An update on CE compliance

Update 23.45pm Mar 30: minor changes to the article to reflect the fact that recent versions of the BeagleBoard-xM appear to have acquired a CE mark.

Update, 8.40pm Mar 28: element14 have put out a new update to their FAQ, enlarging on what’s happening with CE compliance. It’s well worth a read – head on over!

Update, 6pm Mar 28: we have spoken with BIS this morning, and they have confirmed that, given the volumes involved and the demographic mix of likely users, any development board exemption is not applicable to us; as a result, even the first uncased developer units of Raspberry Pi will require a CE mark prior to sale in the EU. As we mention below, we are working with RS Components and element14/Premier Farnell to bring Raspberry Pi into a compliant state as soon as is humanly possible.

Following on from last week’s discussions, both RS Components and element14/Premier Farnell have now informed us that they are not able to distribute the Raspberry Pi until it has received the CE mark. While this differs from our historical view (as we’ve said before, we believed that the uncased Raspberry Pi was not a “finished end product”, and could be distributed on the same terms as earlier versions of the BeagleBoard and other non-CE-marked platforms), we respect their right to make that decision.

The good news is that our first 2,000 boards arrived in the UK on Monday and that we are working to get them CE marked as soon as is humanly possible, in parallel with bringing the remainder of our initial batch into the country. Pete and Eben have been burning the midnight oil – literally; I only exchanged about three words with Eben yesterday, and those were when he got back in from a long day’s hacking at two in the morning. On the basis of preliminary measurements, we expect emissions from the uncased product to meet category A requirements comfortably without modification, and possibly to meet the more stringent category B requirements which we had originally expected would require a metalised case.

We’re also talking to the Department for Business, Innovation and Skills (BIS), to better understand the terms under which other non-CE-marked platforms are permitted to ship to domestic end users in the UK, and to obtain a definitive statement as to whether we can distribute on the same terms. We should say the UK Government in general, and BIS in particular, have been incredibly supportive of the project so far; they are looking into this as a matter of urgency, so hopefully we should have another update for you soon. With graphs in. We know you guys love graphs.

All this means that we’re waiting on one of two things – the results of further EMC tests, and whatever BIS comes back to us with – before RS and element 14/Premier Farnell can give you any firm delivery dates. We’ll let you know as soon as we do.

Finally, here, as promised, are some pictures from the factory (taken by one of the observers RS sent in – both RS and element14 have people onsite to oversee production) of the boards that the Foundation now has.

Raspberry pis in the factory

Raspberry pis in the factory


At the test bench

Each board spends a short while at the test bench. These are functionality tests which are performed in China, not the compliance tests we're doing later this week in the UK.



Once a board has passed, it's ready to be put in an anti-static bag and sent out.



Stephen avatar

Awesome! thanks for the update, it helps pass some time.

Tass avatar

Oh wow – that’s a sight for sore eyes!! Excellent update – you’ve made my night!

eben avatar

You should see the pallet of boards. 40 boxes of 50 boards each in about a cubic meter. Quite an emotional moment for me and Jack when we unwrapped it.

Alexander Langer avatar

You should have taken pictures, videos, get a TV team and do live-coverage ;)

spamel avatar

I think is a site for sore eyes

liz avatar

Oh, well played. ;)

JamesH avatar

That joke won an award for Tim Vine I believe.

Chris Rowland avatar

It’s for sale, register now!

Paul avatar

Does this affect shipments to non-EU countries that don’t care about the CE mark?

eben avatar

Excellent question. North America is out of the question (FCC requirements largely mirror CE ones), but there are areas of the world which are less strict. We’ve discussed with element14/Premier Farnell the possibility of diverting their entire share of the initial batch to these countries, but will probably not do this as the howls of anger from EU customers would be deafening.

Mario avatar

Mexico doesn’t need regulations.. can you send mine first? :D

greg avatar

I don’t believe that FCC and CE are mirror’s. The europeans have been on a crusade to screw SME’s for the last few years egged on by a Test and Compliance lab industry that is now more healthy than the electronic SME industry!

The europeans require ROHS and Immunity testing as part of the CE stamp. FCC part 15 only requires emissions testing. But it’s the list of exemptions that you should watch. Most devboards come under unfinished commercial test equipment exemptions, meaning its not a consumer product and that any person using it will be using it for the development of something (ie commercial activity) whether or not it actually leads to that or not. European EN5022/24 and the mess that are the supporting guidelines say something similar. Tip1: read FCC part 15 first – its written by humans. Then try to read the european directives. Tip2: move the sales operation to china and let people order online direct from there.

tgoldblatt avatar

There isn’t really an exemption for dev boards (which aren’t test equipment by the FCC’s definitions) – what exists is an acknowledgement that component boards that will become part of a larger system may only be meaningfully evaluated for compliance in terms of that larger system. In that case, the purchaser of the board is assumed to be a manufacturer or system integrator that will assume responsibility for EMI compliance. In the case of Pis (and for that matter Beagleboards), that is largely not the case – the purchasers are largely end-users, who (even if they are doing further development) are neither knowledgeable enough about EMI issues nor equipped to test for them, and therefore should not be sold non-compliant equipment, since they will be on the hook for things they cannot reasonably be expected to be competent to deal with. (And moving everything to China doesn’t help, as Customs can (and will, upon FCC notice) refuse to let the boards into the country.)

greg avatar

There’s no evidence to suggest Beagleboards are replacing home PC’s or consumer items. In fact we used one recently as a development platform before we layed out our own hardware. I have no problem thinking that those who use these will be using them as part of a larger project. As such it is reasonable that these devices not be considered finished products but rather development boards

tgoldblatt avatar

Greg – as I said, they may actually be used for development, rather than as end-user devices, and some percentage may be used by professionals. However, perusing the projects list on, it appears a large number of the projects are being done by hobbyists. Even if they are doing development, it is likely that a high percentage of these are not knowledgeable on EMI issues, are not doing their (hobby) development in an industrial setting, and likely don’t have the facilities to test for or avoid EMI.

Since nothing exempts the user (whether an application level “end-user” or a developer) from having to avoid/prevent EMI (not being a “finished product” isn’t a magic “get out of jail free” card), they are still on the hook. (The requirements of Part 15 are basically to deal with reducing the responsibility on those non-knowledgeable, non-equipped users – if they cause interference with appropriately certified compliant equipment, they still need to deal with it, but their liability (legal and financial) is reduced or eliminated. That’s why such users should only use compliant equipment.) If you are selling equipment that has to be used by a certain class of users and/or in a certain way, but you _know_ that you are selling to those not meeting those rules, you are in the wrong. (Not sure whether you are _legally_ in the wrong, but you are certainly riding the edges.)

mark avatar

I’m surprised how quickly some people will move operations off-shore to sidestep local labour laws and/or regulation. Restriction of Hazardous Substances directive (RoHS) has been around for a while; RPi shoud not contain hazardous materials since it’s a new product. Granted the CE directive is a bit of a surprise but these distributors understand their markets and I think they are spot on. Many people’s perception of the RPi is that it’s a cheap home computer equivilent, not a component within a larger development, that’s certainly the message I’ve been getting from the PR campaign, BBC coverage etc.

Abishur avatar

I’m sorry, I just need to clarify something… are you trying to say that the RPF moved their operations off-shore in an attempt to side step local laws?

JamesH avatar

The only missing certs are CE/FCC, board is RoHS compliant already.

Lynbarn avatar

There are so many experts on the technical and legal implications of compliance and conformity on this board, it’s a wonder they didn’t raise the issue (at least by asking the question) of CE/FCC etc. certification several months ago…

greg avatar

I think people need to understand that in a place like China, where the regulations are generally disregarded or non-existant, your phone, tv and all your gadgets still work. Planes don’t fall out of the sky and cars don’t run off cliffs due to inteference. What has happened is that Europe has the heaviest barriers to entry to bring products to market. Its a severe barrier for british and european business because testing each product can easily cost 10K or more. All the products you are buying online from China have no testing. Usually they just put the labels on. This situation has occured because big business has lobied the regulators to get strict on this. The result has been to damage small business. Raspberry Pi is lucky because there are so many orders in place that they can cover these costs. Other small businesses already have things like this on the market. They havent had the attention of the Pi or the backing and have generally disregarded the CE requirements due to cost. The test labs have enjoyed the current situation as they make all their money off large business, who having accepted higher regulation as a barrier to entry for other players have pushed up test lab prices. Hence a vicious circle. SME’s generally ignore the law because they have no choice in order to compete with chinese goods where no testing has been paid for. The regulators are generally under-resourced and don’t care except in extreme cases of complaint.

