Compliance testing - Raspberry Pi

Thanks for the update.

+1

Curious how you’ll ensure such certification without cases, but best wishes nonetheless!

In an e-mail from RS Poland they write about reliability tests. Maybe it’s a bad translation.

Ha! Yes, that’s a VERY bad translation.

Does this mean further unanticipated delays to shipping?

Texy

At this point, we simply don’t know. We’ll let you know as soon as we do.

A more direct question: Is shipping going to be delayed until the boards have passed this testing?

Does this effect shipping at all?

Thanks ~~Liz~~ Eben, that also answers some of the questions on this forum thread.
http://www.raspberrypi.org/forum/general-discussion/new-rs-email-compliance

[edit: sorry Liz and Eben]

Except it was even who posted this

Er, thats thanks Eben, not Liz !!!
T.

The big question is, will people start to get their Pi’s during April? (Yes/No/Unknown)

I believe it’s been told that one if not both distributors have decided to not release the first 10K until they can release a huge batch (a fact which I think is lame and defeating the whole point of getting the first batch… but what are you gonna do?). It was said elsewhere that Liz or Eben are going to be talking to them early next week so we should be getting some updated information shortly there after.

you didnt read the article.

developemnt boards dont have to be CE marked as the end customer doesnt really exist. it will kick about on an engineers test bed for a few weeks before being replaced with a PCB that will be CE marked as a product.

PI’s are clearly going to an end customer (the general public) and need to be CE marked as a matter of law. if its not, then RS will be up for fines and they wont do that will they?

Arduinos are not CE marked (i cant find any mark on mine), but they are arguably MORE of a developement board than a PI.

this is clearly BS
bare PCB is not an end product

its just another lie that allows them to keep my money for over a month while not delivering :/

I bet they never had those “initial 10K batch” :) vaporware

First of all, no one is talking about a bare PCB

Secondly, if you’re trying to be funny you’re not really being successful. The admins (me) and the RPF are really patient with a lot of crap people send our way, but get your facts straight. No one has taken your money. Until the boards ship you haven’t been charged a dime.

If you have a complaint or issue than please air it respectfully, such baseless trolling will get you banned.

Odd, I saw one of those original batch vapourware things only this morning. Seemed solid and not vapourous at all.
Are you sure they have your money – have they actually charged your CC?

… Who are you talking to Tom? Your comment doesn’t fit in with the question or the response.

Arduino Uno (and ptobably others) are now CE marked

Cheers Eben. A good clear statement and very good point about the BeagleBoard.

Got my compliance email today from RS. Hats off to them for their frequent updates :)

why make EVERY dev board out of the first batch complient?
IMO its time wasted if complience ain’t needed… which would lead to the ‘first batch’ status redundant if subsiquent batches have caught up by dispatch.

If such a device does cause interference with another device, the operator is legally responsible for eliminating the interference by placing the device in a shielded enclosure, rearranging connected cables or adding radio-frequency attenuators (e.g., “chokes”) to the cables (which can act as trasmitter antennae) or, in the worst case, moving the device away from the interfered-with equipment until the interference is diminished. Fortunately, the ~1 watt of total power consumed by the R-Pi is much lower than typical computing devices (desktop computers and displays can consume the better part of a kilowatt), so, it usually doesn’t take much to resolve the few cases where interference might occur at ~1/1,000th the power level at which an R-Pi operates. A Faraday cage in the form of a metal case, or made of copper foil or fine wire mesh that can be folded around the inside of a plastic enclosure and soldered to a ground wire, would be perfect, although even aluminzed Mylar plastic sheet can work – but, good luck soldering to that (a ground isn’t usually necessary, though). Whatever you do, don’t let the shield come into contact with the board’s etched signal/power traces or component leads, or you might find yourself at the wrong end of that very long waiting list :(

When I was in the Navy, I participated in an electromagnetic security vulnerability evaluation of computing and network systems in an overseas facility, and it was very interesting seeing a CRT monitor’s image clearly discernable on an oscilloscope on the evaluation equipment cart … hundreds of feet away – a high-gain directional antenna and a preamp did help, though. Fortunately, the entire building was built with a Faraday cage inside the exterior walls, roof and under-flooring, so, our deepest, darkest secrets stayed deep and dark :) It was fun reading people’s computer displays over their shoulders ;)

TEMPEST emanations? I read a fascinating paper about that a few years back… the researchers even came up with a proof-of-concept font which made certain text look like completely different text to the eavesdropper (I think their example made “Oxford” look like “Cambridge” or vice versa). Amazing stuff.

“I can neither confirm nor deny … ” ;)
I’m sure most people will be able to read between the words (it’s only one line, so, that would be harder to read between :)

I have worked in a Tempest-certified facility. The whole building was a big faraday cage. No copper cabling except power lines went in or out. I do not know how they filtered the power lines. There was a printer for writing declassified reports located outside the tempest area, and it needed a centronix-to-fiber interface on each side of the wall so the printer cable couldn’t be used as an antenna into the tempest area. Any piece of electronics containing memory chips that entered through those doors was not allowed out again. even the eeprom chip that holds the bios config was enough that computer main boards couldn’t be returned after service.

The BeagleBoard example is a bad one, since – despite that note – the BeagleBoard *IS* FCC part 15 certified. From their 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.”

The Raspberry Pi can’t escape this, especially since it’s being marketed to the public as a general purpose computer. Does these post mean the Pi has not been certified for anything yet? This is not good…

Just had a look at my Arduino and sure enough it has both FCC and CE certifications too.

We can all have CE labels for our devices ;-)

http://www.dealextreme.com/p/9mm-round-packing-ce-regulated-self-adhesive-label-stickers-black-192-x-15-pack-121807

for 10 000 units it should be about $12 ;-) I mean… CE means Chinese Export, doesn’t it?

Based on text Eben quoted, I’d guess that the BeagleBoard was intended to release without full worldwide compliance testing. Given the text you quoted, it apparently tested OK for Part 15 at some point after the first draft of the manual, which may or may not have been after initial HW release, but FCC is also separate from CE mark testing.

Sadly that means it can not be sold in the USA unless it is a “kit” and I’m sure there isn’t enough work that people can do to call this a kit since it is all surface mounted components. FCC testing and certification is going to cost a pile of money and take several months. I hope there is a work-around for this.

UL testing requirements = bye bye raspperry pi

Are you Psychic?

What are the UL requirements?… looks like it’s class 2 ….

so much hype and news about this product but no actual product as yet. If there are more delays the chinese will knock something out at half the cost before the PI has even made it to market. The PCB design and components can’t be difficult to copy.

We’ve mentioned it before, but it bears repeating: we actively hope that other companies will do what we are doing, even down to copying the thing. We are a charity. Our goal is to get cheap computers into the hands of kids. Whoever’s making them, we consider it a win. So copy away if you think you can.

Then if the definitive goal is to put the cheap computers in the hands of the kids, while not put the entire design on the open and let anybody make it and compete for best price ?
Just find a license that may keep providing some funding to the foundation to focus on its ultimate goal and not have to waste time with sourcing components, manufacturing, distribution and angry developers.

