Self-Certification of consumer electronics

[ricko_uk]

Hi,
if you want to make and sell simple low voltage low power electronics (like a small PCB with display, keypad etc) for consumer use (not just hobbyist).

The PSU I buy in already with all certifications. What about the PCB/enclosure assembly?

1) what are the minimum requirements and certifications you need to issue/get? Which ones for UK/Europe and which ones for US?
2) If any can you self certify them?
3) I think you can self certify the CE mark, what do you need to do to be able to self certify it and be able to stick the CE mark on it?

Many thanks

[AndyC_772]

You need to have evidence that your product is compliant, so that if challenged by the authorities (Trading Standards in the UK), you can show that you were justified in applying the CE mark. You can self-certify, but that doesn't mean just sticking the mark on the back of the box and calling it a day.

As far as safety is concerned, if you're using a mains PSU that's already approved, you may be able to claim compliance by inspection.

EMC is much harder, though. Without testing the complete product, I've yet to hear any possible valid argument that doesn't involve at least some testing at an EMC lab. Most commercially available power supplies are pretty marginal in terms of emissions even on their own, never mind when connected to something that's not just a pure resistive load.

Pragmatically, the chances of your product ever being picked up and questioned by the authorities are very small indeed, unless someone gets hurt, but that's not really a good reason to join the list of cowboys who just ignore the rules.

[Gyro]

If you're buying in a power supply (you don't indicate open frame or wall wart so I'll assume wall wart), then you will at least want copies of their certifications for your Technical Construction File, and have sufficient confidence in their validity, ie. that they haven't self certified. They should have used a reputable test house, particularly for LVD (safety) testing. You can't just switch PSUs on the fly either, as that would affect your overall certification. If it is any sort of open frame PSU then you become responsible for LVD too, its safe enclosing, components, wiring etc.

With LVD taken care of, then yes, you have EMC to worry about. You need to purchase the relevant standard for the equipment that you are building (there are no 'freeware' copies) and verify compliance. On some equipment, you can do the sums and prove that you can't possibly emit sufficient RF energy to exceed the limits, but that certainly won't be the case for a mains block powered device.

You really need to do radiated and conducted emissions, using your chosen mains block. You may also need to do susceptibility and ESD too, depending on what the function the device is and the consequences of it failing. (ESD, you can probably do yourself to a reasonable level of confidence - you probably want to do this anyway if you want to be confident about not getting lots of customer returns).

You didn't say if you possess any EMC pre-compliance test equipment, but without it, I can't see that you would have the evidence needed to attempt self-certification.

P.S. Just to try to add a few specific answers to your questions:

1) There are laid out in the specific standards, based on product type. At the moment, UK and EU standards are still harmonized (according to the BS-EN numbers on the BSI site anyway). US is based on FCC standards. For example, you can find the UK EMC standards at https://shop.bsigroup.com/Browse-by-Sector/Manufacturing/Electromagnetic-compatibility/Electromagnetic_compatibility_standards/.

2) Yes you can self certify, given sufficient evidence that you comply with the relevant standards.

3) The method of self-certifying is by creating and maintaining a Technical Construction File (TCF), which documents the evidence that supports your claim to comply with the relevant standards. This must be available for examination in case of a challenge.

Basically, the test / compliance requirements for your (quite specific) product type comprise EMC (Emmissions - Radiated and Conducted, Susceptibility - Radiated and Conducted, and ESD) and LVD (Low Voltage Directive - Safety). Talking to a friendly local test house can save you an awful lot of time - Consultants (of any description) tend to be expensive!

[ricko_uk]

Thank you Chris and AndyC_772 for your very detailed replies, much appreciated!!

[splin]

This thread about adafruit's certification (or lack of)  is interesting although mostly about their radio products. It seems that Sparkfun don't certify (most)  of their products either but they both seem to get away with selling a lot of stuff:

https://forums.adafruit.com/viewtopic.php?f=8&t=113747

I believe that there is a *lot* of stuff sold which hasn't been expensively tested by an RF test company, by small companies - but obviously not many want to talk about it. How can $25 products, such as test gear sold to hobbyists, which might only sell in the few hundreds at most, cover $10K+ EMI testing cost and still make a profit?

There must be quite a few posters here who sell there own designs.  It would be interesting to have an anonymous poll to see how many do and how many would admit to selling non certified product or product which hasn't been third party emissions tested. After all the chances that a small voltage reference module for calibrating DMMs or an Arduino temperature sensor shield would cause real world problems must be vanishingly small although theoretically possible if marginally stable and oscillate in some conditions. Is Dave's uCurrent CE certified for example?