Collin avatar

Canada has cUL. FCC is for the US not all North American. Will Rpi need to seek out cUL as well?

Austin C avatar

I would’ve waited to order until later on if I hadn’t intended to get one of the first ones, please have the distributors send them out in the order that the orders were received. ( BTW I’m in USA- would not be just EU customers howling)

mike m avatar

While I completely support what you are trying to do I believe that what you are selling is 100% a complete product. You may have deluded yourselves that you could get by compliance by saying that it is not but it is and this should have been prepared for.

You will have my money and do have my support but this launch has not been handled at all well.

eben avatar

That’s certainly a valid viewpoint. As we say in the post, we’re struggling to understand why a lot of very similar (if less powerful) small board computers can ship without the CE mark, while we’re unable to. We’re looking forward to the definitive BIS statement.

greg avatar

Bare boards are never a consumer product, and should never be regarded as such. It will be very difficult to get to Class B emissions (consumer) levels without the benefit of a case/shielding cans. You should sell this as an R&D only product. Farnell and RS have paid compliance people that are not necessarily on the side of manufacturers. Same situation for all the test labs.

tgoldblatt avatar

Just calling it an R&D only product doesn’t work if you are selling it to just anyone as effectively an end-user product. (The Beagle people are doing that, and I think they are in the wrong and they (or their HW partners) could get (expensively) called on it.)

greg avatar

I think there’s a reasonable expectation that if you know how to program / load an OS onto the board that it is essentially an R&D / testing device. Certainly not a consumer item. The reality is that anyone can buy any development platform. The manufacturer need only make it clear that is its intended purpose. If hobbyists choose to buy it, that is their affair.

tgoldblatt avatar

greg – as I say elsewhere, that isn’t really true if they don’t understand what it means. It is similar to anyone marketing a product that they know is going to be used in an illegal way (even if the product itself _can_ be used in a legal way). This is especially true if the seller knows it is illegal, but the buyer doesn’t. The seller _may_ be legally in the clear, but they are morally responsible.

Schaefer avatar

I believe this can be a consumer product as bare boards are sale to all of us at any given time. What I’m speaking about is bare bones computer parts at the brick and mortar store down the street. Looking at the product, I don’t believe anyone is being dishonest here in the marketing of it, it’s a What You See Is What You Get (WYSIWYG). I see a bare bone computer, and as a hobbyist I can’t wait, I’m thinking about creating a cool little case for it.

However, if others that are not computer savvy view the product I’m sure they will be intimated by a bare mother board looking thing… I can’t see my mother or one of my friend that don’t know much about computers purchasing one of these things because they will be too intimidated by it.

cnxsoft avatar

I guess that’s because of the quantities involved.

tgoldblatt avatar

I think (having looked at the disclaimers on both the original Beagleboard and the xM) that the beagle people are being at best disingenuous. While they state the board is a development system that should only be used by developers in with both the knowledge and facilities to evaluate and control EMI issues (and in the case of the xM, use it only in an industrial environment (class A, even though they mistakenly claim it is certified to class B), they clearly are promoting and (through their HW partners) selling it to end-users (who, to a large extent, are neither knowledgeable of nor equipped to deal with the EMI issues) and for use in residential areas. That is, they are putting a legal disclaimer in their documentation, but are knowingly promoting and selling it in ways effectively guaranteeing the disclaimer will be violated, making them guilty of at least contributory negligence in the violation of FCC part 15 and (probably) EMC Directive 2004/108/EC.

Konstantins avatar

It’s not about what someone thinks is or is not. Rather about avoiding dumb side effects of law fool-proofnes.
There should be no obstacle to categorize R.Pi as is as an intermediate product. It’s a bare pcb for * sake!

rew avatar

On the other hand… with over 100k distinct people wanting to buy one, that is really an indication that it is not just developers with EMI knowledge who want to buy one.
When you sell 1000 beagleboards, you can claim that most end up in companies evaluating the processor for use in an embedded application. And that they know what EMI is and what to do about it. But with 100k RPIs in backlog there is a strong hint that they are not only for development.

liamf avatar

Agree 100% with this comment.

The launch has not been a model of organisation.

It’s quite amazing that no-one thought of getting a definitive statement on CE/FCC compliance in advance of promising shipments, given that at least one of the foundation members works for Broadcom who supply OEMs all ’round the world.

You’ll get my money too. And my support.

But frankly at this stage, I am just waiting for the next thing to go wrong. Because with luck like yours …

monkeyhybrid avatar

Oh, come on now, I think you’re being a bit harsh.

We’ve been given full details at every stage of the process, from conception to launch. It’s hit a few obstacles along the way, just like every other product launch in history. The difference with the Raspberry Pi is that the whole team have been completely transparent and open as to what is going on. Other manufacturers don’t give you these details, they get away with delaying release schedules with no explanation or don’t even give an expected release date until it’s ready to buy.

The Raspberry Pi team’s willingness to be so open about these issues is a blessing and a positive sign of what is to come in the future.

Stuart Fox avatar

Well – I paid for my open pandora 3 years ago and after a run of things going wrong and circuit co in Texas almost killing the project with their incompetence- I’ve still not got one and now they are asking for a voluntary extra payment because they are making a loss on what they originally asked for!!!! This launch is going very very well by comparison! – Still supporting the open pandora and excited about getting it – much respect for those guys and you – I’ve been raving about the positive educational and social implications of the Raspberry Pi project to anyone who’ll listen.. not disrespecting anyone – just stating my experience. Good luck guys.

Rick avatar

I have to agree. I gave them the benefit of doubt when the initial delay happened due to the wrong ethernet ports as that was a simple error at the factory. However to not know if your product is going to need to comply to standards isn’t good enough.

We’ve had no real formal explanation as to how long our orders are going to be delayed…we didn’t all get up at 6AM to order something with no known date that it was to arrive. At it stands we’ve not actually been given even a rough estimate for when we’ll receive our Pi’s.

The whole launch has been very poorly handled. I’d expect CE compliance to be one of the first things that was discussed with Farnel and RS.

I appreciate that you guys work VERY hard on this, but it needs to be said that the communication side of things is pretty terrible. Many people (myself included) aren’t willing to trawl through pages of ‘this has already been discussed’ closed threads on the forum trying to find an answer to the most basic of questions.

A blog post really needs to be made saying “if you ordered on launch day, you wont be getting your RPi until at least X” as so far you’ve ignored requests for a rough date or even an update on whats happening.

JamesH avatar

have you even considered that the information you are asking for simply doesn’t exist yet? And that is why it hasn’t been posted? It may be a basic question, but if no-one avtually know the answer yet, how can we possibly post it? And we don’t know the answer yet, because of a number of issues – CE compliance being one it them. Or have you even read any of the posts explaining WHY the CE compliance testing wasn’t originally being done at this stage?
I would say that the communication over the whole project has been better than any other product launch I can think of. In fact, its difficult to see how any MORE information could have been sent out. As to updates on what’s happening – have you even bothered to read the front page of the website for all those updates that have been posted? Jsut because those updates don’t contain the information YOU want (which doesn’t even exist yet), that doesn’t mean that there have been no updates.

Rick avatar

Picture this. You go and place a pre-order for something on, lets say Amazon or somewhere like that and they confirmed your order with an estimated date of dispatch based on the stock levels and pre-order information.

You’d be mighty annoyed if 15 days after your estimated date had passed that youd not received so much as an email saying “really sorry but its delayed”.

Generally if there is going to be a delay, at very least an email is sent out saying how long the delay will be, and why its happened. As I said above, we’ve not actually been given any form of information as to how long the delay is. The last actual update was that there was around a 2 week delay due to the wrong Ethernet ports, even then it was only provided in the comments section of a blog post.

People WONT sit there trawling through pages and pages of forum and blog posts trying to find an answer. People need to have an email saying its delayed – heres how long for, heres why.

JamesH avatar

Well, there have been a lot of emails from RS and Farnell and us explaining all that stuff – and apart from the exact length of the delay (which nobody knows the figure for at this stage), all your points have been covered.

jojopi avatar

Are those BS1363 sockets? Is the factory in Hong Kong?

eben avatar

Shenzhen, so very close.