Well part of the reason is that they don’t want their suppliers being attacked for parts, there’s some NDAs, and most importantly, there’s the broadcom chip, it’s not available in small quantities. You have to get a licensing agreement with them to get anything. That shouldn’t be a problem from anyone who can backwards design the board, but more so for joe schmoe random user. Still, while I say that, the RPF has mentioned their desire to do just that if they provided they’re legally able to do so and reasonably sure their suppliers won’t get mobbed.

Thanks, but I didn’t ask *you*, I’m starting to think that you are RPF’s answering machine. I was not referring joe schmoe, I assume that when you say “them” you are referring to Broadcom. And I don’t understand what you mean by “attacked” and “mobbed”, besides the Broadcom chip all the rest are off the shelf parts.

Answering machine? I kinda like that, a machine that replays the message left by the person you’re trying to reach, it would be rather silly to distrust any message left on an answering machine especially when said answering machine is updated so regularly. I figured you weren’t talking about Joe Schome which is why I only mentioned him off hand in one sentence near the end of the post, everything I said applies to big players too. Apparently, there’s enough ability to take a part list and match it to a supplier (wired magazine was able to, they made a bunch of bad assumptions on price and cause a bit of drama). Moreover, some (we’re not told who just some) of the suppliers worked special deals with the charity. I suppose we do know that broadcom work a special deal where they let them order a smaller quantity than they usually allow. The point is people tend to make a stink if they don’t get the exact deal someone else got, and the RPF wants to make sure it’s suppliers will be reasonably protected from such backlash before they release any such list. (I mean even if no one ever used the list, current companies also getting supplies from them might get upset if they saw they didn’t get as good a deal as the RPF. They shouldn’t since that’s business, but people are illogical beings who make a stink over the most ridiculous nuances… *beep* :-P )

Well, I doubt even the Chinese could make a knock off and then sell it even as cheaply as the RPF is doing theirs, but that’s because the RPF is a non-profit, they’re already largely selling the boards nearly at cost. Additionally, while they would be able to reproduce the board they wouldn’t have access to the code to make it all work so they’d be sitting there twiddling their thumbs looking *quite* silly. Furthermore, since the next batch of boards are, I believe, already having their parts sourced and put together, it would be rather physically impossible for someone else to copy, source and build the boards from scratch before the R-pi is ready. Still, ignoring all that, the RPF wouldn’t really care about being undermined on price so long as the boards go out into the hands of children who needed them and the quality of computer science education as a whole was improved as a result. But maybe I’m being too over literally to your post and you’re just trying to say that while the product most assuredly exists, you’re frustrated at the amount of time it’s taking to reach you. ;-)

It would be really silly to think that the next batch of boards are going into production before you get the results of the compliance tests. Nor distributor or manufacturer will put the money for the components and fabricate the pcbs until they have green light from the certification process.
On the other hand, what “code” are you talking about ?

He’s talking about the boot sequence for the GPU. (My plane is delayed. I get to spend what could be peaceful minutes imbibing airport lounge gin replying to you guys.)

Cool, different but related topic, what is your target total cost/price for the Raspi, with enclosure, pre loaded SD card, power supply, keyboard and mouse ?
PS. Have a safe trip

True enough about the next batch, I do know from what has been said that the parts are being sourced. But my point still stands that CE notwithstanding it would still take much much longer for someone to build the copy, source, and build the boards from the ground up then it will take for the R-pi boards to get released.

Then, since you seem to have the crystal ball … When the boards will be released ?

Sorry, I don’t have a crystal ball, just using common sense to draw a conclusion ;-)

Common sense in electronics design produces short circuits, use math. :-)

Hm? I might be taking the comment too literally (it’s late my time but I’m not really ready for bed), but I’m not actually talking about the electronic design. I mean a x-ray of the board and a magnify glass to get the part numbers will let you physically reproduce the board. The time killer is getting the parts. It’s a simple thing to get 100 pcs of a certain part, but 10K, 100K? That’s going to require a nice lead time (the 6-8 weeks kind) and probably some negotiating and licensing agreements. Repeat for each part and then triple that to get on board with broadcom and you got a clone that will be out… before the end of the year if you’re lucky? I certainly don’t mean to say no one will do it. I’m just saying that when you have two distributors already getting the parts and gearing up to make these things it’s statistically improbable for a clone to come out before the R-pi boards are released. :-)

A long discussion followed this post, but if you’re curious about a “Chinese” solution to the low-cost ARM-computing problem, simply search for “allwinner SoC” or “allwinner tablet”.

As to the question of cost of competing solutions, it can be settled now by answering the question, “Can the Broadcom solution (in the RPi) match the Allwinner SoC’s cost of US$7 [in 100k batch]?”

Am sad that I have to search for alternatives to the RPi, but if I recall correctly, the 10k (or more units initially) will be for developers who would not be asking for more parts or any level of support. Now it looks like the “educational release later in the year” has been fast-forwarded, and before we know it, we will be looking at a total-system-cost, and be jolted that the alternative would look much better.

The Allwinner’s OK, but power management is probably an issue, but since most Raspi’s are powered up, not that a biggie. I cannot comment on the pricing, except to say the Broadcom chip is competitive. Another bit plus for the BRCM chip is the amount of tech support available to the Foundation. ie. the people who designed the GPU are on board and helping out here. You don’t get that with the Allwinner. And no, a datasheet won’t help in these circumstances.

Don’t forget c-tick for Australia. Not that it should probably need testing if it is for development purposes or a component to put into another product. But it is probably a good idea to get this all organised now.

Just to add quickly: Eben and I are hopping on a plane in 15 minutes (I’m writing this from the gate), so we’ll be incommunicado for most, if not all of Saturday, and then at the Beeb@30 do for most of Sunday, but I’ll be checking back in here as soon as possible after that.

Hard plastic clamshell type “Faraday cages”? Shouldn’t be too hard a trick with a few “anti-static bags”, a bit of glue and a small jewelers box you think??
Choke filter for the input power of course,,,

Just thinkin’.

I don’t know about EU/CE rules, but for the US I’m pretty sure there is nothing in US 47 Part 15 (FCC rules about RF emitting devices) that provides an exemption for “development” devices (whatever those might be defined as – clearly, internally used engineering models and prototypes are not required to be certified (although they still need to have their EMI constrained – ever do development inside a Faraday cage?), but those aren’t sold on the open market). In any case, those rules almost certainly would consider a device that is commercially manufactured in quantity 10k for sale to end users (whatever category you want to put those users in) to be a covered device. It appears that 15.101 and 15.102 are relevant to the Pi. I assume CE has similar rules (albeit possibly with the claimed exemption for “development” devices). Whether the existing Pi design would meet the requirements can’t be determined without testing that apparently wasn’t done as part of the development. (It also isn’t clear quite what would be required – normally a (PC class) CPU board is certified in connection with an appropriate case, and I believe PC cases have to provide a certain minimum level of Faraday shielding that allow CPU cards to be certified against a “standard” case. Since no such standard case has been defined for the Pi… But the Beagleboard, which I think has similar issues to the Pi, _was_ certified.)

Whether the RPi developers thought this through or not, it’s pretty clear that Farnell and RS (as not just distributors but apparently partner manufacturers) need to worry about it since there is the possibility that they would be on the hook (likely along with others) if the units require some certifications and ship without them. So I can see them imposing a delay while they deal with the issue. The question I would have is why this didn’t come up earlier rather than not until units were actually supposed to ship?

