It’s inevitable, isn’t it – you’re freewheeling along perfectly happily and then you get a puncture.
As you’ll have noticed, there’s been a bit of a delay in shipping the first batch of Raspberry Pis out to people. This is because of a hardware parts substitution that was made in the factory by accident: specifically, where we’d specified jacks with integrated magnetics in the BOM and schematics, the factory soldered in non-magnetic jacks. No magnetics means no network connection. We’ve known about this for four days now, but we haven’t been able to tell you about it because it meant we had to do some further tests to make sure that nothing else was affected.
Happily, it’s a very minor problem to fix (desolder the dud jack/solder on a new one), and the factory is nearly done working on replacing them on the first set of boards. This means that the first tranche of boards should still go out to customers as we were expecting. There may now be a slight delay in later batches if there’s a problem sourcing enough magnetic jacks (we’ve got teams hunting them down already); all the stock of jacks we believed we had in place and ready to turn into the ethernet ports on your Raspberry Pis turn out not to be the correct part, so we’re having to start again and move through the negotiating/ordering/delivery cycle as fast as we can. Our partners at Element 14/Premier Farnell and RS Components are working hard to help us cater for this, and to expedite supply of the Raspberry Pi.
We are very, very sorry. We know you want your Raspberry Pi as soon as possible (and many of you are being inhumanly patient, having followed us since we launched this website eight months ago). We’ll keep you updated with how manufacture is moving; this is, in the scheme of things, a minor problem, but it’s still a bump in the road and we know that we rely on your goodwill to keep things moving forward.
No worries…sh*t happens :) and this is a minor poo
Yep, minor glitch, you are still gunna do great things for the world of computing
With production ramping up (depending how much influence the Foundation still has on future manufacturing) I hope there is a process for qualifying suppliers. An obvious rule: no substitutions without consultation. If the next batch is 100k+, being struck from the list of qualified suppliers may hurt (if not quite on the scale of being blacklisted by a huge juggernaught like Apple).
Ouch, yes bad.
But the front page article with all the pretty pictures almost make up for it. Very interesting and educational.
The RPi platform shines an educational light on intro computing and programming.
These process explanations to do the same thing for many regarding the electronics mass production industry.
These things happen. I just hope your negotiating isn’t compromised by the position you find yourself in. Good luck
Ha! I know the answer which one is which. The secret is [Snipped by Liz – no fair! You’ve seen them!].
Hey! That’s a bit unfair Gert – you have inside information – you can’t enter! :)
(cong)rats! you beat me by a minute!
No I did not see them. I looked carefully at the two photos and at the X-rays comparing them. Maybe it is because I did a lot of ‘find the ten differences’ pictures in a Dutch newspaper over Christmas.
In that case, you were the first with the correct answer – you win! It’s the one on the left. (I’m not sure what you win. Head over and ask Eben.) ;)
10,000 second-hand non-magnetic RJ45 jacks?
I’m thinking art installation.
Erm, the one of the left is which? magnetic or not? XD
That’s how I figured it out, too. The “ears” inside the magnetic one don’t stick out as far, and you can see that in the x-ray.
That is really weird, I figured the gold one is the one with the magnetics inside due to the curve it creates when bending the wires at the junction point. (Gold ones are kind of fold, silver ones look curved).
What do we learn forom this? Yeah, there’s magnets in a RJ45 jacks. Who knew?
The interesting part to me is this: If no magnets means no network connection, what are non-magnetic-jacks useful for at all?
It doesn’t mean no network connection at all on all devices, but this board has been designed for a magnetic jack. The magnetic bits mean better signal integrity, better filtering and shorter transmission distances for data.
If that were the only problems, you could sell them as Model As with a bonus USB port. The “killer” problem, is that these boards are missing the high voltage isolation that make it legal for you to sell them in many jurisdictions. A conforming ethernet interface will block high voltages.
I suspect these actually worked, at least partially, with short cables when plugged into a laptop. That would explain why no one noticed right away. If my factory were making them, I’d take one to my office to experiment with it.
The regulators require the high voltage isolation, because ethernet cables can be very long. They take long convoluted paths, sometimes even between buildings. Bad things can, and do, happen to long cables. Sometimes they come in contact with power. It is important, that they don’t kill people when this happens.
Which is I believe how you can tell which kind you’ve got without using an X-Ray machine or tearing the jack apart. If you take a resistance meter to them, the kind with magnetics should in theory have a nice high resistance between the pins of the RJ45 connector and the PCB pins, whereas in the kind without the two sets of pins are directly connected.
Perhaps you should use RJ plugs with indicator lights, that way there will be a visible difference between the RJ plugs with magnetics and those without. I guess that this mistake can’t be cheap.
I’m glad you answered this question for me. I’ve been wondering what integrated magnetics do for ages but there isn’t a single explanatory page about it on the net. I was starting to think it was a conspiracy!
For a long time all Ethernet interfaces had a transformer externally on the board. For smaller designs the transformer was integrated in the plug, but they where more expensive. But the with-transformer plugs are now so popular that the price difference has eroded so far that it make no sense to use the without-transformer parts in new designs.
Ah! That explains it to electronics-noobs like me :)
Thanks Liz and Gert!
“Magnetics” in ethernet jargon reffers to the pulse transformers that sit between the ethernet chip and the ethernet socket. These transformers provide isolation to protect from voltage differences between ethernet devices and also block any DC bias currents from making their way onto the ethernet line.
In most PC network adaptors and switches i’ve seen these transformers are a seperate components but in space constrained embedded devices it is very common for them to be integrated in the network jack.
The ‘magnetics’ are in fact tiny pulse transformers, which are present in all ethernet interfaces to DC-isolate the computer from the hub/switch. Older interfaces have a separate pulse transformer component (looks like a DIL IC), but the Raspberry Pi uses an RJ-45 jack with the pulse transformers built-in.
Technically, the current Raspi’s, don’t have the pulse transformers built in (although the Raspi design does…)!! And therein lies the rub….
Did the factory get confused over what part it was using or were they in turn supplied different to what they ordered? Is it worth checking all the other components on the board are as BOM’d where practicable? Perhaps another error may take longer to show?