Oli avatar

Out of random curiosity, what’s the yield like on this run? How many boards per hundred are passing all the tests?

liz avatar

The raw yield was about 95%. The post-rework yield was much better – north of 97%. Obviously, the factory is working to get it even higher as they ride the experience curve for this particular design.

texy avatar

Well strictly speaking the yield was much lower than that – they had the wrong Ethernet socket fitted remember ;-)

Pithonica avatar

Was wondering what those CE mark categories were and found this:

“Category A will be authorised to navigate more than 25 miles from land, Category B up to 25 miles, Category C up to 3 miles and Category D inland waters”

All clear now! Wait that was for the CE marks on boats.

eben avatar

Pastry tends to get soggy beyond the 3 mile mark…

DennisMe avatar

Eben, that’s only if you leave it out in front of the radar too long. (How do radars manage to pass CE, I wonder…).

spamel avatar

Eben, is that a BIS statement or a BS statement?! I reckon they are having you all on a wind up. Surely they should have insisted on this before taking the contract for distribution, it seems to me somebody somewhere got a bit worried and they have played the H&S cards. Barstewards!

mkopack avatar

Was this not covered in the contracts that were done with RS + Farnell? I can understand their reluctance to ship non-CE product given liability concerns, but I don’t understand how they can somehow change the terms after the fact if they agreed to ship it knowing it hadn’t been CE certified.

Also, what’s this going to mean for the US (FCC) market? Are you also going to have to go through FCC acceptance tests? Are those tests also being performed now, or will US customers have to wait even longer to get through the FCC battery at a later date?

And what happened with the other 8000? The last we heard the ethernet port fixes were nearly completed, and that was ~3 weeks ago. Where did the other 8000 go if you only have 2000 in hand in the UK?

eben avatar

Quick answers:

It’s basically an oversight. The two companies’ compliance people weren’t involved at the point where contracts were signed.

We’re doing FCC at the same time as CE; the requirements are basically congruent.

The other 8,000 are trailing the first 2,000 by a week or two.

mkopack avatar

Great thanks…

Pithonica avatar

Even before the madjack problem showed up on shore, Eben was talking about the 10K being in more than one shipment:

“The first couple of thousand are imminent”

March 5, Elreg:

Montala avatar

Once the Raspberry Pi does receive the official CE mark, doesn’t this then have to be ‘engraved’ onto each and every board, as opposed to just sticking a label on them all?

If that is in fact the case, I am curious as how to this can easily be added to the initial batch of boards you have now received.

eben avatar

Nope. We’re lucky that you can get away with a CE sticker for a limited period of time. We’ll add the CE mark (and the little WEEE dustbin and so on) to the silkscreen for the next batch of boards.

tgoldblatt avatar

I don’t think that works for FCC though – it doesn’t have to be etched, but it must be a permanent label and specifically (per Part 15.19b(4)) cannot be a “stick-on paper label”.

greg avatar

The device is too small for FCC labels (credit card size). Placement in the associated manual is acceptable.

ricky roberson avatar

We’re gonna get a manual now? Hooray!

tgoldblatt avatar

Per CFR 47 Part 15.19b:

(3) When the device is so small or for
such use that it is not practicable to
place the statement specified under
paragraph (b)(1) of this section on it,
such as for a CPU board or a plug-in
circuit board peripheral device, the
text associated with the logo may be
placed in a prominent location in the
instruction manual or pamphlet supplied
to the user. However, the unique
identification (trade name and model
number) and the logo must be displayed
on the device.

Sancho avatar

Maybe I am missing something, but I can see the pins for the peripherials soldered on – even though I read here, that only the first batch of ~10 units does have them on and the rest will be just drilled holes ready for home-soldering gurus :)
Don’t get me wrong, I would love it should it come soldered with the pins.
I am just not sure and the image resolution does not help either…

Sancho avatar

I ment GPIO of course… Too late, have to go to bed.

eben avatar

Funny story – turns out I forgot to remove it from the BOM before getting the quote from our CM (at some CMs the extra 26 through-hole joints caused a significant cost delta). They quoted with it, and were affordable, so we left it on.

ailwyn avatar

So will they stay or be removed for future runs?

Grumpy_Git avatar

I’ve never loved a man before. good job on dealing with the whole situation so calmly.

One question, do you sometimes wish it had been a lower key launch so that you could have clear the first 10k as pure development boards, not for retail etc… etc…

ps for a nice FCC getout clause, read the TI Launchpad user guide:
FCC Warning
This equipment is intended for use in a laboratory test environment only. It generates, uses, and can
radiate radio frequency energy and has not been tested for compliance with the limits of computing
devices pursuant to subpart J of part 15 of FCC rules, which are designed to provide reasonable
protection against radio frequency interference. Operation of this equipment in other environments may
cause interference with radio communications, in which case, the user will be required to take whatever
measures may be required to correct this interference his own expense.



Iain avatar

Oh well, at least it gives me a reason to break out the desolderer, doesn’t get much use these days with all this new fangled surface mount stuff :)

jbeale avatar

Great to read this post and see the photos of the assemblies. So close now!

Bikko avatar

Kinda reminds me of a tray of cornish pasties going into the oven

Capt. America avatar

If the CE compliance fails what does that mean for the boards? Trash the batch and ship none? Or else what is the plan in the event that CE compliance results are deficient to obtain certification?

jbeale avatar

they could still ship to the various countries without CE or FCC regs. They could try to fit into whatever exemptions allow the BeagleBoard to avoid CE and the TI demo boards to avoid FCC. They could rework as needed to meet CE/FCC; add a metallized case, etc. Or they could ship them back to China as electronic scrap whereupon they would likely show up on Ebay, maybe even with fake compliance labels…

eben avatar

Yeah, that’s the galling thing. How many products with the CE label on them have actually been put through testing? Most of the stuff in your average supermarket is “China Export”.

Fortunately, there’s basically no chance the board won’t meet spec. It’s well designed, and we did a run in the test chamber a couple of days ago that only just missed despite the fact the HDMI was misconfigured (and therefore radiating like mad).

magnetism87 avatar

It sounds like the distributors are pulling the rug from under you. Why have they raised these concerns so late in the game, and what was their justification for treating the Pi differently from a Beagleboard? Also, compliance testing takes time, resources, and equipment, that I am sure will cut into your bottom line. I hope this does not derail your efforts. Do you foresee this affecting the pricing structure of the Pi models down the line? Is there anything that the community can do to help you guys out? Do you have a donation mechanism, so that people can contribute and help you offset the unanticipated cost of compliance testing?

Keep up the good work!

eben avatar

Thanks for the kind words. We’re incredibly lucky to have received support from a number of high-net-worth individuals in the Cambridge area, so we can afford to cover the (totally unexpected) cost of bringing the uncased device into conformance. It’s a setback, but not one that could derail us or cause us to vary the pricing.

jfedor avatar

The picture is blurry, but are those GPIO pins on the board? I thought the production boards were supposed to not have them.

jbeale avatar

See eben’s reply above. they were left on the BOM when it was quoted; turned out to be affordable, so they went with the pins on.

eben avatar

Yeah – I forgot to take them off the BOM before our CM quoted the assembly, and they were cheap enough, so we left them in.

Frank Buss avatar

Nice, will there mounted pin headers on the board in the next batches, too? Would make it easier for schools and beginners who wants to avoid soldering irons, but doing some Arduino like stuff with it.

Jim Manley avatar

Eben, Liz, and Team R-Pi

First, thanks very much for the update – any news is better than no news (although know news is even better :) ).

You may not be aware, but, I believe that other manufacturers have been using a loophole in the CE standards (they’re not technically regulations, much less law) wherein only digital devices with clock speeds at, or below, 300 MHz are subject to compliance, except for toys and transmitter/receiver radios. This is due to the typical delays in the bureaucracy in that the standards haven’t been updated since the days when computers had clock speeds well below 300 MHz (hey, the ARM you’re using is equivalent to a Pentium III > ~300 MHz, right? Problem solved! ;) ).

As I posted elsewhere:
FCC Part 15.103 states:
“The operator of the exempted device shall be required to stop operating the device upon a finding by the Commission or its representative that the device is causing harmful interference. Operation shall not resume until the condition causing the harmful interference has been corrected. Although not mandatory, it is strongly recommended that the manufacturer of an exempted device endeavour to have the device meet the specific technical standards in this part.”