Raspberry Pi is a charity run by a small group of good people using their spare time to try and do something positive for the education of children by designing this product.

They are not a start-up, there is no VC funding and they have significant limitations to the availability of time and money to design and produce this product.

When evaluating the critical path to releasing this product they researched the dev board exemption and decided on a 2 stage launch to get a smaller number of devices into the hands of people who could help develop the software that would be needed for the full educational launch later in the year.

Because of this they decide to proceed with manufacture a non-tested dev batch of 10k units, they were fully aware tests would need to be run before the second stage educational release.

Finally, it became apparent that demand for the first stage dev board would be so overwhelming they needed volume manufacturing assistance sooner rather than later so they swapped distribution to RS/Farnell late in the stage 1 launch process in the hope it would get more boards to customers sooner and would free up very limited foundation time to work on the next stages of the product.

As you describe, because RS/Farnell need to move into high volume manufacture immediately they wish to conduct the testing now, hence this (short we hope) hold.

I don’t think that you realise that you cannot test a board into compliance. Some boards will never pass and you need to start again. I know, I’ve been there.

There are two parts to compliance: adherence to technical standards, and certification. I don’t believe that anyone has claimed that the ‘Pi doesn’t meet the necessary technical standards, and they would certainly have been taken into consideration during the design stage, so it is perhaps just a matter of getting the paperwork in place. Agreed it is unfortunate that, after all the other delays, this will add further to the time before ‘Pis can be shipped to customers, but unless and until we know different, we
should perhaps assume that the first batch (once the ethernet port has been repaired) is at least standards-compliant, even if not yet fully certified as so.

On a general point, things haven’t always gone according to plan, and there are many lessons to be learned from recent events, but I have every confidence that, given a little more time, effort and patience, the Raspberry PI WILL be released onto the world, to great effect, and the RPF (and their partners) will learn from the lessons, and move on to even greater things.

Communications have suffered of late, whilst Liz and Eben have been on holiday, but I’m sure Liz will sort that out once she is back, and it does go to show just how small the Foundation is. Not quite two guys working in their garage (which is just how Apple, HP and others started!), but they are a VERY small team trying to do a BIG thing. I’m not leaning on their charity status as an excuse, but I am asking some of those who are dissatisfied with recent progress to remember what it is that the RPF are trying to achieve, rather than just trying to satisfy their own need for instant gratification. Like your first beer, your first kiss, or even your first paypacket, your first ‘Pi is something to enjoy in anticipation as well as in achievement…

It doesn’t really matter if you are a saintly charity or a satanic multinational – (at least in the US, and I assume it is true in the EU also) if you are commercially selling a product, you are responsible for it meeting mandated EMI regulations. If you cannot do that, you shouldn’t be selling the product. Saying that “this isn’t the real product – that one will sell in a case later in the year and be certified – this is just a `test’ or `development’ board” while selling 10k of them to whoever asks doesn’t fly either – you are still bound by some rules. (Yes, there are OEM board makers who can push the certification off onto whatever manufacturer/packager buys their products (though if they radiate badly, they aren’t going to find customers), but that isn’t who are buying Pis. A few Pis may be going to people who meet some of the relevant definitions of “test” (which doesn’t include actual _use_) but not many, and I believe no one who is actually buying one meets all them (at least in the US).

(And by the way – for many of the products I’ve worked on in the past, 500 units would have been considered “high volume manufacturing” – 10k would have been _way_ more than the conceivable lifetime build…)

I don’t know about EU/CE rules, but for the US I’m pretty sure there is nothing in US 47 Part 15 (FCC rules about RF emitting devices) that provides an exemption for “development” devices .

Wrong, I have got many TI development boards that state they have NOT been FCC tested.

Well, most boards aren’t tested by the FCC, but by independent labs. That said, there are certainly boards that aren’t certified (including some classes of boards) – this isn’t an “exemption” for some instances of a particular board, in the sense of removing the requirement because of a declaration by the seller about how those boards will be used, for example “for development”. In the case of non-certified boards, either the system integrator is responsible for certifying the system as a whole, or the end user bears responsibility for any EMI. This isn’t a trivial responsibility – in most cases a user will get away with it, but if there are complaints, the FCC can assess users that cause interference using non-certified systems or components fines in excess of $10K per instance. The TI development boards you are referring to are almost certainly “true” development boards, used in a commercial setting, and (one hopes) by people with some understanding of EMI who are making some efforts to control emissions using shielding of one sort or another. Most of the Pis are likely going to be used in residential environments (which have much stricter interference rules) and without special efforts to control interference by users who probably don’t know how to or are even aware of their responsibilities. Not quite apples to apples.

I find it very odd that it could be possible to be delivering this device into the marketplace without it having been compliance approved. To be talking about compliance testing now on the pi, which was anticipated to be being delivered at this point does not make sense?

The issue was that originally the boards were going to be released as development boards and therefore didn’t need the CE stamp (Liz is a lawyer, she did her homework ;-) ), but due to how many boards there are being ordered and the fact that the first 10K will be released with the main bulk release the distribs decided they wanted them all CE marked. Which… you can’t really blame them for, if someone wanted to call foul they’d be held responsible for selling the board and sadly the world is full of vultures who have nothing better to do than sit around looking where they can find some technicality to call foul on rather than make an honest living.

Edit: I say “the fact that the first 10k…” in the above post, I just want to correct that as it is not a fact (I was in a rush and forgot to check over what I had typed) this is something I’ve heard from one of the distributors, but until I can provide a hard link from where I heard it that needs to be taken with a (large) grain of salt.

and if they fail?

Sucks for everyone?

If they fail, there will be a further delay.
Surely that is obvious?

I truly believe in what the Foundation is trying to achieve and appreciate that it must be one hell of a stressful time at the moment. But, over the years, I’ve had boards that simply would not got through compliance testing without more layers being added or being retracked. What is the Foundation’s position if that is the case?

Redesign the board I would think. 3 month delay perhaps. I’m guessing – I don’t know details.

Ebay!

That might be a usable approach in the EU with regard to CE certification, but I don’t believe that there is any equivalent category in the US relative to FCC Part 15 certification. You can certainly have unreleased items that are distributed (not sure about _sold_) to partners, (sub) contractors, etc., with the onus being on them to avoid EMI issues during development work, but once something is sold commercially, it must have an appropriate certification – even if it is just a self-certification.

and when little Timmy uses his new PI next to granny’s life support machine at home, and the PI emits so much EMI the machine stops working!?

If granny’s life support machine is that sensitive then it would never have passed the required CE immunity testing in the first place.
– Where you yut the thing in a box and squirt really, really high levels of radio at it. High enough that it would count as radioactive exposure if you were in the same room.

So I wouldn’t worry about that.

EMI is more about not interfering with radio signals like the TV, Radio 4 and (more particularly) all the other useful low-power radio transmissions.

I’m assuming that the Foundation are not redesigning a new board for the education market. If so, why wasn’t one of the boards that was made before Christmas sent off to a certified test house before production. If it was, and it passed, then please scan in the certificate to put all of our minds at rest.