There are very few parts on the board, and they have now all be thoroughly checked – one of the reasons for the delay on the announcement was needing to give the board a thorough going over.
Confused? Simple mistake ? Pulling a fast one? Not known.http://www.raspberrypi.org/wp-admin/edit-comments.php#comments-form
hmmm check that they’re fitting authentic components too, including all ic’s – there are counterfeits of everything these days…..
there are applications that require electrical connection to the wires, and therefore use a jack without integrated magnetics combined with an external pulse transformer. see e.g. Power over Ethernet.
Actually, there’s a point. Could future Raspi’s run off PoE?
I was wondering if PoE would be possible on the next revision. I was thinking remote controllers in factory environments.
“Magnetics” is electronics jargon for “Transformer”. There’s an isolation transformer in the jack, to protect your pi from nasty voltages that an ethernet cable might pick up.
To repeat, there are no magnets in the jack.
As Mark B said, schmidt happens.
I bet your heard dropped when you found out, but as you said, its a minor issue. I’m sure people will be fine and probably wont even notice a delay from the sounds of it :)
(Sidenote: Please, please ,please can you get rid of the broken password thing on comments – took ages to post this :( )
Workaround I stumbled upon a while back: If you have a forum account and login to that, you don’t need to type the anti-spam password to post a comment.
Well, we do have to make it clear that this could very well cause a delay – we still aren’t clear on whether it’ll be possible to get the parts we need for the huge numbers of orders we have in time now, because all the sourcing the factory had done for us was on the wrong part. Like you, I hope like hell it *doesn’t*, but we have to be realistic.
And yes. You should have seen me, Pete and Eben over dinner last night. It was as if we were at a wake.
Don’t take it too badly – I’m sure the vast majority of people waiting on orders will be understanding. You’re dealing with techies here, we’re used to things going wrong :)
It made the BBC News, the report was quite sympathetic I thought.
I think the “gold colored” version is the magnetic version…
The little kinks in the side spring tabs on the X-ray image of the jack with magnetics match the left-hand jack in the top picture.
Thanks for being so upfront guys. Easy mistake to happen and competely out of your hands. Would rather receive a product that works than have to return it and get another!
Like many things in life I’d rather it be right than first. X-Raying components – how cool is that!
Most modern dayo electronics are x-rayed. This is a non-distructive testing looking for things like voids in solder joints.
I know this, as the images are from one of the X-ray machines I helped design :)
That can happen… btw correct one is on the left :) I won’t post any hints for that :)
Now we will see if RS and Farnell can redeem themselves!
They’re both being brilliant – they’re scrambling to try to help us find the parts we need.
i’ve got a lot of old NICs would it help if i unsoldered the connectors and sent them to you? i think i could scrape up about 20 or 30 of them :)
Kinda wished they kept a few hundred and sold them to people willing to DIY them! :)
The old nics are likely to have the wrong ones that are already on the board.
the 2nd one but YMMV or even mine
Doh! stuff happens
Yeah, you better sort it out! I want my Raspberry Pi :D
NO worries, Liz – still a while until mine is due for delivery anyway :-D
Not being a hardware bod, I’m assuming that “magnetics” are ferrite core like thingies that prevent electrical garbage entering through the ethernet connection and that connectors without them are designed for applications with more circuit board space for an alternative solution?
Its a good thing that your QC checks found this error out before the boards got into the wild and caused worse problems! Terrific work!
Another very important feature a transformer adds is that you have no common ground. Long cables common ground: absolute disaster!
makes no difference to me, since registering on RS for pre-orders i have not heard anything more, so i didnt even expect them to come through at all (yes, im a half empty kind of guy)…
What are the magnets for? Do they ensure the pins hold together tightly?
Magnetics should be read as ‘transformer’. Copper windings on one side, copper windings on other side. Ferrite materiel in the middle. Offers electrical isolation, impedance matching and suppresses noise radiating into the wires.
They’re sucking bits through the wire.
I think it’s got to do with shielding. Look above, where Liz tells it’s about signal integrity and stuff.
Cheer up Liz (and Eben, and everyone else), we know you’ll get things sorted very soon !
Heh, i know it, too! :D
It was pretty easy though ôo
You are GOLD! Glad that you’re bound to retur-urn, There’s something I could have learned, You’re in-de-struct-i-bu-ulllll, Always believe in… etc.
I’m pretty sure an apology isn’t good enough….
So true. Funny how it seems always in time, but never in line for dreams….
Thanks for the transparency! :)
Hahaha! You sure can’t get more transparent than an X-Ray that’s for sure! :)
ahah… good one!
Mitigating problems and risks is a real job by itself, and delivering a high quality product is *never* an easy thing… Keep up the good work!
It isn’t magnets or magnetic.
Magnetics refers to the presence of transformers and chokes which are used to isolate the Ethernet wires from the RaspPi’s power supply. and each other and probably to reduce high-frequency noise.
Without them you would effectively tie the RX and TX signals together and probably turn the entire network into an aerial for Radio 2 reception.
Hmmm. Didn’t know the RasPi had a FM receiver.
quick add that as a “feature”. Jazz FM on Cat5e
Jazz FM on Cat5e.
… “where, although it follows the orginal 32 bar A-A-B-A structure, instead of providing a harmonic departure from the A section, the bridge resolves the rising chromatic pattern.”
Speaking of JazzFM, did any of you see the horrible error they made a few weeks back?
There’s a (NSFW but hilarious) link to the audio in the Guardian article.
Jazz mag, anyone?
(Sudden inexplicable urge to switch from Jazz to ShowTunz).
We were promised robots
No big deal, nothing to worry about :)
No use crying over spilt milk :)
Out of interest how was this spotted and how did it get past the initial checks you did when the boards were manufactured a couple of weeks ago?
The boards build a couple of weeks ago where runs as a small series for free by Pete Lomas’ manufacturing plant. Pete is one of the trustees of the foundation.
The initial checks ensure that electrically everything is sound; they don’t check that the network works. (Because if everything is electrically sound and we take it on trust that the jack is the correct one, we know the network will be fine.) Our mistake was not imagining that the part might have been swapped out for something else.
Thanks for keeping us updated, hope you find the parts soon.
Its like the time I printed 1,000 newsletters only to find out I spelled the title wrong. Disheartening at first but it all turned out OK in the end.