As for the EU/CE, the emissions standard EN 55014-1, which covers consumer digital devices (not including radio transmitters/receivers) does not address digital emissions above 300 MHz since effective radiated power for such devices is typically milliwatts, or less. It does include radiated emissions limits and methods, but only for toys. The experimental/educational nature of devices such as the R-Pi means they are not being sold as toys, and they’re not intended to be radios! YMMV, so start looking for copper mesh, conductive paint, etc. :)

Farnell/RS are using the FCC’s strong recommendation to ensure compliance so that a bunch of boards don’t get returned because some unsuspecting customer is living next to a neighbor with nothing better to do than find something about which to complain. I don’t often quote Brits who are alive today, but, some of the members of the Dead Poets Society were prescient – as Shakespeare said, “First, we kill all the lawyers.” In no way am I promoting the killing of lawyers – go arrest Shakespeare, OK? :)

I really am looking forward to seeing the “splatter” diagrams, polar graphs showing the emitted radio frequencies by angle of orientation, colored by frequency, around all three axes of rotation.

eben avatar

Oh yeah. We’ve actually got some shiny graph goodness now, but we’d miscalibrated the HDMI output so it makes us look bad. More in a few days when we’re rerun with the proper settings.

tgoldblatt avatar

I don’t see the 300MHz value in the current (2004) version of the CE EMC regulations (2004/108/EC). Further, my reading of those seem to indicate that the PI would not be in any of the exempt categories, would not be considered “inherently benign equipment”, and would be subject to them as an “apparatus” since (even though it is a “component/sub-assembly”) it is available to end-users. (Note that this is because an end-user is assumed to not be qualified to deal with EMI/EMC issues. Components/sub-assemblies intended to be exclusively used for an “industrial assembly operation” (which I take to mean a commercial manufacturer) as a part of a further assembly, _are_ exempt (because it becomes the responsibility of that manufacturer to ensure compliance).) This section is why I think the Beagleboard documentation is disingenuous at best – they are claiming it is intended for usage in ways and in environments that make it exempt for the seller, but it isn’t actually being promoted and sold that way (certainly not _exclusively_).

I’m also not quite sure what your reference to Part 15.103 is supposed to be saying. It does say there that for exempted devices, meeting the technical standards are recommended not mandatory, but the Pi doesn’t match any of the exempted devices in that section (that is, it isn’t exempted).

Finally, as far as I can tell, EN 55014-1 applies to home appliances, tools and so on, with motors, switches, and relays (that is, electrical, not electronic devices), but doesn’t apply to home digital devices.

jaygee avatar

It’s great to see RasPis in bulk production at last and to hear that compliance testing is progressing. Can anything be said at this point about how component sourcing is going now, especially the ethernet connector that was causing worries when the wrong type was fitted? I am still waiting to be able to order (having initially chosen RS) and Farnell are now quoting 135 days. I am sure many are in this position and it would be really good to have some news on how scaling up component sourcing to allow production to meet the demand with reduced lead times is going.

Keep up the good work – it really is appreciated!

eben avatar

Quick summary: no current concerns around component sourcing (as you’d expect given who our distributors are). I’ve seen very large orders go in for all the potentially troublesome components (particularly the ASICs and connectors). I don’t think there’s going to be a backlog for very long :)

rew avatar

Then WHY are they quoting a 135 or 143 day lead time?

I have the impression that you guys got a good deal on the 10k batch because you chose a relatively small factory that is able to do 10k in about four weeks (they promised 2, but in the end we’re more than 8 weeks along now).
So that “small” factory is now working like mad to be able to produce 100k units in 20-40 weeks?

liz avatar

Because they had a system error which outputted the wrong date. I’ve just put a post on the main page about that.

Brian Driver avatar

Reported from the RaspPi Facebook fan page:

CE compliance – psh. What’s a little EM radiation between friends? It’s not like I’m going to wallpaper my office with these things.

Wait – new project idea!

John Sousa avatar

The Beagleboard is FCC Part 15 Class B certified so I’m not very surprised the RPi needed similar certification, especially since it’s being marketed and sold to non-EE end users.

Question: s FCC compliance testing being done at the same time as CE? If not will this be an issue with US sales?

liz avatar

Could you give us a reference for that? It’d be very useful. avatar

Search for “FCC” in the PDF at

John Sousa avatar

A reference for Beagleboard compliance?
It’s stated in page 2 of the System Manual:
“NOTE: This equipment has been tested and found to comply with the limits for a Class B digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference when the equipment is operated in a commercial environment. This equipment generates, uses, and can radiate radio frequency energy and, if not installed and used in accordance with the instruction manual, may cause harmful interference to radio communications. Operation of this equipment in a residential area is likely to cause harmful interference in which case the user will be required to correct the interference at their own expense. All accessories used with this board must meet FCC certification to maintain compliance of this equipment.”

jbeale avatar

Note that FCC “Class A” is commercial, and “Class B” is consumer/residential. See also:
Example label on a fluoro light ballast: “This equipment has been tested and found to comply with FCC 47 CFR Part 18, Non-Consumer RFI/EMI (“Class A”) limits. This Ballast should only be installed in a commercial environment. Do not install this ballast in a residential environment.” avatar

Also, for FCC documentation regarding the requirement:

It seems that selling “non-compliant” devices can result in federal seizure of the devices and steep criminal penalties, but only if the non-compliance leads to interference that is reported to the FCC. Presumably that risk (as well as risks in other pertinent jurisdictions) is why someone at RS/Farnell decided to prevent them from getting out.

jbeale avatar

My boss used to run his own company designing various PC interface cards. He was undercut by cheap Chinese clones with no RFI suppression and fake FCC labels. He took the import boards to a compliance test lab and found very high levels of emission, way above FCC limits. He contacted the FCC and they had zero interest in pursuing it. FWIW

tgoldblatt avatar

They used to (30+ years ago, not sure when it tailed off) pursue EMI issues _very_ seriously, with teams sent out with sophisticated DF’ing equipment to locate the radiation. Between industry pressure, redirected priorities, and funding limitations, this is no longer the case. Whether they have no _interest_ in pursuing it or just no resources to do so is unclear.

Roland avatar
greg avatar

The first version did not have it. Subsequent to the device becoming successful they have included it.

Please post the reply from the BSI when you get it. It will make interesting reading. Hopefully some common sense will prevail

Abd Al-Latif Whiting avatar

Now I wonder, it’s getting awfully close to 1st April. Will all become clear then :-)

eben avatar

Damn. You rumbled us :)

Abd Al-Latif Whiting avatar

Although I have not been in production for a few years, isn’t CE marking a self-certifying mark, and that you only need to prove compliance when challenged?
Shouldn’t this product have gone to a test house for certification at the development stage. After all, should there have to be a redesign it would have saved a lot of expense.
Of course, technical people rarely make good salespeople, and although the technical people managed to get us all hyped up, the introduction of this amazing device was not handled correctly. Maybe it should have been kept under wraps a bit longer , until it was fully developed and tested.
Like myself, people I know are now on the downward curve, and our enthusiasm has taken a bit of a knocking after weeks of anticipated sales. I especially regret getting up at an early hour, waiting for the Farnell or RS sites to open, and never getting anywhere.
I think that I will have to put this project to the back of my mind, until I am certain that I can actually place an order, for a stocked item.
Written with the best of intentions.

Cesare Montresor avatar

omg omg omg!
(I tried to think about something smarter but that’s the best I managed to do ^_^)

OlivierW avatar

You really should drop RS and Farnell and work with companies like Adafruit, Sparkfun or Seeed Studio. They’re probably less annoying and can understand better what you want to sell.
Since the beginning, RS and Farnell don’t even know how to sell the Raspberry Pi : first it’s on their website, then a waiting list. Farnell, wanted to charge me $70 with taxes and next day shipping (while the Raspberry Pi are still in China). Well, they really don’t know how to handle all that stuff.

Travis Rosenbaum avatar

While I don’t necessarily agree with dropping RS/Farnell, I wholeheartedly second the idea of working with Adafruit and SparkFun as additional suppliers. They are a PERFECT fit for reaching the type of community you are serving currently.

shirro avatar

Does the RS and Element 14 concern for regulatory compliance extend outside of the USA and EU? I am curious to know if C-tick is under way for Australia and if it will cause similar delays?