Hi Stuart, please see my reply to tgoldblatt.
TL;DR version – RPi is a small team, with limited time and resources and at the time a 2 stage launch with stage 1 not needing certification made sense. Cheers.

I come at it from a different angle, a design and manufacture angle. You only ever want to produce 1 board. If this current one fails compliance testing then the education boards will be a new design, not what you want. Testing would not have taken more than a day at a reputable, certified test house and cost significantly less than the money raised from the auction of the 10 boards.

The tone of the RS email was suggesting that this is a reason for the current delay. However, the Foundation are suggesting that the first batch do not and will not need to be compliance tested and could be shipped immediately. Very, very naughty of RS.

There is no exemption for limited quantities where products are made available for sale to the general public. Since when has complying with the Law been “naughty” ?

The bottom line is this….you cannot test a board to compliance. Sometimes boards fail and you have to start again. So, if they fail, will they still ship the boards or scrap/rework all 10k. If they still ship then there’s no need for delay and they could ship today. They are either compliant or not. If they have to rework the boards again then it shows that they should have done compliance testing before starting the production run.

No but there is an exception for boards labeled as development boards. The issue is that the distributors became worried that the shear quantity of boards being produced would have someone going “wait a second there’s far too many to call this a development board” So the plans were forced to accelerate things up to getting them CE marked. *sigh* the price of success.

Abishur – while that might fly in the EU, at least in the US (and I assume other places), as Magnetron says there is no such exception. “True” development boards (prototypes, etc., used only internally or by development partners and not sold) obviously don’t need to be certified (though there are still liable for EMI), but you cannot just label a commercial product as “well, not really commercial” and proceed to sell it without meeting FCC requirements. If this approach _does_ work in the EU, then perhaps these early boards should have been restricted to sale in the EU. However, I really doubt that this (commercially sell 10k boards claiming they qualify for a “development” exemption in EMI certification) would really work in any geography.

At this point I barely trust any of the updates the way they conflict like this, although RS has at least been reasonably prompt. It seems we’re basically now being told the testing which could and really should have been done months ago (despite not being required for our particular RPi’s) is going to be done (despite not being required…) after the initial batch is already made up and being “sold” (a term used loosely at best right now…).
A very strange way to operate, adding what, another week or two delay to our ordering (which we still haven’t even heard WHEN we will be able to do yet).
And this is the launch that was being hailed as a great success? Really..?

On the whole I wouldn’t be arguing bad faith. Had the original plan been get 10k out over a few months, then work out any bugs before volume production, that make some kind of sense. On the other hand some comments in discussions on other products being “vapour” in retrospect seem a bit unkind. Getting something to market even with a full-scale engineering staff is hard. I’m happy to give it a few more weeks.

It’s interesting that my post about the same time (between 1.30 p.m. and 2 p.m.) did not make it through moderation.
I think that RS are not being truthful when they suddenly claim that delays are because of “compliance testing” with no mention of the dodgy network sockets.

What evidence do you have that they are not being truthful? Or is a guess?

RS did briefly mention a minor manufacturing hiccup a few weeks ago but have said nothing about the dodgy sockets since.

I am not guessing about their motives but, over many years, two of my animal senses have become highly developed: instinct and smell.

… perhaps they don’t mention it anymore because it has now been resolved, and is therefore a non-issue? I’m only saying – I don’t know if that IS the case, but it is a possible explanation.

Reading between the lines, the Foundation hadn’t expected to do compliance until they went for the formal commercial/education launch, expecting this early batch to be considered “development” and therefore exempt. I imagine someone at RS legal has since said, “hang on a minute, 10k run is hardly dev boards, we need to cover ourselves and go through compliance”. Its likley a formality, they know the power levels involved and the components all pass in other implementations, but the process has to be conducted and the device signed off.

Compliance testing is never a formality. It’s caught me out in the past and other more experienced engineering friends. It’s the most stressful part of my job as an engineering manager of a small electronics company.

What I don’t really understand why there are so many folks overreacting on a defensive manner with some lame excuses like RPF is a “charity”, “they intended this or that”, “the end goal is this or that”, etc, etc.
I don’t believe anybody is blaming or accusing RPF of bad faith, or whatever, what is it clear is that the launch was not as successful as many kool-aid drinking heads pretend to believe, totally understandable, RPF screwed up, was there a better way to do it, probably, but the past has to be just a learning lesson for whatever is done in the future, and what it is quite clear is that RPF needs help to move this to the next level, and not talking heads or carbon based answering automatons, cut the hype level a little bit and bring back the end mission to the front page and as the main topic of discussion.
Trying to make everybody happy and putting an unfinished product for the general market and creating so much expectation will backfire and become a nightmare for RPF, the developers and the distributors.
No harm if with a wide criteria you qualify who is and who is not a developer, and yes include some end users as test subjects, and you say, listen the final product will be available in Q1 2013, meanwhile we are offering a limited number (which still can be a big number) of developers the opportunity to test, debug, develop, contribute, etc, so the end product becomes much much better.
This is just an opinion, and I hope the best for RPF and their mission, as a hardware/firmware developer and computer science and networking nerd, I’m still excited to have at some time the opportunity to contribute with more than 30 years of experience. I’ve two young kids and I’d love to let them learn to program and how computers work using a Raspi, and have them feel the same thrill as I felt when I put my hands on the first microprocessor kit, or the Sinclair, or TI-99, or Commodore-64 or Atari, or Apple I/II, you name it.
Raspi has the magic to bring back those days and let dinosaurs of electronics and computer science like me to transmit those feelings and knowledge to new generations.
So cut it out with the defensive mode and switch to constructive mode.

I can’t speak for others, but I know that when I post, I’m just answering questions. People want to know why they didn’t plan for this, the answer is because they intended them to be development boards. That’s not a defensive answer, that’s a fact. I agree with the charity response. The RPF has never gone “but we’re a charity” and I dislike it when people try to cheapen their work by saying “oh but they’re doing amazing for a charity” They’ve done good work for any type of group that’s this small and have never tried to twist things around just because they’re a charity (which technically they’re a non-profit, but that’s a matter of semantics)

Hi Abishur, my apologies for bringing up the “charity” point. For what it is worth, I have never imagined that the term was being used to excuse anything, simply setting a context that some readers may not understand the foundation also have day jobs. Having set up a business myself, even working full time, to get anything done you have to go with your best judgement at a thousand decision points and hindsight is a cruel critic. Mea culpa.

I wouldn’t worry about it, I don’t remember anything you said being offense. What I was specifically referencing was the fact that some people kinda do a sly attack at the RPF by using the word charity patronizingly. Like “those poor people they can’t help it, they’re just a charity after all” as if the people there were less intelligent or the problems encountered would have been avoided by a *real* company (a lot of people tend to forget that the founding members of the charity are business veterans who have run a nicely sized business before ;-))

TBH, many problems encountered would also be encountered at ‘real’ companies. Product would still be ‘late’, but we wouldn’t know about it because we wouldn’t have been privy to all the going’s on prior to release. And the product would be twice the price.

Jamodio, please refrain from insulting the moderators on this forum (refer previous post, not necessarily this one). Keep it polite and in turn I will.