I wonder if anyone ever gets prosecuted or at least told to stop selling product unless dangerous or it grossly violates emissions regulations and causes real problems? I'm talking about small companies selling their own designs here - not re-sellers of cheap, poor quality Chinese tat or larger companies selling large volumes of consumer product.

[Someone]

I believe that there is a *lot* of stuff sold which hasn't been expensively tested by an RF test company, by small companies - but obviously not many want to talk about it. How can $25 products, such as test gear sold to hobbyists, which might only sell in the few hundreds at most, cover $10K+ EMI testing cost and still make a profit?

There must be quite a few posters here who sell there own designs.  It would be interesting to have an anonymous poll to see how many do and how many would admit to selling non certified product or product which hasn't been third party emissions tested. After all the chances that a small voltage reference module for calibrating DMMs or an Arduino temperature sensor shield would cause real world problems must be vanishingly small although theoretically possible if marginally stable and oscillate in some conditions. Is Dave's uCurrent CE certified for example?

This space has exemptions in FCC part 15 subpart B (see attachment below) so its possible to avoid some of the more onerous compliance standards.

[OwO]

Focus on the safety aspects because these are what can land you in trouble real fast (e.g. a unit sets your user's house on fire). I'll assume you are using a properly certified PSU, so let's look at your design:

- Does it handle large (> 5W) amounts of power?
- Any batteries?
- Flammable enclosure/materials?
- Does it interface to other machinery that can be unsafe if your device fails? e.g. controlling large motors, heaters, etc

For EMC, you should absolutely focus on susceptibility. Don't let emissions requirements distract you from what is really important, which is the fact that your device *can* malfunction if placed near a strong RF source. For example, I've seen PSUs go overvoltage with a nearby strong transmitter, soldering irons lose temperature regulation and keep the heater on, etc. These are the things that most designers completely overlook and can lead to safety liability real fast. I highly recommend building several transmitters for e.g. 10MHz, 100MHz, 915MHz, 2450MHz, of at least 5W each, for susceptibility testing. Emissions otoh as long as they aren't completely egregious generally won't land you in any amount of trouble, so just do some probing with a spectrum analyzer and be done with it. All cities are already littered with a shit ton of RF noise anyway, so no one is going to care if you are emitting slightly more than what is allowed.

[OwO]

In the US all test equipment are exempt from FCC emissions testing, so if you can make the case that it is an instrument you are in the clear already. If not, you are selling abroad into the US and it doesn't matter, since the buyer is the importer. That's not an excuse to skip safety testing, but I don't see any problems with skipping FCC certification because ~everyone does these days. Same thing with CE, do the necessary safety testing and certification as required, but EMI can be done with the cheapest SA and a wideband antenna.

[AndyC_772]

I highly recommend building several transmitters for e.g. 10MHz, 100MHz, 915MHz, 2450MHz, of at least 5W each, for susceptibility testing.

In every single case where I've seen a product fail immunity it's been at a very specific frequency (which may depend on the length of the cable attached to the device, or perhaps it happens to hit a point where the device's internal filtering has a notch).

Fixed frequency transmitters would be completely useless in all these cases.

[OwO]

No, only if you test at low powers would you need to hit a resonance to induce failure. For certification purposes you care about whether it fails at exactly the amount of power stated in the standards, but for your own testing you can simply up the power to guarantee a weak design will fail. In both the PSU and soldering iron case it wasn't frequency selective; I can get the soldering iron to fail on anything between 100MHz and 1GHz. Mind you, that PSU *is* FCC certified and passes the standard stipulated immunity test, but for my uses it's unsafe and IMO such a product wouldn't have been sold if I were in charge.

[Gyro]

I believe that there is a *lot* of stuff sold which hasn't been expensively tested by an RF test company, by small companies - but obviously not many want to talk about it. How can $25 products, such as test gear sold to hobbyists, which might only sell in the few hundreds at most, cover $10K+ EMI testing cost and still make a profit?

There must be quite a few posters here who sell there own designs.  It would be interesting to have an anonymous poll to see how many do and how many would admit to selling non certified product or product which hasn't been third party emissions tested. After all the chances that a small voltage reference module for calibrating DMMs or an Arduino temperature sensor shield would cause real world problems must be vanishingly small although theoretically possible if marginally stable and oscillate in some conditions. Is Dave's uCurrent CE certified for example?