There is a saying: “Of any two options, choose the third”. I believe it is the Jumping Jack Flash from the Quantum Pi of the future . . . or is it the past http://youtu.be/GnFci5UQKuU
Easy! Just ship A boards instead, they don’t need magnets at all.
Top post! Never thought of that…..
That would be unfair: you would get an A version with TWO USB’s for whilst you were promised only one. Power consumption is another reason why a lot of people want an A.
2 USBs on an A board.. bonus! ;)
Ah, but we can’t do that because the A doesn’t have a LAN chip, and the USB hub which allows us to put on an extra port is in that chip…
Oooh, a limited edition $35 model A+
I want one ;)
Or post the boards out with the fault and the correct part.. some people wanted an unpopulated board.. to test their soldering skills! ;)
Sounds like a plan to me – easy fix :)
‘unauthorised’ component swapping/substitution is not unknown in many Chinese electronic plants. Why hold up production and miss billings when you can put in an ‘alternative’ component when shortages occur. They are all at it!!
I don’t think we’d necessarily have minded too much if the “alternative” had been something that worked!
the point is not that perhaps that an alternative will ‘work’ but that is it within a required spec set by the designer. Some ‘alternatives’ are not so obvious in their effect on a design.
They’re totally different parts, but I do wonder if someone thought that the superficial similarity meant they were the same thing. I guess we’ll never know…
If it quacks like a duck and if it walks like a duck…….!! ;0)
The removal and replacement – can (and will) that be automated, or is it going to require manual intervention?
I believe it has to be done entirely by hand – I’m only 90% sure on that, though.
The ‘rework’ will be manual and is the nightmare scenario for the manuf. Puts up the costs and reduces the margins if it gets out of hand. Hell mend them – can kill a product if it happens on too many units.
Well, they are the ones who made the mistake…..they need to sort it out.
At least it wasn’t a 100k batch…..
Can be cheaper to scrap the units than rework them in many cases.
In countries with low labor costs, especially China, if the work can be effectively done by a human, it often is. It’s the opposite of the ‘everything done by machine’ attitude that you might expect in manufacturing.
The jacks might have been mounted by hand in the original assembly.
It’s fairly common for boards to have only the small surface mount parts placed by machine, reflowed (heated to melt the solder paste), and then all through-hole and large parts inserted and soldered by hand. This has the advantage of not needing to wait for the large parts to heat up during soldering, reducing thermal damage. If radiant heaters are used, there isn’t a risk that large parts will shadow adjacent components.
yeah this worries me about whether or not they’ve done this with every single other component on the board, loose fitting usb sockets? cheap power jack that breaks after 5 insertions? PANIC!!!
We’ve checked all those things (this is what we were taking the time over) – Pete has an insertion testing device, for example, that he’s been setting to do tens of thousands of insertions and removals of all the ports.
Λ l є x
Good you guys test all things and only find this one.
The X-Ray are cool!! You x-ray all the components? :D
We x-ray the whole board, but we needed to get them to the UK before we could do that. I’ll have to see if I can get some more x-ray pics from Pete (who owns the machine to do it), because they really are fascinating.
Yes please, those X-ray pics are cool !
Some X-rays may make an interesting sidebar in the school workbooks too.
yes, this is especially true for an education project. Seeing parts inside a small plug is really cool. And it is a remainder to not think that if something is going in and the same way out it can not be changed inside.
It was a very interesting post! Thanks a lot – both for the post and bump in the road …
Can not wait to see more X-Rays – and the videos about the manufacturing of the RaspPi and its parts ;-)
“No magnets means no network connection.”
…so I guess these connectors are reserved for the A version of RBpi?! (-:
What would they connect to?
Already have the first replication on the internet:
Raspberry Pi suffer manufacturing hiccup
God, that was quick.
Shows you how much interest there is in Raspberry Pi ! :D
FYI; the “Magnetic” does not refer to actual magnets in a RJ45 jack, there are none. It refers to the magnetic windings/loops, working much like a power transformator between the board and cable. It separates the two circuits somewhat and facilitates beter signal integrity.
Is that “signal separation”.
Impedance matching transformers built into RJ assembly.
Interesting tech spec sheet from hirose jp that shows what people mean here, for those interested; Not necessarily the same part but same “pulse transformer” design.
Not a problem. These things happen. I’ll just have to get a cheap iPad3 instead!
/crawls into corner and sobs
iPad3 – whats one of those? – just keep crawling, GoG! :)
No that would be stupid. This was hard to forsee going wrong. Its like your HP printer outputting your document on cows instead of text. You just dont imagine that detail going wrong.
(disclaimer: the following sentence contains 37% joke but is based on a true story) .. Its like working with graphic designers. They introduce things you dont want all the time, you cant forsee all their mistakes.
If the delay looks like being a long one how about offering a batch of the boards without the connector fitted? It would be a simple job for anyone with basic soldering skills to fit the correct part at a later date. No more difficult than fitting the GPIO pins.
As we are on GPIO headers and soldering :
I have seen a first production batch board. It had the GPIO connector in place (Male at the top). Eben told me that he does not expect that to change for the first X-(thousand??) boards as it is in the BOM. (Bill Of Materials) Any change to a BOM causes manufacturers pain, even omitting a part. So we have to see what happens. It would be nice if Farnell / RS would keep it in place and lose a bit of profit as it makes extension boards available to non-soldering-iron users.
“it makes extension boards available to non-soldering-iron users.”
Unfortunately it also makes putting an extension board underneath a whole lot trickier. I know one soloution to the mounting problem I was considering was to sit the Pi on top of a baseboard.
Removing components is much trickier than fitting them in my experinace.
Aw, man, I am deep into design of an expansion board that is designed around the header being on the “bottom” / non-component side of the RPi board. When you say X-thousand, are you saying that “some” of the first 10K will be without the header – or were all 10K manufactured with it? I thought I read somewhere that the header would be left out to save its expense and improve the overall RPi costs. just out of curiosity, what changed?
I like the idea of the header being in place. if it is able to be placed by machine (not sure) then the cost impact would be almost nothing.
although a bottom header might be convenient for compact designs a top header will be great (and more usual) to give access to experimenters.