Looks like this post makes sense and is not offensive: “No harm if with a wide criteria you qualify who is and who is not a developer, and yes include some end users as test subjects, and you say, listen the final product will be available in Q1 2013, meanwhile we are offering a limited number (which still can be a big number) of developers the opportunity to test, debug, develop, contribute, etc, so the end product becomes much much better.”

@Raffy. I did say in my post that it wasn’t necessarily this post that had something ‘offence in’ It was indeed another post. I have simply asked the poster to refrain from attacks on people in general. Which have happened on other posts.

You guys are hypersensitive on any type of comments when you don’t agree or like what others are saying, and very often overreact taking the comments as attacks or insults or whatever, and by doing so assuming that we know nothing or read nothing or answer questions that have not been asked with assumptions or second hand information in an authoritative manner which fuels the discussion. There is an old golden rule on the Internet to minimize flame wars, read twice, answer once.

So stop answering then? We ARE protective – that’s the (unpaid) job we have taken on. We believe in this and don’t like to see insults, or baseless or repetitive attacks. Constructive comments are frequent here, as you can see (it doesn’t get deleted, unlike, for example, Apple’s website). But I see no problems with attempting to defend the position of the foundation by refuting those comments. If that makes us appear hypersensitive, tough – that’s just how it is. If I don’t agree with something, I WILL post about it to attempt to refute it, as will other mods, admins or members of the public. If you have a problem people defending against your posts, well, that would make you the one being oversensitive?

Thanks for the update.

Hope to see some of you at Beeb@30 :-)

If the first batch of the board is not FCC compliant and some lawyers like to complain – too bad, then the development board should be simply not delivered to the US.

Unfortunately I get the impression in some posts, that you (RPF) sometimes don’t get hints how to improve rather than claims. Personally I like your strategy of getting some development boards out that software development can start before the final release of the PCB – you are volunteers and engineers, not politicians or a profit-oriented company. It’s not clear for my why some “electronics experts” want force you to release a perfect product immediately, without understanding your already posted arguments (non-profit organisation, overwhelming demand, …).

Guys, I think you are doing a great job without becoming rich that way. Thanks a lot for your passion, please continue!

Amen

Entirely to opposite of getting rich in fact – just a few people putting their houses on the line to produce what was originally expected to be a small run of development boards, leading to larger runs in the future
Circumstances have changed dramatically. People need to accept that, and whatever needs to be done to cope with it. It may not be pretty, but there are UNPAID people trying very hard to get things done as soon as possible.

All credit to the Foundation but RS are not coming clean with prospective buyers of this development board.

Correct me if I am wrong but RS & E14 are just the resellers and not the manufacturers of this item. What exactly aren’t they coming clean on?

Unlike the Foundation, I believe that RS are not giving us accurate information about reasons for the delays.

Good luck on the EMC stuff, hated being involved in that stuff :(.

EMC compliance testing makes me happy, because I expect to use my RPi in the same room with my amateur radio station. Extra EMI I don’t need.

How do do you EMC test a bare board with no case?

Same way as you test a board WITH a case.

I’m sorry. That would mean the board would fail the test if you used the same EMC standards.

Also, EMC performance would be different when the end user/developer fitted a case that was not part of the original EMC testing.

If the end product is witohut a case they sould do the testing without the case I believe.

The bare board is intended to pass, with or without a case.

Own design case? That would be the responsibility of the user.

I really must take issue with all those who appear to believe that you can or should properly EMC test a board without any case or enclosure.

Any EMC testing under those conditions would be invalid unless the board is specified for use only in a laboratory or inside a specified, industry-standard rack or enclosure.

This board is a development board very similar to others on the market which are sold in large numbers worldwide.

If there is NO EMC shielding included in the case, and the board passes EMC validation without a case, what difference is the case going to make?

If the board DOESN’T pass EMC validation without a case, then you may have a point, and a (shielded) case would need to be included in the testing. Let’s wait for the results, shall we?

h people would stop comparing the RP with development boards like the Beagleboard. It is very different in many ways. With the RP all you need to have a working system is a power supply, keyboard, mouse and monitor just like a normal p.c. The only real difference is that the operating system is on an SD card and not a hard drive.No specialist knowledge is required. It is a single board computer with ethernet connectivity and will be used in domestic environments where EMC can be a real issue.

I still don’t see how you can properly EMC test a bare board without a case. I am also a radio amateur and I used to be a working electronics engineer.

Unless a case provides electromagnetic shielding, it makes no significant difference to the board’s EMC.

How many of the Rpi cases you’ve seen so far provide EM shielding? I bet the vast majority on offer will be simple ABS or acrylic, and hence EMC testing might as well be done without any case present.

In fact it *should* be done without one, to catch the worst case scenario (in both meanings of the term).

I’m sorry. That is not correct.

I really must take issue with all those who appear to believe that you can or should properly EMC test a board without any case or enclosure.

Any EMC testing under those conditions would be invalid unless the board is specified for use only in a laboratory or inside a specified, industry-standard rack or enclosure.

This board is a development board very similar to others on the market which are sold in large numbers worldwide.

My Arduino didn’t come with a case. Do you suppose it was tested with one? Would the certification then be invalid since it was sold to me in a condition different than as tested?

First 10k were destined as development boards for the educational release which would have certification. Sheer quantities sold means that is no longer appropriate, according to RS.

Firs 10k were intended to be Dev boards for the educational release, which would be fully certified. The sheer quantity sold means this is no longer appropriate, according to RS.

That doesn’t answer the question. How can you test them without a case?

Regarding the dodgy network sockets, no mention at all in the RS email: Much has been happening behind the scenes since you received our last update, not least the start of compliance testing on the Raspberry Pi. Why are we doing this? Compliance testing is an essential part of bringing any electronic product to market. The Raspberry Pi is being tested to make sure it conforms to all the regulations that apply to electronic devices around the world. This means that we can be sure the Raspberry Pi we deliver to you meets the correct standards and is as safe as you would expect any electronic device you purchase to be. We’re working with the Raspberry Pi Foundation to manage the testing process as quickly as possible, while ensuring all tests are carried out to guarantee safety.

If the case is plastic, there is NO difference between cased and uncased, for radiated and conducted emissions. Radio waves pretty much ignore plastic. If you need shielding, you would add a metal shield (although vacuum-deposited metallized plastic for EMI shielding does exist).

That statement is not correct. An unscreened plastic case may or may not attenuate radiated signals. Care must be taken with screened cases to minimise emissions where sockets and connections pass through the case.

@Bill
The board in a non-conductive, non-shielded case (eg ABS and most other plastics) will either have the same or lower emissions on all frequencies than the bare board.

Thus if the bare board passes, there will be no issue for all non-conductive casing. Some may attenuate, most won’t. Gain is impossible, and gain is the only thing that matters.

If the bare board does not pass, then you can’t rely on the casing providing the necessary attenuation.

However, the same can’t be said for any kind of conductive casing, even one intended to provide shielding as it may inadvertently become an antenna with some gain at one or other frequency.

In other words – testing the bare board is the right thing to do.

– That said, USB is nasty. Good luck with that port, you will need it.

I can connect a power supply, keyboard, mouse and monitor in my beagleboard. I can write OS image into SD and stick that SD into board so where is the diffference?