Jesse avatar

I think I speak for a large number of people with this request:
Have you got a higher resolution version of the first picture? It looks amazing, haha.

liz avatar

Sadly, no – they came embedded in another document, so that’s as big as they get!

Samuel avatar

RS only took expressions of interest and to let you know when you could order it with regular updates on what was happening. E14 you could order it on their site even though they didn’t have it, that didn’t make sense, the RPi’s hadn’t even been assembled and tested then they messed around with the price causing a bit of heartache with those who have paid already.

The mistake most people were making here is the product was never readily available to sell and though they weer ordering the product, but all they were doing is expressing their interest in the product and added to a queue that is attempting to be in some sort of order.

I for one can’t wait to mess around with this when I do finally get one. :)

yaconsult avatar

No one has “paid already”. When you place an order, the vendor may have made a small charge to verify that the credit card account is valid. Since they don’t verify the charge after it’s approved, the charge expires due to lack of confirmation and you were not billed any amount. So you have not paid anything – only confirmed that you submitted valid credit card details. Vendors are not allowed to charge for items that they haven’t shipped.

It was believed to be ready for sale to developers and experimenters. The much greater than anticipated number of orders caused the distributors to realize that all these orders couldn’t be going to only developers and that they needed to go through the compliance process to cover their own butts. The vendors do this stuff as a matter of course and are certainly more familiar with the requirements than anyone else – including the foundation.

My ship date was listed as March 30th until a couple of days ago. Now, no one knows – we have to be patient and let the process take its course.

rew avatar

There are plenty of shops that allow you to buy stuff they don’t have in stock.
Go to a car shop and ask for a fuel pump for your ’77 Chevvy. They’ll order one for you.
It surely makes sense to have people order the boards before they are in stock. Now you know for sure that the boards you get will be sold. One of Farnell’s worst nightmares is that they will be left with an unsellable stock. You occasionally see that they have some components or boards in stock which will never sell. They also have the successor or that component which is cheaper and better. Who will buy the older component? Nobody.

rpiguy avatar

Ugh. This is frustrating and discouraging.

Just checked my order status on the Newark website. At last check, a few days ago, it said it would ship Apr. 3. Now it says Aug. 16!

The timing of the ship date change makes me think maybe it’s due to compliance testing.

Abishur avatar

Please read the post before this one as it addresses that e-mail (short answer, it was a mistake)

Morgaine avatar

A product range with which it makes sense to compare Raspberry Pi for certification purposes is Arduino, as Arduino boards are produced in very large quantities and the company is similarly based in Europe (Italy).

The leaflets accompanying my Arduinos say that the boards have RoHS, CE, and FCC (Class B) certifications (as well as something called Life Zero Impact). Since the Raspberry Pi is substantially more consumer-oriented than an Arduino, it should be expected to have at least that level of certification.

Steve avatar

Part of me is starting to hope this is all part of a VERY elaborate April Fools’ day prank. I’d be very sad not to get a RaspPi but, man, would I respect that kind of commitment. At the very least you should consider an April 1st vapor ware “confession”. That kind of trolling would get you some major PR, and there’s no such thing as bad PR.

mtrx avatar

Thanks for the insight on the comliance testing, not a very nice move from the distributors. I’d still love to hear why RS has removed some of us first batch customers from their update lists though, as discussed in the previous article. Since they are planning to open a dedicated store page for raspi soon, I doubt us unlucky ones who have been dumped from their lists will get that update email either.

vaalpens avatar

Yep, same here, also removed from the initial mailing list . also no updates and I registered my interest with RS components within the first few minutes the option was available.

RS Components in South Africa is clueless, I really hope we don’t get double delayed down here because of RS…

Sadpanda :(

piopio avatar

I’ll just say that the “distributors” (who are actually -manufacturing-, as per the licensing model) are just doing things right, as compliance testing is required to sell the RPi, other claims are just hilarious. RS and Farnell are a business, and will handle the matter how you expect from them. Oh, they’ll do it for the profit, first of all they’ve put lots of money on the line too.

Stuart Lea avatar

Thanks for the update…As far as I can see the RPF have become a victim of their own success. As soon as the wider media got hold of the story of a $35 computer (with a nice desktop and 1080p) then the board became a “must have” for everyone who was vaguely interested in computers, including non-developers. I’ve no doubt that if wasn’t so attractive it would be shipping now as a dev board. And, in turn, this the reason why RS and Farnell got wobbly. It’s very easy for all of us to say that it should have been tested earlier and I’m sure, given their time again, the RPF would do that.

I’ve been as frustrated as anyone with the delays, but these guys need congratulating for their efforts so far. They’ve risked their own money and credibility to bring this here and been exposed to abuse from some parties. Most people don’t have the guts to run with their dreams and it makes them bitter over time.

If there was a PayPal account to buy them a beer, I’d be contributing in a shot. They need to sit back with the drink of their choice and know that people appreciate, no, love, what they have done.

This is the end game and they need a round of applause from us all.

Pew446 avatar

So my ship date of April 2nd is no longer April 2nd? Great.

Pew446 avatar

But thanks for making the pi anyways. :)

Mattt avatar

Excellent news, thanks for the update.

SN avatar

Nice to see a big clutch of the little beasties in the UK. Best of luck with the Compliance Testing and I look forwards to seeing the first “I got one!!!” posts on the forum, hopefully next month ;-)

lobster avatar

Fellow Raspberrians,

A pallet of raspberries = a super punnet?
How wonderful. (Eh Ma Ho as the Tibetans say)
Like gold. Only more valuable. Our future is underway.

Are we learning Python while we wait? I is.
Are we preparing for a world of Raspberries?
We iz.

Onward and upward.
Bravo team Rpi

antiloquax avatar

It’s great to see the RPis rolling off the production line. While it is frustrating to have another delay, we should not forget that we are getting an amazing machine for very little money. I’ve been watching some of the videos on YouTube and was fascinated by Eben’s comments about interest from “developing world” countries. When production gears up fully, the RPi is really going to change the world – people who have never used a computer before will be able to plug this into their TVs and instantly overcome the “digital divide”! Also I’d echo Lobster – I have a lot of programming to learn… :)

azbesthu avatar

I have a feeling that image shows the first batch and the mass production will be started after the compliance tests succeeded.

Pascal Harris avatar

Is that my one I see there? Top rack, row 3, column 2? I’m now wishing that I’d been selfish and ordered more than one. I am so excited – and my plans for this little computer just keep getting more, and more, extravagant.
For me, this is the most exciting computer of the 21st Century!

David Slee avatar

Dear Eben and Liz,

Keep your chins up! As true entrepreneurs you are attempting and against the odds achieving the almost impossible: to bring a high tech product through prototype into mass manufacture and worldwide availability at the lowest price point on the planet and in the public gaze. All as a charity and thus without the buffer of profit.

I am thrilled that some time this year I will have my Pi and get perhaps a few months of Python experience under my belt after 10 years away from hardware design and coding and then be able to buy another three or four Pi for my teenage son and teenage nephews, nieces and cousins who all want to have some part in creative computer science but lack the platform on which to learn like I did some twenty-mumble years ago with PETs, ZX-80, Spectrum and Beeb. Then I can give something back to the youngsters – the benfit of my experience to mentor and help their Pi journeys. So take your time and get it right.

The nay sayers will continue to criticise the decision to go with Farnell and RS for distribution and suggest alternatives. Keep your nerve. Schools are corporate buyers and will already have RS and Farnell on their approved suppliers lists and trust them to deliver solid reliable “education grade” product. A little delay is worth it to assure success in the big roll out. Don’t forget these first 10,000 are “Series Zero”; there are still manufacturing wrinkles to work out. The protoypes got you to TRL9, reaching MRL9 will take more time and effort still. But you will make it driven by hard work, midnight oil and caffeine overdose.

Finally, the greatest respect to you for baring all with the in depth reports on magnetics and EMC testing. Yes, I’m a geek, but x-rays of boards and polar radiation plots are meaningful and educational. Your charitable mission of education is already well under-way – you are educating and refreshing my skills and I can explain and show to my son how this process works with your candid blog as the “documentary footage” of engineering product design in action.

Keep up the good work and feel free to ask if you think there is anything the user comminty can do to help you right now.

Stuart Lea avatar

If only there were more posts like this. They’ve taken a risk and gone under the microscope during the process. I cannot believe that people didn’t expect problems and delays along the way….