Looking at the nice X-rays, I was struck that it would be nice if you get a full end to end manufacturing video (even throw in a clip of it dropping on the doormat at home?).
I know my son would definitely get a big educational buzz from seeing how “his” R-Pi was manufactured – pick ‘n’ placers running full wack, robotic handling units, the works.
Just a thought…
Depending on whether the manufacturers are happy for us to do that, I’ll see if it’s possible; I’ll be doing a site visit in the next few months. (It may not be something we can do due to cleanroom rules and the factory’s preference.)
Be good to also include how the SoC is made too. I know when I first joined Broadcom the Introduction to the semiconductor industry was the most interesting course I went on.
Yes please! :)
Be very careful what you eat on your visit !! ;0)
I am a veteran of only eating in horribly expensive (and sparklingly clean) or properly recommended (and sparklingly clean) restaurants in China. A flight back to Heathrow a couple of years ago after a dodgy Shanghai dumpling, followed by an immediate transfer to Budapest, during all of which all I wanted to do was die, has cured me of street food for good.
You should ask Eben about the time he had the prawns in Mumbai, and the awful thing he ended up doing with a banknote with Ghandi’s face on it as a result.
Yeah – I heard about that !
Sure theres some ‘factory mods’ on my BBC micro too :P How bizzare! :P
So when can we expect another batch (another 10 000 or more?) will be in the factory?
It’ll be a bigger batch, and it’s entirely dependent on our ability to source those parts. I understand it’s not economical for them to build everything except that part, move everything off the line into storage and then back onto the line again to place a single jack on the boards, so everything needs to be done at the same time.
Just got my Farnell order info, it says that estimated delivery date is 23/07/2012. I didn’t expect that I have to wait 4 months :/
There have been a LOT of orders….
Quote JamesH: “There have been a LOT of orders….”
Allot or “A Huge Lot”… Due to popularity.
The RaspberryPi seems to be the best Bang for the Buck since the Commodore for personal computing. Others that barely come close are nearly $200 for something that you can develop on! There are E-Readers that are under $200 but you can’t change a thing on them without them glitching in some way or performing well.
No phy in the ethernet plug, oh crap…
What a waste (all the solder and desolder plug will be put in trash i guess…).
Take your time, after 6 month waiting, we aren’t at 2 weeks more. :)
but we want a good working pi!
No worries – thanks for being upfront with us about it, and thanks too to those who spotted the mistake in time!
I’m resigned to the possibility that it could be a couple of months anyway before I’m finally holding a Model B in my sweaty mitts ;-) If that’s the case, I’d far rather the machine was “right” before it left on its way to me (if that makes sense), so I’m cool with the news…
Take care of yourselves, everyone :-)
This shows the difference between the RP foundation approach and everyone else, getting the truth out there is far better than saying nothing, respect.
Now if only RS would acknowledge I have registered an interest…
This is just another trial by fire that further forges your status as geek heros that have been there, done that. You’ve overcome this challenge, and you’ll overcome the next one. We are all proud of you and your tenacity and perserverance. Onward and upward!
I can’t draw cartoons, but I have one in mind;
A LARGE pile of ‘Pis on a workbench, one full crate labelled “RJ45s (magnetic)”. one empty crate labelled “RJ45s (non-magnetic)” One ‘Pi on the workbench, with an operator busy with the soldering iron. He (or she) is singing to himself (or herself):
“10,000 green pi-boards, sitting on my bench, and when this one’s finished, there’ll be 9,999 green pi-boards … “
To the tune of ‘My Sharona’.
Let’s say that this is the second “event” that occour in China manufacture, that delay the shipment.
The first one was the smaller crystal event.
I am assuming this is the 1st 10k batch yet you RS and Farnell have been saying they were all done and in transit. Did it get picked up in the UK or were being misled?
Not trying to be difficult just getting conflicting info all the time.
Good question, i thought QA testing was meant to be finished two weeks a go today? And then the boards were being shipped?
It was – but it was done remotely, and we checked that everything was electronically sound, not that the parts were the ones we’d specified! We didn’t check that the network was working until they arrived here (electronically, it should have been fine if the parts were what we’d asked for), and then it took a while to work out what the problem was because we couldn’t imagine that the type of jack specified in the BOM and schematics might have been swapped out. (We found out when Gordon took a pair of pliers to a port in a desperate attempt to find out what was going on…)
There’s certainly some mixed messaging going on, because it’s definitely been stated that the batch had already shipped:
on March 5, 2012 at 11:25 am said:
What? The first 10k boards were shipped. Anybody who managed to order from the first batch will get them pretty quickly.”
Sorry about that. James wasn’t aware this was going on – only about six of us here at the Foundation knew there was any problem, and we didn’t let everybody who volunteers for the Foundation know about it until yesterday, when we were able to be confident the problem wasn’t any bigger than it actually was.
Sorry, I’d made an assumption on shipping from something I heard from a bloke in the pub who said he’d once met Eben in a massage parlour in Amsterdam. Or something like that.
I see so they came to the UK now gone back to China.
RS have said they are expecting delivery 9th March then will invite orders shorly after (Had a email from them) and Farnell were expecting a delivery the same time to start shipping 12th (Got a order in). I assume they didn’t just read James’s post?
I’m on a much longer lead time (end-April or May) so a bit extra won’t bother me personally.
But how come the factory get to change things without asking you?
That doesn’t sound like a terrific idea on mature reflection …
Will RS/Farnell be using the same factory, or will they be moving to one the can exercise more control over?
Well, they are supposed to use the components they are told to use. In this case, that didn’t happen. I’ve been led to believe this isn’t uncommon in many factories in an effort to increase profits. in this case, if that was the intention (I have no idea whether that is the case – it could just as easily be a simple mistake), then it has definitely backfired.
In my experience it never seems to be a mistake. Every time without exception it’s the factory trying to make a saving…..They are buggers for swapping out high stability/accuracy capacitors and resistors with cheap shit and then acting daft when you pull them up on it.
Sorry to hear that.
For me it doesn’t make much of a difference I guess, since I haven’t even gotten the mail with instructions on how to order.
If any of the card is destroyed in the re-solder process, I am interested in buying a broken card at full price. I’d like to have one to make exact measurements.