Now, take this as a constructive critique, I even -tried- to be short…
Please, stop saying “charity” “non profit” “little margin” and such as a liet motif for everything ppl point at: I find it hilarious, is someone here able to do some math? Even with a profit of 2usd/board (and it’s just to put a number), with 100.000 unit actually sold/pre-ordered there’s a nice 200.00usd, how’s that a little? I think that if you are releasing something, you must do things to make it compliant with normative, it’s a given, and you do it before production, cause 10.000 units is production (at 35usd/board it’s 350.000usd), and the rick of a -non compliance- are too high. You what? Forgot that? Postponed?
Anyway, there is another aspect of this: always talking “charity” “non profit” “little margin” undermines your credibility as a foundation, and this is aggravated by Farnell and RS, who are basically giving the impression that the RPi was launched before being ready.
In the end, I do hope the RPi get tested, certified, approved, seal of God, (and sent to customers, possibly not too overcharged) but if this doesn’t happen… well, I don’t want to think of it.

Piopio, *please* read through the comments as this has been answered, a lot, in the comments on this post. the compliance issue was originally non-existent. They were releasing the boards as development boards in full compliance with the law. It was the distributors who have pushed the issue of getting them certified because of the extreme demand for the boards. Now I can’t blame the distributors for doing this since they could be held liable if anyone started making a stink, but neither is it logical to go off on the RPF because the plan was forceably changed by the distributors mid-stream here. As for the profit comment, the reason it is ultimately a little margin is because they put up their own homes as a liability so most of that profit is eaten up paying their startup fees back.

Reply more down, blame antispam :P

I’d just like to take this opportunity to point out that there is little margin for the Foundation, who are non-profit charity.

Oh, on profit not side: I’m not against the foundation making profit out of this, little or big, even if it won’t go into charity, I’m just against fanboyism. You seem just one.

Cheers! It’s a leitmotif of mine. :)

piopio said: “I’m not against the foundation making profit out of this, little or big, even if it won’t go into charity”
That’s a very serious accusation. The Foundation are a registered charity. If you suspect them of fraud then I suggest that you contact the police and the Charity Commission. If you have no basis for this comment then perhaps you should apologise and shut up.

I accused nobody, of nothing, anywhere. Read the whole phrase. You jumped at the gun so fast, for nothing. Seriously, ppl here put meanings in what one writes that aren’t there.

There will be profit on each board – not much, but it is there. However, to get profit, first you must have sold the boards. Which so far, has not been done so the Foundation has no income from anywhere at the moment, its all being done from donations, and the Founders own money. As to the certification, covered extensively elsewhere.

You are making a number of assumptions here. For one thing – the members of the RPF have sunk their own money into producing 10,000 units – not 100,000 – so even at $2 ROI per unit, that will only bring in $20,000. Personally, I would expect their margin is probably a matter of cents rather than dollars per unit, but assuming $1 – when the first batch are all sold, they will have $10,000 extra, which will pay for another 294…

Apart from that, you say:
“… always talking “charity” “non profit” “little margin” undermines your credibility as a foundation, …” Please bear in mind that it is NOT the RPF that say this, but posters on this group (okay, so that includes admins, such as myself, but they are not, in the main, RPF members). Besides, the Foundation IS a charity, registered as such in the UK, and it HAS to abide by statutory regulations as to what it can and can’t do, and I’m not sure how these comments undermine RPF credibility. Remember, this is an EDUCATIONAL charity – aiming to improve certain aspects of EDUCATION, …

it’s right, and I was just referring to posters, not the Foundation itself. And it was really intended as a constructive critique; if I had to say what is passing in my mind, they’ll just nuke me.

EMC testing – so fun, many an hour spent in a shed just north of Cambridge…

Whilst I’d rather have the board I’m not going to complain.

I just hope the post above about thee distributors holding batch 1 until rolling production is active are malicious rumour, bearing no relation to reality.

Abi, I did read all comments, and I didn’t ask any question, if not as a form of speaking (as in rethoric).
I don’t really see how 10.000 boards, sold to anyone who can do an online order is a “development” scenario, in most EU it’s not, and compliance must be done before selling, before manufacturing, on so called pre-prod samples. Because in most EU certifications -are part of the product itself-

Even if it’s rhetoric when someone goes “You what? Forgot? Postponed” I’m going to answer the question as the asking of it reflects a lack of information. Liz is a lawyer who did her homework, 10K boards can be sold as development boards. But as I also said, I understand that due to the fact the release of the 10K were delayed, and the unprecendented demand for the pi, that the distributors decided to push up the time table for compliance.

>I don’t really see how 10.000 boards, sold to anyone who can do
>an online order is a “development” scenario, i

It might look weird to people from outside of electronic business, but this is how it works.
TI does it all the time, Look up $4.30 TI MSP430 Launchpad, they probably sold over 100K of those.
All you do is put this in documentation :
“FCC WARNING
This evaluation board/kit is intended for use for ENGINEERING DEVELOPMENT, DEMONSTRATION, OR
EVALUATION PURPOSES ONLY and is not considered by TI to be a finished en‐product fit for general
consumer use. It generates, uses, and can radiate radio frequency energy and has not been tested for compli-
ance with the limits of computing devices pursuant to part 15 of FCC rules, which are designed to provide
reasonable protection against radio frequency interference. Operation of this equipment in other environ-
ments may cause interference with radio communications, in which case the user at his own expense will be
required to take whatever measures may be required to correct this interference.”

Where’s the off button?

There isn’t one. Just unplug it, or switch off at the wall if you have switched sockets.

I hope he means “where’s the off button for this thread.”
Personally I’m looking for it….
:D

:D

Price increase?

No.

It seems to me that RS and Farnell are being very correct and professional in pushing for full certification. After all, the board is going into homes everywhere, not just into the hands of electronics engineers and radio amateurs.

The Pi was presented repeatedly by the Foundation as a general-purpose micro PC with good media performance, and this ensured that interest would be universal. For all intents and purposes then, the broad audience was planned, and the volume of interest a foregone conclusion given its tiny price.

Given that Eben & Co. are leading-edge engineers and the same board was expected to go into cheap (hence unshielded) cases for education, I fully expect the compliance testing to go well. If it doesn’t then Eben will be kicking himself severely, no assistance needed.

Fingers crossed …

Right, keep up the hope Eben with another delay. (humor intended :) )

Yes, just like they did in case of TIs Launchpad, they waite… oh wait no, they just SELL IT with no certs because it is a bare PCB.
http://uk.farnell.com/texas-instruments/msp-exp430g2/kit-dev-msp430-launchpad/dp/1853793

Two comments in this thread from Abishur cause me concern

“I believe it’s been told that one if not both distributors have decided to not release the first 10K until they can release a huge batch (a fact which I think is lame and defeating the whole point of getting the first batch… but what are you gonna do?)”

“the fact that the first 10K will be released with the main bulk release”

Did I miss an announcement about this, or a forum topic, or an email from my supplier? News to me and I visit the board and my suppliers website several times a day

Arg, I saw it on one of the distributor’s site but now I can’t find the original source to my info. I also try to make it well known that I take things from them with a grain of salt, but sometimes I’m in a hurry and forget to word it in the right way to express my uncertainty of that piece of info (sorry)

That will be rational on the part of suppliers, especially with the constant checking on margin by end-users, ie, minimize costs by shipping infrequently/in large batches.