As I said in my previous post, they should set up a PayPal account for us to buy them beer as a thank you for their efforts so far.

poppe avatar

This coming from the guy that called out the late Compliance testing…

Stuart Lea avatar

Yup, that’s me……I have a tendency of moving on and seeing the big picture and not living in the past…They are doing a great job, but not perfect and I stand by the fact that they should have realised that going on telly would give RS and Farnell the willies and they would insist on compliance testing. I’m disappointed that they didn’t see that one coming. We all screw up and that, in my opinion was a screw up.

But, if I get one in the next month or so I’ll be happy and, all things considered, they will have done a great job I’d be happy to buy them a beer.

poppe avatar

Well, hindsight’s great.

I don’t think anyone REALLY could have predicted the way the R-Pi was going to blow-up in the media.

You said it yourself “I cannot believe that people didn’t expect problems and delays along the way.”

JamesH avatar

I’m defo. up for beer.

steviewevie avatar

Well said. And I’d love to chip in to the “buy them a beer/painkillers for Liz’s knee” fund too !

Gregor avatar

Exactly. And it also doesn’t help the case if you “advertise it” on media and explaining how any kid can tinker with it. Perhaps the barebones boards could pass as development platrom if a cased certified version was there. I am surpirsed no one though about certificates. Then again no one ever does. People always assume you can just move any kind of goods freely arround the world.

liz avatar

We did think about certificates, actually (and we took legal advice) – the conclusion was that the Raspberry Pi wouldn’t require CE marking until it was a cased product for education. As you’ll see if you read the post again carefully. This was kind of sprung on us; CE marking was in our business schedule, but for later on.

piopio avatar

ehm, no offense here, but really, get legal advice somewhere else, your credibility taking a hit there…

JamesH avatar

No offence, but are you a lawyer specialising in this area? If not, I suggest you may not be the best person to comment.

The problems have arisen mainly because of demand quantities which means advice sought for a batch of a few 10 of K’s isn’t necessarily applicable to 100 of K’s

piopio avatar

For what it’s worth, I had my fair share of.

This 10k myth is nonsense, but if you want to think so, you’re free to.

Abishur avatar

Achievement got: Banned for trolling!

Joe avatar

Cant you just drop RS and Farnell and sell these from a simple Paypal shopping cart site? the majority of buyers wont care about CE marks – we just want to get our hands on the boards!! put up a Paypal “Buy Now” link, and start posting them out :)

arm2 avatar

That would open up how batch of other problems and is impractable

Alexander Langer avatar

Just hope you thought of making an unboxing video! ANy idea how long compliance tests can take? I mean, put it in a chamber, get your data, let someone sign some papers.. 1-3 days?

JamesH avatar

It’s more of an unwrapping video – they are in anti-stat bags.

Burngate avatar

I keep looking at that photo, and I keep thinking, they’re not Pis, they’re Easter eggs

Dao Ming Si avatar

Hi! I live in Guangzhou, China. I saw that your factory is in Shenzhen. Any chance I can get it right away? (I saw the pictures, and seems there’s heaps of them stacked!)

simonIOW avatar

Shhh, people will be asking you to hop over the fence…

Topher Brink avatar

And all of those ones in the picture are for me? How nice of you all!

Fraggle avatar

This whole sorry saga comes down to one thing: not having the strength of your convictions.

The original plan was sound. Disappointing for many of us, but sound. Get the first 10k – see how quickly they sell, then decide what to do next. Would have worked fine. No-one would have cared whether they were certified or not.

Ironically, the RS/Farnell plan was also perfectly sound. Get certified now, let them handle manufacturing. Ramp up is much quicker after a small startup delay. But it only makes sense if you’re expecting more than the initial 10k to sell quickly.

The problem was sticking with Plan A right up to the last minute and then switching to Plan B. All the confusion and crossed wires can be explained as a result of changing your mind from “10k might not sell that quickly” to “10k will go in 30s, better get the next batch ready”. I understand why you did it, but if you’d stuck with one or the other, you’d have a lot less frustrated people not really understanding what’s going on, which is a shame.

Not having any clue where I am in RS’s list of registrants (I was really ticked that you tweeted that the registration page was wrong, else I’d probably be near the top!), I have no idea how long I’ll have to wait. Oh well, still looking forward to some Pi when the time comes! :-)

JamesH avatar

The problem with plan A is that the Foundation is not equipped to deal with the level of demand. We are unable to buy the next batch without the proceeds from the previous batch, so ramping of production would take ages. 10k batch, 15k batch, 20k batch, 25k batch – only 70 k supplied after 4 batches. Instead of getting your boards within a couple of months it would be more like a year! Lets say demand is currently 250k boards. That’s about 8 batches. At say 1.5 months per batch. 1 year minimum. Add on top of that the need to have a website that can handle the demand….which costs money….and a full time employee to run it all …which costs money….

Fraggle avatar

Don’t get me wrong, I understand the logic for bringing in RS/Farnell. On release day when I saw the plan, I thought “That’s genius!”. Then it turned out to have been a last-minute change (as evidenced by RPF never really knowing what RS/Farnell are doing, CE requirements not being discussed, etc.). *That’s* the bit where it went wrong. Sticking with Plan A would have been interminably slow, yes, but everyone knew where they stood. It’s not the problems, but the confusion about the problems that’s making you look bad.

JamesH avatar

Actually, it wan’t a last minute change – you don’t just get people like that on board overnight. It was a couple of months work to get that all sorted. Or not all sorted as has turned out.

Zoltan Arvai avatar

I think this is the major problem: communication.
The foundation tells an official plan and says that plan is going well and this is the only plan.
After weeks of consistent communication turns out that a very different plan is going on for weeks or months.
I know the foundation tries to have a very positive communication for marketing purposes.
But this not-always-honest communication is very frustrating for the community that following the happenings.
It seems the communication has a long delay (or lag) before turns out what is really happening.

Otherwise, this project is very needed for young and not-so-young people who has interest in computer sience and it could start a new age.

Stuart Lea avatar

Sorry, but to produce 10k of anything as a startup takes balls. Focus groups and the likes can jump around all they like, but you never know how many you are going to sell until you open the doors to the shop. I know and it’s scary…

We can all look back now and say how things should have been done based on the number of orders taken and expressions of interest, which probably top 500k and to say that they lacked “the strength of their convictions.” But that would be wrong.

Fraggle avatar

You’re misunderstanding my point. I’m saying I agreed with RPF’s original postion and they could have sold the first 10k in the manner in which they had originally planned, for all the reasons they originally enumerated (which I agree were valid), but at some point late in the day they decided to bring RS/Farnell on board. I don’t see why the foundation couldn’t have sold the first 10k themselves while still talking to the distributors. I don’t see what was gained by using them from the off when they clearly weren’t ready. There’s no point talking about 500k when we still haven’t seen the first 10.

Nevertheless, what’s done is done. It will all be sorted in the end.

scep avatar

“I’m saying I agreed with RPF’s original postion and they could have sold the first 10k in the manner in which they had originally planned,”

That’s where the misunderstanding occurs. You are talking about the original position — the position where Eben and Liz and the rest of the team would have quite happily packaged and sent 10000 RasPis from their kitchen.

This was before they had to answer thousands of emails a week; deal with a forum of 15000 people; travel round the world promoting it; decide which of the several interview requests per day they should accept; do those interviews; and still deal with the manufacture and testing and logistics of a prodcut that has hundreds of thousands of people queueing up to buy it. On top of a day job ;)

So no, they really couldn’t “have sold the first 10k themselves”, not once it all snowballed.

simon avatar

“This whole sorry saga comes down to one thing: not having the strength of your convictions.”

Tell you what bud – go and invent something innovative, something that tries to solve a problem and is potentially world changing. Remortgage your house and give up your job to follow this dream even though people are slagging you off and you don’t know if it will be a success. And then when you have a few problems we’ll all come along and post on your forum (you will have an open forum where you post everything that is going wrong as well as right, won’t you?) that your whole sorry saga comes down to one thing: you not having the strength of your conviction.

Post like yours make me puke.

Mike foxton avatar

If the device passes the CE tests for emissions and susceptability, i don’t
Think you will need the static proof bags for shipping.
I designed the Dragon home computer in 1981; you need careful PCB layout to meet the emissions testing, but it can be done. Good luck.


jaygee avatar

I hope the anti-stat bags are just a standard packing rather than a necessity. Those of us used to handling circuit boards know to at least touch a ground point before any further contact with the circuitry. However many of these boards will go out to people who are not used to any special handling precautions to prevent static damage so I hope that the Pi will prove to be reasonably static proof or there will be disappointed customers or, worse still, customer returns.