As I will be very far down the waiting list, when I am finally allowed to order, I don’t expect to receive a working card for several months.
Please consider offering a broken RPi B card for sale!
Yeah, same here. Got no email from either RS or Farnell and would be more than willing to buy a broken board if you’re left with any. Trying to fix it would get the thoughts off the rest of the parts I’ve gotten for a project, now lying forlornly on my work desk. :)
yeah I’d be interested in a broken board too, for housing design work.
If you have RPi B production card in your hands, would it (please!) be possible to present high res pictures (top and bottom)?
I don’t mind waiting if the end result is near-perfect. I had already decided to wait until I could just buy the Pi and not go on an “interested” list at one of those other companies.
Just more interesting stuff to go in the book ;>
Thanks for the update and being very transparent about the situation. Very few companies would release this kind of information, so good job.
I’m interested in who picks up the bill for this mistake, yourselves or the factory?
Maybe I can help out. I still have one of these jacks with magnets lying around here. As a little compensation for sending it to you I would only ask for getting one of these first batch boards.
Gold prongs… (the images were in the same order :)
I guessed because i’ve seen only the gold pronged ones, and they look higher quality.
Liz, please ask the factory to save a couple of hundred desoldered faulty eth jacks and sell them on the RPi store. I would like to own a piece of the RPi history. I already have in my collection the hand-cut RPi stickers from the 1st batch :-)
Based on four years of living and working in China I think the answer to the “which is which?” question is simple: whichever costs the Chinese factory less money. They would never accidentally put a more expensive component on a board. But putting a cheaper, inferior component on is encouraged, especially if the client doesn’t notice and it saves the manufacturer money. Bonus points for sourcing the original and replacement parts from a relative.
As other’s have said, thanks for letting us know – it’s far easier to be patient when you know what’s happening.
Glass half full thinking – at least you found out about this now when it’s easy to fix. Imagine the howls there would have been if you’d shipped 10,000 boards with iffy connectors. Nightmare!
“Anything worth having, is worth waiting for!”
I’ve seen this kind of thing happen before, the devil is in the detail.
Sometimes there are worse consequences. We had to cancel a project after 3 months of work as we realized some components would not be able to pass the customer requirements although the datasheet seemed to indicate otherwise.
I bet there was copious use of a wide range of expletives! Well, it’ll give the guys working on Fedora a chance to catch up… Though I’m poised with my keyboard and power supply by the letterbox!
TBH honest, the delay won’t be that long. Not enough to do a LOT of work on a distro, but enough to do a bit.
Fedora news got lost in this storm but it is downloadable:
Don’t worry – we’ll be putting a big post about it up tonight (and getting it in place on the downloads page).
I don’t mind waiting. I don’t even know when I can get my hands on a rpi board as both suppliers are not shipping to my country.
Nevertheless this is a good education, I knew and I’ve seen isolation transformer on ethernet/modem cards but never knew it can be integrated inside the modular jack 8O
Well I’m going back to try running reaspberry pi fedora remix on QEMU.
Liz I think this is a good time to put that positive news on the front page since Seneca has release the official image+installer. :D
We’ll be doing that this evening. :)
I download the image through torrent and it was fast and had no problem. I guess installer fetching the image from Liam’s load balancing system? It’d be interesting to see the download stats after the news got out properly :D
Seneca states on the Fedora Remix Installation page that microSD cards with an adapter will not work with the Raspberry PI.
Is this true?
At the moment there seem to be some conflicting opinions on this. We’re trying to find out.
I don’t think it should matter if it is SD, MiniSD or MicroSD… they all have the same contact points but in different sizes… The Main thing to look at is which Types and Brands and Classes(4-6-8-10)??? that will work with the RaspPi as some will include an extra unremovable partition for some reason or unsupported File System. A supported Brand/Part list would be most welcome! And quite possibly a simpler program to copy the RaspberryPi Image to the SD Card!(GUI is more common than command line)
Regarding the differences between SD and Micro SD I seem to recall that a SD card is required to support SPI I/O as an alternative protocol but the Micro SD spec does not require this. This means that some embedded SD implementations might fail.
What a pain for you guys, I hope you can get this sorted without too much trouble. Thanks for such a detailed post, it has been so interesting following the Raspberry Pi story from idea, to design, to manufacture.
Don’t worry about any issues you guys have. I’m using the Pi as research for my school about the new IT lessons they have and the Pi seems to be the best solution for them. Great Work so far guys!
It looks like Intel, that other small understaffed, undercapitalized provider of computers is having outside vendor issues of their own. Their troubles go all the way down to the silicon.
The Phoronix article you linked mentions the pain that the closed source PowerVR drivers and closed documentation gives the Linux community, and hence the relief felt at Intel moving to their own graphics hardware.
Broadcoms’ VideoCore IV graphics in the Raspberry Pi’s BCM2835 SoC has both closed source drivers and closed documentation, just like PowerVR.
What you say seems accurate. Looking through my parts bin of development boards that do much less, and cost a lot more that the RPi, I’ll let folks that work closer to the kernel ponder that. From my standpoint in user space on the external side of the GPIO pins: WooHoo!! An inexpensive embedded Linux system!
The spring shapes give it away ;)
I’m disappointed that the error was made, but massively grateful that it was spotted and is being handled.
I could source and resolder a new connector (although maybe not if 10k others were trying to source it as well) but not everyone could. The only reasonable decision is to keep them until they’re fixed.
Will future batches be remote tested? Are you getting hdmi, Composite, audio LAN, USB and keyboard bashing robots in place (of course they’ll need webcams etc – they could be built around an RPi ;)
Now we’re working with RS and Element 14/Premier Farnell, who are large enough to have bodies on the ground in China all the time, things will be being checked out there by their guys. We’ll also be making some site visits ourselves to satisfy ourselves about conditions and build quality.
The one on the left in the “visible spectrum” ( :-) )shot has te same squiggly EMI connector brush thing as the one on the left in the X-ray.
That’s a drag. They are hard to tell apart without the X-ray.
Wait a sec… “brush” is just a spring like XAPBob said. Coffee time.
I feel amazing having the chance as a medical student to work with amazing electronics like this! I don’t mind waiting a couple more months.