This space has exemptions in FCC part 15 subpart B (see attachment below) so its possible to avoid some of the more onerous compliance standards.

Yes, hobbyist stuff and prototype boards fly below the radar in the EU too, that's how Sparkfun and the like get away with it.  Unfortunately the OP indicated Consumer in this case. Which ups the approval stakes.

As I mentioned previously, find a friendly local test house. They will know all the specific standard that apply [EDIT: and have them!], and if the item is simple, has a small number of interconnect configurations and operating modes (those take up time). You may be surprised at how little it actually costs you to do it properly. Accompanied testing allows you to sit with the test engineer, perform quick investigations and fixes and hopefully a first time pass (as long as you implement the mods - they will take product photos).

One company we used was out in the countryside, had a nice open field emissions setup, and all the chambers located in a converted pig farm. At one point, due to high demand, they were operating 24hr 3 shifts. If you took one of the cheaper night slots, they even threw in free pizza. I remember one time there was a load of different sex toys waiting to be tested for the importer. Happy memories.

[Wilksey]

You can worry about certification until the cows come home,

I know a company who have been operating for 30+ years, and apart from the equipment it builds for the Uk government nothing else is certified, the components inside are already certified and all radio equipment is using a manufacturer approved antenna, therefore the assumption is that the RF is not being abused so it should still comply with the original cert criteria, much like AdaFruit, the only difference apart from the radiated emissions and immunity EMC testing is a product file, which has datasheets, information about decommissioning, WEEE information, product and component information to support the claim etc.

A lot of local councils buy the products and most of the products are "self certified", and not once has it ever been questioned or challenged.

So i'm not saying what is right or wrong for you to do, and this is industrial electronics not consumer, but, if you were to self certify I very much doubt anyone would ever know or even care as long as you don't do anything knowingly that could take it outside of the operating characteristics.  There are people who think changing a resistor it will need to go back through all compliance testing then there are some that will change the antenna path and not worry about testing and rely on the OEM testing.

You would be surprised how many just don't put the declarations on or just stick them on and hope for the best, ultimately it comes down to cost, it's not cheap to do compliance testing, so if you are only selling something for a few £'s then it might not be viable.

[fcb]

These are the things that most designers completely overlook and can lead to safety liability real fast. I highly recommend building several transmitters for e.g. 10MHz, 100MHz, 915MHz, 2450MHz, of at least 5W each, for susceptibility testing.

NO! Don't build transmitters, 5W or otherwise - this is a shite way to look for susceptability and super illegal! Susceptability testing should ALWAYS take place in a chamber, and you need something like a swept 20W source (power is misnomer, it's really field strength that matters).

If you are just starting out, then probably find a competitior product(s) and see what they are claiming compliance to. If you need EN standards - buy them for a small fraction of the price from www.evs.ee (the Estonian standards agency)!!

If you are buying a COTS PSU - make sure it has all the CoC/CE stuff.  In most cases that's the big risky EMC bit.

You can do a huge amount with a basic spec-an and some DIY field probes. If you can post schematics/layouts on-line then this forum will give you a huge amount of help/advice re: passing emc and other gotcha's.

Also,you might be surprised by the number of products outhere don't have TCF's and testing to back up their CE status. This might be a function of the virtually zero enforcement that goes on................................................................................

If you are in the UK - drop me a PM/email, we have an EMC chamber and the toys sat idle 98% of the time and happy to throw stuff in it for genuine scrappy boot-strapping start-ups for a small charity donation.

EDIT: Sorry Wilksey - just read your post - not trying to parrot your line on 'surprised by' 

[Wilksey]

No worries, just goes to show it does happen more often than some think.

Speaking of PSUs, I had "dealings" with Traco - usually a good brand not cheap, not overly expensive, all the right stickers in the right places etc etc.
One of the products we did take for testing was transmitting more noise than the analyser could read at it's settings, we had literally just the PSU at one point and tried different load resistors to make sure it wasn't a loading thing, in the end we switched out to XP Power, we had brief exchanges of conversations with Traco and they asked for the test results which we sent them, in the end they just started ignoring us, so we sent 400 PSU's back to the supplier after speaking to them and them not getting anywhere with Traco (we sent them all of our original correspondence) and they gave us XP's for the same price, they were on paper £15 more each than the Traco for the same specs, they looked quite similar too, but they were very quiet!  Been running them ever since and switched all of our products away from Traco.