Hopefully this has been considered and checked but at the very least an included note about basic bare board handling precautions to avoid static damage might be a good idea to minimise problems. (It does beg the question of whether the device will ship with any documentation at all – I hadn’t expected this in the initial experimenter phase but now that interest has gone wider, expectations, as with CE testing, could change.)

mahjongg avatar

Especially Farnell doesn’t ship anything, not even a simple resistor in anything else but antistatic packaging. Its simply standard practice in the industry.
So yes, “… just a standard packing rather than a necessity.”.
Indeed, it means exactly nothing that they do it with the RPI.
The RPI isn’t any more “static sensitive” than for example the average USB hub.

Stuart Lea avatar

An uncased USB hub – possibly yes. Cased certainly no.

piopiopio avatar


Peter Green avatar

“Its simply standard practice in the industry.”
Actually it seems to vary a lot between suppliers

RS seem to put nearly everything in full on conductive anti-static bags (metal or carbon based). I’ve seen a complete boxed evaluation kit place inside an anti static bag by them before. They also have this habbit of pre-packing things in standard sized bags unless you select “production packaging”.
Farnell seem to use the pink “static dissipative” bags for non-sensitive components and most things are packed to order (though you occasionally see stuff that was clearly prepacked).
Rapid, mouser and digikey all seem to put the non-sensitive components in ordinary plastic bags.

arm2 avatar

Liz, Eben, Pete et al. nill carborundum!
You mention category A & B requirements what items/usage needs the higher spec category B?
The EU Directive on EMC at
Doesn’t list Categories.
n.b. Now if the Foundation had known that changes A, B & C to the plan would be forced on them, things would be different. We all have hindsight but it is only the critics who think the foundation should have been prescient!

~Neil~ avatar

Hindsight is 20:20. Or as Danish physicist is reputed to have said: “prediction is very difficult, especially about the future”.

~Neil~ avatar

Ooops…. that’s Niels Bohr.

Lars T. Hansen avatar

In danish it is written: “Det er svært at spå – især om fremtiden”
“Det er” = “it is”,
“svært” = difficult”,
“at spå”= “to predict”,
“især” =”especially”
“om” = “about”
“fremtiden”= “the future”

As a Dane i can tell you that it came from a member of the danish Parliament “Folketinget”, from 1935 to 1939. Unfortunately, it is not know, exactly which member of the parliament had said it.

Source of information (in danish):—isaer-om-fremtiden/

scep avatar

As a non-Dane I can tell you that it precededs Bohr, Petersen and the Folketinget ;)

Jose avatar

This saga with the CE markins shows what is wrong with todays regulations. You should be able to sell a development board without any restrictions and certifications needed. Jobs and Wozniak would have failed to start Apple from a garage, because of safety regulations and so on.

On another subject, can you just put the boards in eBay?



Stuart Lea avatar

and someone keeps thousands under the stairs and takes a truck to the post office every day.

It may be painful now, but going with RS and Farnell is for the best in the long run.

Snar avatar

While you are talking to BIS about the “requirements” of EMC testing, ask them how they can purposefully allow powerline networking products (PLT) that have been shown to exceed EU testing limits (CISPR22) by 40dB (1000x) and cause interference totally unchecked?

Conformance to EMC standards such as EN55022 is non-mandatory. If you want the pass certificate, you can even get a test-house to falsely declare a pass, which is how PLT manufacturers have claimed conformance. (I have proof via FoI requests)

The law is the EMC Regulations 2006, which state the essential requirements as:

“2) Equipment shall be designed and manufactured, having regard to the state of the art, so as to ensure that—
(a)the electromagnetic disturbance it generates does not exceed a level above which radio and telecommunications equipment or other equipment cannot operate as intended; and
(b)it has a level of immunity to the electromagnetic disturbance to be expected in its intended use which allows it to operate without unacceptable degradation of its intended use. ”

And Ofcoms view on this is that there has to be a victim receiver for the law to be breached. Even if it is found to be an issue, they will act “proportionately”

It seems to me that if you have pots of cash and piles of junk to peddle, you can bring pretty much anything to the market unchecked, safe in the knowledge that BIS and EU will not allow barriers to trade (even if they happen to be legal!). On the other-hand, if you are bringing a product to market with intrinsic value but lower volumes, expect to get picked on.

pauldow avatar

Back in Ye Olden computing days, some people found that an AM radio would pick up different tones based on what was coming out of a line printer it was sitting on. They then wrote code to output to that printer, and they could play a song based on the combination of lines of characters. It wasted a lot of paper, but that’s just a side effect of development.

You could market the RPi as an experimental music device that is intended to emit RFI.

FoxTossing avatar

A there-pi?

Jason avatar

On a slightly lighter note, once the CE approval are you going to have to find the running temperature of each Pi for VAT purposes?

If it runs at ambient air temperature those Pi’s would surely be VAT free, only those running hot would have to have VAT charged?

eben avatar

Heh. But I believe the famous bagel case

establishes the precedent that as long as the Pi only gets hot in order to be more crunchy, it is still VAT-exempt :)

Babak avatar

Блять, вы заебали, когда наконец начнутся продажи???

mahjongg avatar

Naughty naughty.
This won’t help you getting them any sooner.

Lynbarn avatar

In fact, even in Russian, it could get you banned!

Kevin Wiltshire avatar

Liz were your 3 words to Eben “Glass of wine?” :)

Sandrokan avatar

Sorry Liz but I really think that the Raspberry pi MUST be CE compliant. It has been designed for children, not for electronic geeks or I’m wrong????????

jbeale avatar

This has been discussed a few times… the first 10k batch was always intended “for developers” to get the needed software work done, so there is a working set of packages in place that kids can use. That could include hardware and other accessories also. Then the “educational” release for schoolchildren was (and still is?) planned for later, which would include a case, and CE/FCC testing.

Scott Bridges avatar

But weren’t the electronics geeks children at one time? There has to be someone around here that is older than CE compliance. That old non-compliant stuff didn’t kill us. Personally, I’d pay extra for a R-pi that radiated just enough to make my mobile phone unusable. I could enjoy some quiet computing time. Maybe someone will create a pi project that creates a phone free bubble of about 3m. Quiet dinners, face to face with real people, what a novelty.

Doep100 avatar

LOL, yes that would have its advantages i guess. Being a radio HAM myself and intending to use a couple of PI’s in my radio shack, I would love for them to be very “RF quiet”. But hey, I’m A HAM, so detecting and shielding RF interference is kinda my thing.

George avatar

Oh, I thought it was designed to turn children into geeks?

Midnight Caller avatar

Let’s hope the testing does not take too long, I can’t wait to get my grubby little hands on one.

Adam Hamilton avatar

Sounds like you should just fulfill the non-Europe orders first.

Tass avatar

Thanks for all the updates! I know there are a lot of people out there that are frustrated with the recent delays but as I see it, the worst is over, right? The boards have *sailed* through compliance testing (i.e. the foundation were right to take the approach they did), the first sub-batch are here, the others will follow shortly and best of all, Farnell & RS can start ramping up production on a clossal scale!

liz avatar

We’re very confident they’re fine, but we can’t *definitively* say that until they’ve been through the testing chamber. So it’s closer to say they’re sailing through. Please cross your fingers for a fair wind.

Tass avatar

Just read Pete L’s post on e14 (very informative BTW), so now start to see the larger scale of the testing. Crossed & waiting!

Manuel F Martinez avatar

Hey, look it, mine is there
Glad we all are seeing our boards there.

Imagine if all them were dispensed outside a building!

Kimomaru avatar

Thanks for the update! No worries, mate, do what you have to do to make the launch a success.

janin avatar

When you decided to declare that the Raspberry Pi was a finished product to avoid UK import taxes, what made you think you could also avoid the CE certification?

Abishur avatar

Actually this is a misunderstanding of several events. The first is that in the course of pricing they decided instead of importing every part into the EU and having to pay taxes on each component they would import the board and pay a single import fee for a single product. (so they weren’t trying to dodge import taxes, they wanted to only have to pay them once) The second is that the issue is not “finished” vs “unfinished” product when it comes to CE testing. CE testing is about what the boards are being used for. The original expectation was that while it would be open for anyone to purchase, the main people showing interest would be developers as such the boards were classified as developers boards and didn’t need CE testing. As things turned out, however, the demand for the r-pi has far exceeded everyone’s expectations by leaps and bounds and the distributors lawyers went “wow, with this much demand, I think we should push a head the CE testing”.