Thanks for being so upfront about the issues you’re having here – it really helps to understand what’s happening.
It must have been really depressing to see people (all over the world) getting upset about shipment dates – but I for one appreciate the effort you are putting in am am sure RasPi will be a phenomenal success…
BTW, would be great if this site allowed us to like/unlike other people’s comments… although admittedly it’s probably not one of your priorities right now ;-)
As so many other people here have suggested, this WordPress blog needs to be nuked from orbit and a more usable system put in its place. It’s not like there’s a shortage of good ones. The continuous “Click back” faults (which continue without change after Tuesday’s hardware upgrade) and the lack of useful features are just two of the problems, but there are several others. It’s a a pity, because good communications are very important.
As so many other people have replied, this WordPress blog is due to be replaced shortly, using the phpbb platform.
So, will all future orders from this place include one randomly pulled and torn apart by volunteers looking for substitutions?
I haven’t ordered one yet, I figured I’d wait until the turnaround was somewhere in the weeks instead of the probable months, but I’d be interested in getting one with the upgradeability of the first x-thousand. Has anyone proposed manufacturing an in-line isolator? I barely know resistor code, and can only add up parallel resistances if I think about it for a moment, but it seems to me you could put that transformer in a pigtail. Maybe the company would offer a discount to avoid having to desolder so many parts?
Ha! No. We’re keeping that one.
Doh, That sucks, but better that you found it now and can fix it before 10,000+ of these things are out in the wild and we all start screaming about it either not working, frying the RPi’s, or worse, frying our other network equipment. MUCH easier for you guys to fix them now and get it squared away…
I just hope you can find the parts you need quickly…
If this was a result of a factory making an unauthorized substitution and, worse, not telling you, perhaps a delay to reevaluate this particular factory is in order before going from 10K to 100K batches?
The work you are doing is beyond fantastic, but there are few things that will demotivate a kid faster than getting an educational “toy,” spending time setting it up, and discovering it doesn’t work. That means that what this factory might consider an acceptable defect rate might not be appropriate for your project.
If this factory wants to keep your business, I’d think at the very least it should send you the existing batch for free, and do up the next 10K from scratch. If it doesn’t want to keep your business, that might be best for all concerned. Maybe a slight price increase combined with a switch to a British (or Chinese) factory with a Caesar’s wife reputation would be best?
Sounds like you have a good supply of model A+. 2 USB ports. No ethernet!
As part of the transparency can you please clarify the number of boards are being fixed with the in stock correct parts.
Is it the “couple of thousand” that INQ were told were “imminent” before this issue surfaced? Or the full 10K? or somewhere in between. Trying to understand whether the sourcing delays kick in after 2K or 10K. How many parts have been re-sourced?
Hope the resourcing goes well and glad you have farnell and rs involved in procurement. Still hoping cases and bulk orders will be available before Sept, but one step at a time.
If that is the reason, I totally agree.
Well you are discovering the downside of offshore manufacturing. I spent many years purchasing PC systems and PC parts some of which were made to my design (the world first “green” motherboard) all were from the Far East
You must keep a scrupulous eye on what is coming off the production line. Substitutions are common most of which are OK but some, as you have discovered, are not OK. Sometimes the problem is greed, sometimes it is “benefit to customer” (I forget the Mandarin for that expression) and sometimes it is a genuine mistake but always the play of different cultures comes to the fore. I was purchasing a high value each month so I needed my own local team of inspectors who came from a different company.
You must watch every thing and do some batch testing. I would have thought that Farnell and RS (both of whom are current suppliers to me and our clients but on not so grand a scale) should take care of these issues. Look at your contract. Question absolutely everything (re the time I found plumbers solder in one of the solder baths – had to refit ~500 power supplies on sites across Europe) check everything and dot ALL the Is and cross ALL the Ts.
Just because it is lower cost manufacture does not mean that benefit comes at zero admin cost. Good luck and welcome to the world of the micro penny.
PS Why have you chosen an RJ45 connector with a shield connection since I assume you intend us to use Cat5e UTP?
Wise words indeed..
Delays or no delays, I will eventually get my pi, and when I do, I will make it do all sorts of cool things!! You Guys Rock!!!
In 2010 TI announced their MSP430-based Launchpad development tool Global in size though this company is, they simply could not deliver their product to satisfy demand. I waited months for my kits to arrive. The Raspberry PI folks are in good company! :)
Seriously, the open, candid approach taken here is a testament to the philosophy of the product and of the folks behind the scenes who have laboured to produce something
I sure hope thing else is wrong – Chinese manufacturing can be excellent or questionable. It reminds me of John Wanamaker’s quote: “Half the money I spend on advertising is wasted. The trouble is, I don’t know which half.”
I hope the correct version has gold-plated connectors, and that other connectors are gold-plated as well. I understand the difficulty in getting the correct magnetic Enet connector — there are many different pinouts and you need the one that matches the PC board. I found this out when I first looked at the RasPi PCB artwork and was puzzled that the receive and transmit pairs didn’t seem to go to the correct pins. With a bit of searching I found out that there are all sorts of different configurations :-(
Seriously ?? You want gold-plated on a $35 device ???
Come on…. Don’t you think that’s just a little too picky ??
Everything is gold plated- I have devices a seventh the price with gold plated connectors.
No sweat! I think we’re all learning here!
Pandora had many similar problems with factories in China. In the end they gave up and moved it all to Germany.
Looking at their forums they also had problems with quality control and DOA boards.
All units really need to be tested properly before sending out. The failure rate will be key to success.
OpenPandora’s story was at the top of our minds all the time we were in development, and was a big factor in our deciding to seek a manufacture licence agreement. I still get all shaky whenever I think about it; they’ve had awful luck, and I hope it turns around for them soon.
I’m guessing the coppor coloured one is the magnetic one?
Good that you do some fine testing of the boards. Thanks.
Ok, who put a lemon in the Raspberry Pi?
my uneducated guess went with the real one being the one of the left, for the sheer fact it looks genuine, like you’d pay money for it! the one of the right looks horrendously shoddy :o like something you’d get off fleaBay 10 for 1p with Free P&P from China!!
Glad you found it before they got sent out, you guys have suffered enough already!!
