticky LED strips that you can buy with like a remote control and they blink all different colours. How a small strip of LEDs can create so much broadband noise was staggering. Both I and the customer, he was unpleasantly surprised. I was pleasantly surprised because it meant I got to help him out and earn a few shekels which is always nice. All of this can be avoided to an extent. You can follow good EMC practice. You really have to consider it at the start of a project. Unfortunately most people think about EMC at the end of a project. I've designed this really cool thing and it's got buttons and bells and I need to ship it and I've got this EMC. What's EMC are? I better go to the test lab and then they find that actually they really should have thought about it at the start because their circuit board might be a mess. You have to design EMC in. You have to look at a circuit and think right okay when we're going to do this test we're going to be injecting noise onto this port what can we do to do that? Well we could put some shielding on it or we could add components to it to help with that. You need to address it sooner not later because the cost of doing a spin of a circuit board and the time delay especially if you're time critical you know you could really wreck a project. If you've got problems you can look for help. There are a lot of test labs out there. There are a lot of consultants out there. There are some consultants with test labs who are very helpful. Hello that's me. If you're time critical you need to test early and try and find as many problems as you can. If you're time and money critical well good luck to you sir. I wish you all the best. A quick word about safety, electrical safety. The radio equipment directive says you have to be safe. Normally with the low voltage directive which says what we have to do regarding safety you only apply it on things that are above a certain voltage. I think it's 75 volts AC or something something in that band. The minute you put a radio in it doesn't matter if you're powered by a 1.5 volt battery all the safety requirements apply to you. There might not be many but they apply. So we're looking at electric shock, thermal burns, radiation, RF exposure. You need to consider RF exposure if you're making something with a radio in that is being worn or carried like a mobile phone or a microphone get it off, get it off. It has to comply with specific absorption rate testing so they actually strap this device to the side of a dummy filled with what's called vitreous humor which basically mimics the dielectric constant of your head and then sweep an antenna around inside it to try and map out the emissions and the amount of power that your head or body part will absorb. So a quick wrap up since I've just been given my five minute warning. Once we've done all of our testing we need to compile a technical file. You need to keep your technical file for 10 years after you stop selling the product. That's a long time. Not many people have set up to think 10 years in the future so I think you've got to think about that. A declaration of conformity that's your legal document. That is where you as the manufacturer take responsibility for your product. You say this is my product. I take responsibility for it. I've done all of the tests. You don't get that document right. That's where people are going to start coming and sniffing around your technical file and wanting to know what tests you've done. The key thing is the manufacturer is responsible. This is a very important point. You can go to any test lab, you can go to five test labs, ten test labs, you can have all the certificates, you can have the certificates coming out of your trousers. You're still responsible. The test lab will just hold their hands up and say we did the test. We did the test on the unit that was provided. It's not our problem if it fails. You're responsible. You should do a risk assessment. Not many people do. They're not that common for EMC. They are for safety. You need to control the production to make sure that the unit that you tested is like all the other units that are rolling off your production lines. You need to make sure that you use the same components. Build them exactly the same way. You need to think about the ROS directive. So you're not putting too many LEDs, cadmiums, mercury, hexavalent chromiums, all these nasty chemicals into the supply chain. You need to think about your Wii directive, whether you're recycling things properly, whether you're painting your tubes or not. After all that, you can finally apply the CE mark. Actually, you did that earlier because you had to send the label off to the printers, right? So again, all we're trying to do here is mitigate risk. All the testing and everything we do is about reducing our risk. Some people are more comfortable with more risk. Some people are very anti-risk. There's a really wide spectrum. If you're making something that's like a car, a plane, a train, something that's life critical, you want to be fairly risk-averse. If you're making something that's got a battery in an LED that blinks on and off, you can probably afford to be fairly risk-tolerant. Obviously, anything electrical safety, that but times 10 in the conservative direction. There is market enforcement. Trading standards will look into complaints. It's very much a complaint- driven architecture. Germany have got a particularly harsh way of looking at things. They are used to have a finder's fee, so if you thought that someone wasn't compliant, you would shop them and claim a bounty. I don't know if they still run that scheme or not. The FCC have recently taken action against a company called Hobby King for making non-compliant Wi-Fi transmitters. They proposed a small fine of 2.8 million dollars. Your competition will shop you in. If you're in a competitive market, they will say, he's not compliant with this, so you should investigate him. Those in glass houses shouldn't throw stones. The EU conducts market surveillance every year. They mostly go for low-hanging fruit widely used equipment. What are the results? This is where it's time for a bit of audience participation because it's time for play your compliance cards right. I hope we've all got our lungs warmed up. No peaking at the laptop if you look because you can see the answers. 2011, the EU looked at LED lighting. Let's give you a start. Let's assume that everyone passes about 50% at the time. Do we think it was higher or lower than 50%? This is your big chance. Higher or lower ladies and gentlemen, lower. You're quite correct. 17% of the LED lights were compliant with emissions and documentation. 17%. That's not a lot. Okay, what have we got next? We've got laptop power supplies. Any guesses? Is it higher or lower than 17%? Lower. It's higher. Well done, the gentleman in the front. Okay, 24% though. One in four, that's not very good. Solar inverters. Higher or lower. Lower. Lower. We think lower. 9%. Ouch. Then the next one is... I'm not going to put it up yet because I don't want anyone peaking drones. Okay. Come on. 9%. Sorry. Negative numbers. Yes, they're taking drones off the market on purpose. Anyone want to guess? Higher or lower for drones? Higher. 8%. Lower. Oh, that's terrible. Even Brucey is disappointed in the drone manufacturers. All right, last one. It's a nice easy one. Induction cooking plates. Are they going to be lower than 8% I wonder? What do you think? Higher lower? Higher. Yeah. We finally got above 50% compliance. Well done, induction cooking plate manufacturers. You have saved the day. So these are actual figures from the EU market surveillance. Actual figures. So that means that even for fairly simple appliances, generally speaking, people are getting it wrong and widely wrong as well. So even a CE mark. These are CE marked products. So thank you very much for your attention. This is my shameless plug. You've just given you a brief intro to what I do on a daily basis. These are the things I help you with. You find me on LinkedIn or my website and I'll be around at the tea break if you want to chat and ask questions.