Hope this helps.

janin avatar

They did however say this ( : “If a British company imports components, it has to pay tax on those (and most components are not made in the UK). If, however, a completed device is made abroad and imported into the UK – with all of those components soldered onto it – it does not attract any import duty at all”. The intent to pay less taxes seems clear (I’m not saying it’s a bad thing).

It does say “An ARM GNU/Linux box for $25” at the top of the page, not “An ARM dev board for $25”, and also in the FAQ : “The Raspberry Pi is a credit-card sized computer”. It was marketed as such with demos of the board running XBMC, MAME, Fedora, etc. so arguing now that it is a dev board seems a little disingenuous.

jbeale avatar

Ok, but did the original BeagleBoard have CE? For reference… “The USB-powered BeagleBoard is a low-cost, fan-less single board computer that unleashes laptop-like performance and expandability without the bulk, expense, or noise of typical desktop machines.”

Abishur avatar

Well now you’re wording it differently than your first post. In your first post you made it sound like they were trying to something shifty and under the table, but in the second you recognized that they were trying to pay less import taxes. (maybe not how you intended it, but how it came off :-))

Secondly, you can read all over this site and the recent post by Pete L on compliance on Element14’s site and see what people mean by dev board. I think you have an overly literal interpretation of the word when that’s not how it used in a legal sense. ;-)

Nicholas Tiong avatar

Will this affect delivery dates for customers outside the EU? Ie. Australia?

Magnetron avatar

The C-Tick requirements for Australia are for emissions only and based on EU and International Standards so if the product complies with the EU requirements I can see no reason why it won’t be fine for Australia and New Zealand. It is likely that FCC will be achievable too although the limits are not identical. Good luck with the testing !

Ellis Birt avatar

As I understand it, your original plan to sell the board as a development board to hardware and software developers to start an ecosystem of projects leading to the education (& consumer) launch later in the year.

Can you not still do that with the first batch of boards by asking purchasers to declare their intentions as part of the selling process. If a purchaser were to fraudulently declare their purposes as being development, neither you nor your distributors could be held responsible.

That way, if hardware modifications are required, you can make those changes before subsequent batches go into manufacturing.

Tass avatar

I’d be happy to sign/waiver whatever I needed to, but by the sounds of things the bulk of the work is done and it won’t be that much faster than if we just waited for all the testing (provided everything goes well)

Ken Hilke avatar

I live in the US and do not care about CE certification when and where can I get one?

Abishur avatar

I *believe* the answer is that they’ll have to do our equivalent of CE testing (FCC) but from what I’ve heard they’re fairly similar so if it passes one it *should* pass the other. Other than that we purchase it from the same peps as everyone else :-)

elmicker avatar

I recall you saying you were very tight on board real estate, possibly in the context of saying personalised branding on large orders was a no-go. How’re you going to cram CE/FCC marks in there?

Abishur avatar

The tight real estate reference was speaking more to the data/electricity lines going through the board (thus the need for 6 layers), but even if they couldn’t fit the litany of certification they’re going to have to get for the board, I believe it is acceptable to have the certifications recorded on a separate piece of paper shipped with the product :-)

jbeale avatar

Looking at the top side of the beta version, there’s some unused area available for silkscreen text & logos near the left hand side:
and I think there’s even more room on the bottom side.

eben avatar

Indeed. I think we’ll squeeze it in somewhere.

vimal kan avatar

Ordered Feb. first Farnell said expect delivery end of march then said end of April now saying wait…. seems the guy at raspberry are a nice bunch. I hope it does not turn out to be Raspberry Pi In the in the sky for their sakes,

Stephen Hammond avatar

If you sold them from China wouldn’t that negate the problem ?

Jason Kridner avatar

For the record, as a member and leader within the community, I agree that it seems a bit silly to expect development boards to have FCC and CE compliance certifications. Nevertheless, when you start reaching beyond the basic set of design engineers who are knowledgeable enough to avoid interference with other electronic circuits, there starts to be a seed of reason—though I honestly don’t think the Raspberry Pi (or the products for that matter) crosses the threshold.

I do hope that the more zealous members of the Raspberry Pi community will begin to take a slightly more reasoned approach to communicating about other open source focused low-cost development solutions as the claims have been consistently outrageous. While it is certainly possible to build a $35 development board (computer) with a TI AM335x processor, we made other choices we thought best served our community in the launch of the BeagleBone—and those choices were not focused on making more money of the sale of the boards, except perhaps by selling more of them. Paying for operations like certification was one of those choices.

I personally look forward to getting a Raspberry Pi when it is available to me and to whatever benefits it provides to open source on affordable, low-power ARM processors where we all benefit. Our communities have significant overlap and we’d both benefit greatly from more accurate communications in the future—and perhaps even more corrective action on miscommunications.

The project has been fairly consistently praised for its openness and honesty and I hope to continue to set an example I’d be happy to see the rest of the open source hardware community follow.

JamesH avatar

Hi Jason, thanks for the comments.

Just out of interest, I’d be interested to know which claims you regard as outrageous, as I would like to put those right if possible. As with the Beagle project, the foundation is trying to be as open as possible, but we do like to get our facts right. Feel free to PM me if you feel there are aspects that should remain private.


Jason Kridner avatar

Frequently when the project has been mentioned, the data has been inaccurate. Objectives, costs and performance have all been mischaracterized. I’d just suggest verifying any claims and looking for ways to do better side-by-side comparisons to best enable developers to make an informed decision on where and how to participate–many of whom will happily participate in both communities to advance our shared objectives and choose the best tool for individual tasks.

JamesH avatar

I’m afraid that we don’t have time to verify all claims made by posters to the forums (or to posters on other forums!) – however wherever the mods/admins see incorrect data (that we know is invalid) we do try and correct it. That said, we do like our own facts to be correct, so I have emailed Eben to look in to any inaccuracies that may be promulgated by the Foundation itself.
I personally have no idea of the capabilities or costs of Beagleboard products, so refrain from commenting, but I am sure, as you are, that each product line has advantages and disadvantages over the other.

eben avatar

I have mailed Jason to arrange to have a chat.

Every time I try to say something nice about BeagleBoard it comes out wrong. As an example, my comment to Gareth that someone is charging over the odds for components came across as an allegation that their members are all driving Ferraris paid for by overpricing their hardware.

Bollocks. More evidence that I need media training :(

Bill Holly avatar

All it needs is an instruction to mount the board in metal case before use “to prevent possible random outside radiation getting into the board”.
OK, so April 1 is just past tense.

George avatar

Just had a look at my Microchip 32bit Development board (Explorer 16) and cannot find any CE or FCC markings either on the board or in the documentation? Looks like you have paid the penalty for success – you should have known how we hate success in the UK!!

Mike avatar

Why not self-certify the units as passing and leave it to someone else to prove they fail. I suspect no-one would bother.

Paul avatar

What about the ones that are shipping to the United States?

Tony avatar

So it’s Liz that puts the cream on the raspberry Pi

capt jack harkness avatar

spoken with BIS this morning, and they have confirmed that, given the volumes involved and the demographic mix of likely users, any development board exemption is not applicable to
Im not sure why volumes matters and especially the demographic mix.
Im pretty sure this is a HEAVILY male, balding clientele with a touch of young geeks so unless many women in their 80’s are interested, I fail to see where demographics come in.
And the volume speaks of the interest towards the project. If 100 geeks are into something, its quaint but if 10,000 geeks from around the world are into something, somehow its a problem?

The fact that people are buying the first batches without cases hints that this is for geeks to test and play with as opposed as Joe Consumer who wants to plug in a cheap computer in his 60 inch TV.

Im not saying that I wont be buying some later on for those kind of uses when cases are out but Im not thrilled that a projects popularity is given as reason for not receiving an excemption ESPECIALLY since many of these Pi’s WONT be shipped to UK.

Malcolm avatar

Great news about the commencement of deliveries. In can say with some certainty that there is going to be a shortage of raspberry pie in this part of Australia (possibly globally) when my raspberry pi arrives – due to the party I have planned

Comments are closed