Guess I’d rather have one which works a few weeks later, than one which doesn’t sooner.
Mind you – I’d rather know I was getting one at all… Still no news from RS.
Of all the possible things to go wrong, this seems like a pretty minor one. Glad to hear that you’ve had some success in securing the correct part, too! We’re patient :)
Also, I have to remind myself sometimes that the net goal of the RasPi is not in any way to give hobbyists a cheap machine to play with, and that that’s just a wonderful side effect of the hard work that’s been done to make public education better in general!
Yeah – we spent some time trying to cheer ourselves up over this by discussing the much worse things that *could* have gone wrong and didn’t. A business like this has so many moving parts it can make your head spin sometimes.
Who has their own X-Ray machine? :-o
The way I hear it, anyone who wants to hook an old tube, (valve?) to a neon light transformer, and also wants to forgo any future children.
It’s good to see the comments being more supportive then with the release. They will get out, we know that. It’s just a matter of time, and it’s better this be discovered now then after they have shipped.
Just a notation. I seem to not be able to see the bottem xrays.
Would be possible to sell part without the Ethernet connector? I can buy one here from Farnell or RS and solder it by myself.
So, I will not have my ideas delayed :-)
No that won’t be possible, I’m afraid. The sub-standard units are being repaired, and all new batches will be correctly furnished. In the future, some such variations MAY become available, but at this stage, I would say highly unlikely for some time (possibly years?) to come.
A little bit of love from Lady Ada :-)
That awkward moment when you only see positive comments and the comment section is moderated…
How so? AFAIK, first posts have always been moderated on the blog -part of the anti-spam defences.
Really? Your comment’s here.
We do moderate comments here – we always have done. If we mod a comment, we always leave a message saying why (although given the number of comments we get here, we have to do this very seldom.) We moderate for civility; we expect people to behave as decently here as they would if they were talking to each other in person, so if they are repeatedly oafish or trolling, they are warned they may be banned; and we have to moderate out bad language because kids read this. We do not moderate out dissenting opinions; on the contrary, we encourage debate here. We do not moderate out criticism. Have a look at some other comment threads; you’ll see plenty of dissension and criticism on all of them.
What *does* get modded out without comment is comment spam. But I am pretty sure you’re not interested in a stream of messages from me or the spambot saying we’ve just removed something about cam girls and fat pills.
To be honest, I’ve been surprised at the lack of negative comments. I was prepared for the worst, and apart from one or two comments on the forum, it’s all been very civil. Thanks everyone for understanding!
That awkward moment when you actually read the blog and forums and realise that you are utterly and completely wrong…
I was surprised there were less rude comments than some others i saw about a week or two ago somewhere or other here. @scep: who are you addressing that to?
Brian (it’s in his reply container!). I forgot the wink smiley :)
The journey is all part of the fun
in case there are any interested aussies reading, i was just able to pre-order a model b board from http://au.element14.com/raspberry-pi/
Knowing Chinese manufacturers and having bought many electronics from China, I would almost say they did this on purpose to increase their own profits.
I actually hope they did that, because the manual labour they need to put into the fix is going to cost a substantial amount and will increase the chances of breaking the device. If it was done accidentally, they’d have to pay exactly the same yet they would not actually learn that fake components are bad.
This brings back memories of the Bad Caps drama.. MILLIONS of devices had to be recalled due to cheap Asian rip-offs (I can’t remember if it was China or Taiwan actually). I can totally understand why manufacturers have been reluctant to use electrolytic capacitors instead of solid capacitors
How large is the first batch anyway? I’ve heard 10 000, true?
I ordered mine last week and seeing how busy things are I’m assuming that I won’t be getting one from the first batch. In that case, since I already ordered, I’m also assuming that the factory has received orders from RPi for the next batch already and are probably working on it. So how much have been ordered so far?
Thanks for being so upfront about this and even turning it into an educational experience for us! This is what Raspberry Pi is all about.
And I am really impressed with the comments so far on this news. Way to go Raspberry Pi community!
Pretty clear from the angle of the flaps on the xray that the one on the left has the magnetics in :)
I really appreciate the transparency here. It’s neat to see the hiccups along the way, it makes me feel like I’m part of the team or something.
These things happen, we understand!
Perhaps not bad that I’m getting a second wave RP.
Bummer, but hey at least you found it before I spent days checking my network looking for the problem…
And here we are, thinking that *designing* the thing was the difficult part…
Well, I’m sure Pete and Eben would agree with you there!
Form long experience in the telecoms business the design bit is the fun part.
the supply chain/logistics part of the process requires a lot of attention to set up correctly and then to ensure that it continues to operate correctly. not fun but very important.
It will be interesting to understand how this problem came about. Part substitution is normal business but should require approval in advance. Factories are only authorised to build to what is in the bom. there can be issues where the bom spec is not detailed enough but this would seem to be a fundamental difference.
The one on the left with the golden pins!
— Tom Mcloughlin
I’m gonna say it’s the one on the left. The attachment prongs on the sides are bent in a different direction that doesn’t look like it’s locking something in (as the right does), but instead pulled into a groove by magnetics.
Such minor problems will happen in every manufacturing process. I guess the part on the left is wrong and shoul dbe substitued by the jack on right side. The left one looks like made of bronze and is not magnetic and should be replaced. Perhaps the manufacturer thought that the bronze one looks much better and makes the RapsberryPi more high quality looking.
Actually, the one on the left is the magnetic one, and I thought that was gold plating/coating.
Do we have to expect mails from Farnell/RS with changed delivery dates?
I expect they will be sending something out when we’ve nailed the sourcing problem for the correct parts.
Interesting how sanguine the community seems to be about this – Just telling us all about the problem has defused the backlash.
I’m still looking forward to getting my paws on a pi, but I’m not really in any hurry – I have too much to do as it is. Like the rest of the folks, I’d rather it was right than on time.
Glass at 50% capacity and holding.
Yes indeed, AllyR!! I hope the Foundation has noted this after the success of this announcement about a problem.
Open and honest communications is what keeps communities happy and working together, rather than against each other. I’ve seen it go downhill in a number of projects when a company closes ranks in what it thinks is a good PR move, but isn’t. When a company shares its problems without spin, the community stays on its side. Secrecy and spin creates an “us and them” split, and ruins the relationship.