OK, it’s a little too late for those who have already gotten their panties into a twist, but, here’s the bottom line about compliance testing – it is _not_ legally required. As I mentioned in my earlier post, if you decide to operate a device which has not been certified to meet the FCC/CE compliance standards and (here’s the legalese weasel-wording) the device causes interference with other devices, you, as the operator, are solely responsible for eliminating the interference, up to and including ceasing operation, if necessary (almost never the case – shielding and line RF attenuation devices are typically sufficient). The reason that businesses pursue compliance is that it eliminates any concern on the part of prospective customers that they might not be able to use their purchased device as desired without any hassles from disgruntled neighbors (just once, I’d like to live next to some gruntled neighbors ;) ). It’s simply good business to pursue compliance.

As I also stated earlier, the ~1 watt total power consumed by the R-Pi very likely results in much less than that much in effective radiated power (ERP – oh, excuse me, must have been the delicious spicy noodles I had for dinner :) ) in the RF spectra. So, given that the boards appear to have the kinds of ground planes, guard traces, signal trace routing, and component placement that generally result in compliance (I’m betting the layout software automagically provided a pretty good starting point), there probably isn’t anything to worry about. Any issues can easily be ameliorated with appropriate shielding and/or cable attenuation.

As for a maximum number of “development” boards, there is no legal limit, it’s simply a matter of how many potential product complaints/returns does a producer want to risk if residential customers wind up with those pesky disgruntled types next door?

Just take a chill pill and don’t call anyone in the morning, OK? That also goes for the admins – you guys are just providing waaaay too much bait for the trolls. To quote the Bard, “Methinks the lady doth protest too much.” Moderation in all things is best … especially in moderation, ironically! :)

That is completely false:

1) CE compliance testing IS legally required, even for development kits.
2) Testing has to be done by the manufacturer, importer or entity first placing the product on the EU market – not the end user.

Gary Nevision, Head of Legislation & Compliance at Premier Farnell, has covered this very well in his Element 14 legislation-related forums.

I’m surprised – even incredulous – this wasn’t discussed before the RS/Farnell distribution agreement.

Perfect! The first boards can all go for orders outside the EU. More for us in the US, Australia, etc.

What about the FCC? :)

I’lll second that.

John – Your comments don’t apply to the FCC, and it clearly states below that the _operator_ is responsible for resolving interference, but, it’s strongly recommended that manufacturers perform compliance testing:

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, you’re also wrong. 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.

You should read the actual standards/regulations, rather than relying on blogs and similar blathering by self-appointed “experts” who would actually be producing useful products if they knew how, but, instead, they can only spread misinformation.

@Jim
You don’t seem to understand how law works. You can’t just pick a random quote of one regulation to ground your statement on. You placing your opinion – based on this selective reading – over a legal expert’s opinion (who works at one of the RPi distributors) is laughable.

Try reading ALL the regulations involved in CE compliance for a start:

EN 55014-1/1997, EN 61000-3-2/1995, EN 61000-3-3/1995, EN 55014-2/1997, EN 60335-1/1988 and EN 60335-2-56/1991

Btw you’re also wrong on the RPi’s power consumption, it’s actually 3.5W. Maybe you’ve done some selective power rating too?

OK, since we are starting to have a go at each other I suggest that people (who are not lawyers working in this area) stop commenting on this subject, and leave it to the lawyers, parasitic scumbags that they are (YMMV). No-one posting here is qualified enough to make definitive judgements. And when I say suggest, I mean stop.

not true, you dont need to test anything uif you arent selling end product
Farnell knows this , they are selling devkits that have ZERO certificates
http://uk.farnell.com/texas-instruments/msp-exp430g2/kit-dev-msp430-launchpad/dp/1853793

http://www.ti.com/lit/ml/slac432a/slac432a.pdf

“Texas Instruments (TI) provides the enclosed product(s) under the following conditions:
This evaluation board/kit is intended for use for ENGINEERING DEVELOPMENT, DEMONSTRATION,
EVALUATION PURPOSES ONLY and is not considered by TI to be a finished end‐product fit for gene
consumer use. Persons handling the product(s) must have electronics training and observe good engineer
practice standards. As such, the goods being provided are not intended to be complete in terms of requi
design‐, marketing‐, and/or manufacturing‐related protective considerations, including product safety a
environmental measures typically found in end products that incorporate such semiconductor components
circuit boards. This evaluation board/kit does not fall within the scope of the European Union directi
regarding electromagnetic compatibility, restricted substances (RoHS), recycling (WEEE), FCC, CE or
and therefore may not meet the technical requirements of these directives or other related directives.”

Please dont spread disinformation

.” Moderation in all things is best … especially in moderation, ironically! :)”

Nothing ironic about that, and as an admin, I agree, but we do have to trawl through a lot of dross and many-times repeated comments making exactly the same point.

Maybe if all contributers were able to pre-post moderate their own words, life would be moderately better for all of us :)

Quick question since I’m not 100% familiar with how it is handle in the EU, what is the situation in relation to the hazardous material contained in boards like this ?
The picture that has been sent by RS via twitter shows a technician using a fluoroscope X-ray station for material analysis.

For electronics RoHS is actually mandatory. If that was the question. Or you mean you want to know how they do the thing?

The board is RoHS compliant already. Its the EMC stuff that is ongoing.

That’s nice to know. -1 test to pass.

Thanks piopio, yes that was the question, I didn’t know if RoHS was mandatory or not, so it makes perfect sense that RS seems to have done material analysis.

The RoHS stuff was done by the Foundation, not RS.

I didn’t say the Foundation didn’t, I said “seems” RS did some material analysis derived from a tweet sent by RS on March 12.
https://twitter.com/#!/RSElectronics/status/179552684578312192

Ah, sorry, I though you were implying that RS had to do it because the Foundation hadn’t.

This is really sad news! I hoped that now thousands of RPs are going to be manufactured and that they soon will be sold. Let’s pray that the PCB layouters did a good job and the RPi will pass the tests. If not? What will happen to hundreds of boards which are already manufactured? Or is the whole manufacturing process stopped? Well, the RPi Foundation did a great job so far! We should all keep in mind that you are all volunteers and earn no money doing this and it would be very hard for you to pack and ship 10000 RPis. I can understand that you needed a producing partner – but I really hope that you find a way (if the design did not pass the tests) of releasing the boards already produced to the public. Unfortunately I gone with RS and registered early in 29th of February….

So… When is the Re-Launch Date? My Pi is ordered for the end of July. Am I going to see it in August?

The RPF is trying to get this nailed down. Liz said that they would talking to the distributors sometime early this week. Now obviously if they’re legally prevented from telling us anything in that meeting we won’t get any new info, but I think we should be getting some form of update at some point this week.

I know it’s really aggravating having to wait the way we are (oh how I know ;-)) but if I force myself back and look at the data then the best case scenario then let’s say that when we got the information about the ethernet jack issue that the entire first 10K were almost done (something I think to be unlikely based on the fact that they said first set rather than first batch and the fact that it wouldn’t make sense to do all 10K before the factory confirmed that swapping the boards solved the issue), then let’s say they got done with retesting the electronics two days later (Friday March the 9th), then the factory would come back on Monday and ship them up. So let’s say they shipped Tuesday March the 13th. That would mean in this *example* we’re looking at 9 business days in the best case scenario since they would have shipped. Goodness knows I’m waiting for some LEDs I ordered from china and it’s been more than 9 business days! :-(

“Now obviously if they’re legally prevented from telling us anything in that meeting” I think “Contractualy prevented” is what you meant!

lol, is that the correct legalize? I fear my knowledge of legal matters is only broad enough to grasp the concepts of legal matters, not the correct terms to go along with those concepts ;-)

This time next year we will look back on all of this and laugh 8-p

“We did a first trial run in a test chamber this week” .Not tested by gordon freeman i hope ;-) .Keep up the great work everyone .

CE parking or Chinese export marking isn’t worth the paper its written on! The number of very dangerous devices that come out of china with the CE mark on is HUGE. Its a self regulated standard and is essentially useless! Shame the RPi is possibly going to be delayed due to some paper pushing…

CE compliance testing is a pain, but an essential pain. I don’t think anyone wants dangerous good being sold on the open marking. You do raise an interesting point about the Chinese Export CE mark that most people are not aware of.

http://en.wikipedia.org/wiki/CE_mark#China_Export

Sorry, “market”, not “marking”. Should not submit entries before my cornflakes.

no way will a bare PCB like this pass EMC testing! While im sure it will not emmit much EMC, im sure it will be affected by external sources!

heres to another 6 month delay, when RS realise they cant sell a non-CE marked product as a finished product.

WRONG!

Most electronic devices these days do not need a “faraday cage” around them to be compliant.
I have built numerous devices that were encased in a plastic enclosure that passed EMC tests with flying colors.
Obviously on the board level you have to know what you are doing, but considering that the Raspberry PI uses (primarily) a SoC that is known to be able to pass EMC tests, I have all the confidence that is at least possible to build a device with it that doesn’t depend on complete shielding to pass EMC tests. In fact IMHO the smaller the board the easier it often is to emit low amounts of electromagnetic radiation.
Remember that almost all the fast switching signals are happening inside the SoC, not (normally) exiting it.
On the other hand, I do not see any ferrite beads around the SoC for the decoupling of the power supply, and near the I/O connectors (to prevent them to become antennas) but perhaps these are on the bottom side of the board. Actually I’m more worried about ESD immunity than EMC. I see no protection against ESD impulses entering through the power supply.
On the whole, I’m not too worried.

Tell that to all those Fluke 87Vs that died when gsm phone rang near them :)

no protection against impulses on the RPi i guess, but easy to fix on the power source I suppose.

For anyone that is interested, here is a link to the EN 55022:2006 document that gives an example of the level and complexity of compliance testing. From the description in the document, I believe it would apply, for example, to a Raspberry Pi sold to a domestic user. http://www.kh-emc.com/biaozhun/EN55022-2006.pdf

If the first batch doesn’t quite meet FCC or CE certification requirements, perhaps they could be donated to non-profits in developing nations? Low power computers could be very useful in many places, and whatever small amount of interference that a RPi could generate would be of little consequence in a location where there is little or no other technology for it to interfere with…

We have an in-house facility for EMC radiated and conducted emissions (Test Cell, LISN, etc.). Would love to get my hands on a PI to see how far away from compliance it is….

;oj

I hope we’ll be able to share some graphs with everybody soon. We *love* graphs.

I’m amazed at the number of people posting absolute, definitive views that are diametrically opposed. I suggest anyone who is so certain of their position as to be unshakable study this: http://en.wikipedia.org/wiki/Dunning-Kruger_effect

The fact that that the authors were awarded an Ig Nobel prize for a similar study says a lot, but I’m sure you already had doubts about it and didn’t take it as a definitive view… right?

Silly question but why is the Raspberry PI diffrent from say – buying a graphics card? – not everything in computing comes in casing?
And we can all go build a computer without a case surrounding it. I always looked at the Raspberry PI as a component not the finaly product, thats for us to come up with and as they said – by the time they go into schools they would have cases.
Just wondering where the diff comes in and why people are worried it wont pass?

I believe things like graphics cards are also CE marked as they are not Dev boards but consumer items. They will also be put inside a case for usage.
I use my Raspi as a final product…and the volumes to be sold make it less likely they could be regarded as dev boards.
They should pass btw.

Thanks for the reply – the way some people were talking it was as if it needed a case in order to pass the test.
I just couldnt get my head around that one. Allready read the update where you guys have allready done some of the testing and are fairly happy, congrats and roll on the RS ordering website :)
Also congrats on doing all this out in the open and sharing the experience with us all – not every company would do this.

Dogs

I actually work for a compliance test lab in the US. And I agree that the quantity we are talking about here and that they are marketed to the public for general sale makes it so they should be tested.

CE marking in Europe is a self certification process. You can stick a CE mark on anything you like without even doing any testing.

But if the sh*t does hits the fan and your product starts behaving badly, the persons with their signature on the Declaration of Conformity are going to cop it. The persons responsible for signing the Declaration of Conformity are the product importers, not the manufacturers. So in this case RS and Farnell.

Amen to that!. Just slap on a CE sticker, start selling them and testing at the same time (if you are confident enough that it will pass tests). CE is just statement that product is compliant with “all applying norms”. The worst part of certification is knowing which ones do.

After years of assisting clients with reliability, obsolescence, legislation issues it never fails to amaze when NPI’s hit snags like this. I remain hopeful the knowledge/experience is not confined to just another blog and someone asks me ‘how can we do this better?’ – I should really do more marketing.

Fraser
http://www.ilscomplete.co.uk

News

Latest posts

Supercapacitor free flight planes | HackSpace #65

Raspberry Pi Pico Game Boy Interceptor | The MagPi #128

Building a pick and place machine with RP2040 and Starfish

How to build your own Raspberry Pi webcam

Next Post

Pricing matrix update from element14/Premier Farnell

Previous Post

Programming the Raspberry Pi webinar with Eben, April 4

Share this post

196 comments

Inky-

p4trykx

texy

Charlie

texy

Stephen Hill

Abishur

Tom

rasz

Abishur

JamesH

Abishur

Matthew Slack

Funar

Lee

tzj

Ian Shepherd

birger

John Sousa

John Sousa

birger

jbeale

Greg_E

Tom

JamesH

Nich00001

steve

Abishur

Abishur

Abishur

Abishur

Abishur

Abishur

JamesH

shirro

tgoldblatt

PassingThrough

Stuart Lea

tgoldblatt

tgoldblatt

Ian Ramsay

Abishur

Stuart Lea

Abishur

JamesH

Stuart Lea

JamesH

Greg_E

tgoldblatt

Tom

Tomo2k

Stuart Lea

PassingThrough

Stuart Lea

Magnetron

Stuart Lea

Abishur

tgoldblatt

Topher Brink

Bill Harris

JamesH

Bill Harris

Winston

Stuart Lea

Abishur

PassingThrough

Abishur

JamesH