The Foundation did quite well here, and so everyone is supportive.
(But perhaps internal communication needs a bit of attention, as JamesH misinformed us about the units being shipped simply through lack of accurate knowledge about it, as Liz noted. These kinds of faux pas are picked up rapidly in such an attentive community. :P)
Full marks here.
I hope Farnell and RS have the same commitment to quality as the Foundation. It would have been easy to ship out the faulty units, and only discover the fault later, with less diligent checking. Though some have said they would have been happy to get a broken one and fix it themselves, I would rather that the supply chain starts out with the right attitude on sticking to the spec. Otherwise what dodgy part will you get in the next batch?
Thanks very much for the transparency and honesty – we were beginning to wonder why anyone, much less 10,000 people, hadn’t posted the unboxing of their shiny new Pi boards.
As a fellow engineer, I can definitely feel your pain encountering yet-another speed bump. However, when did they stop teaching Murphy’s Law in engineering school or by mentors in the workplace? I suppose some good news is that the First Corollary to Murphy’s Law didn’t happen, “If more than one thing can go wrong, that which will cause the most damage will occur first” (as far as we know ): ). That could have been the case if boards with the incorrect connectors had allowed voltage offsets to interfere with or damage other equipment.
This is an example where precisely the kind of technical education we’re trying to foster with the R-Pi is critical – someone in China either didn’t understand that connectors that look alike can be very different, or purposely ignored the difference (probably some business idiot with no technical background who over-ruled the engineers with the necessary knowledge – see also both Space Shuttle disasters).
There’s a very good reason why Chinese manufactured products are cheap (very different from inexpensive) – and if anyone thinks this was an “honest mistake”, you haven’t been paying attention. This is a country run at every level by a government that allowed melamine plastic powder to be used as filler in baby formula and lead in paint on children’s toys. This is a government that allowed concrete school buildings in earthquake zones to be built with no steel reinforcing material. Businesses have to have senior military officers as non-contributing executives and major shareholders so that they can profit undeservedly on the backs of slave laborers. This is the government that not only allows, but, encourages, the “accident”-guaranteed conditions in factories recently revealed at Apple’s contract suppliers (and those are the _best_ places to work – try to imagine life in garment factories). This is the same bureaucracy that uproots tens of millions of people from their ancestral homes with no recourse so luxury condos can be built that only the government elites can afford. This is a police state that oppresses every minority with the intent of driving their cultures into exile, but, preferably, outright extinction. All of this is going on and much, much, more, such that there are over 100,000+ violent outbursts each year by their own people, and climbing every year – but, no one outside hears about them thanks to the complete control the government exercises over every avenue of communication.
Former CEO Andy Grove best summed it up in the title to his book, “Only the Paranoid Survive” – the Chinese manufacturers cannot be trusted any farther than an Ethernet connector can be tossed. This is what our hard-earned money is supporting while we’re destroying our own industries with economic starvation on top of excessive bureaucracy. The good news is that automation via robotics is more than leveling the playing field as well-designed robotic systems can outperform even the cheapest laborers in the world, especially as economies of scale ramp up into the millions. We’re starting to bring jobs back to the U.S. because of this advantage, but, robotic machinery needs to be designed, installed, operated, and maintained by people with more education than the old manual assembly line jobs.
Hence, we get back to the real purpose of the R-Pi – educating _everyone_ on the importance of at least familiarity at a fundamental level with the technology on which our very lives now depend.
gosh that’s a tough comment.
my experience is that with the right processes in place manufacturing in china can be of very high quality (quality = doing what was specified)
Jim Manly said (above): “Chinese manufacturers cannot be trusted any farther than an Ethernet connector can be tossed.”
Yes, that is a little harsh – SOME Chinese manufacturers may be untrustworthy, but then again, many are okay, and some indeed are excellent. I haven’t any direct experience, but I do know of a railway modelling company that has produced some exquisite example of models in brass, that are absolutely perfect, delivered in good time, and (relatively) quite cheap.
By the same token, there are many western-based manufacturers who fall far short of ideal too.
Do we really need such a political rant here ? ;0(
China bashing. How constructive ! And the arguments sound so much first-hand experience…
Why don’t you blame the Foundation to have the RPi manufactured in China while you’re at it ?
Do we really need such a political rant here ? ;0(
A delay is a delay, I’m Suprised one hasn’t happened yet! I can’t wait to get my order in and have one (in june from the looks of the preorders lol)
I am outraged! Rabble rabble rabble, want my money back, etc.
it’s not like you guy’s are a 10 years old company these thing’s are bound to happen the important part is how you deal with them ;)
BTW i’m getting 4 :P
Geez this thread is interesting. Just wondering now which RJ45 connectors are in my router.
Thank you for this information.
Unlike at the launch, where information was clearly hard to come by (for you guys as well) this is a great case of full disclosure. I’m sure some people will be disappointed by the small delay, at least we now all know what the problem is and that you’re fixing it, so we don’t have to be concerned.
Now, if you can tell the difference, then it’s a badge of honor if you get a Raspberry Pi with the re-soldered connector. :)
I’m a little disappointed by RS Electronics… the last time I checked (yesterday or so) their FAQ stated that they are going to inform the people who registered by the end of this week about their registration status.
Now this information has totally diasappeared from the website which equals a delay (in my opinion). I can live with a delay in the process but I’d like to be informed about that.. so a little more transperancy would be appreciated.
Nevertheless, a big big thank you to everyone involved in the foundation. I cannot even imagine the amount of time and work that went into this project, but the results are very imnpressive so far. Keep going!
To the team: take the phone off the hook, take the website offline and have a good weekend.
Things have a habit of coming out right in the end. Chill everyone.
Eben and I are actually going on holiday (kind of – we’ll be working through the evenings while we’re away, but this is more than made up for by the fact that we’ll be skiing in the day) on Saturday. I am actually kind of looking forward to the 12-hour plane journey in a way I never have before, just because there won’t be any internet access so I can sleep without feeling guilty.
it was with some sadness that I read yesterday that a carrier is trialing wifi internet access on trans-pacific flights.
one of the great things about loooong flights is that you are out of touch!
Whoever found this mistake BEFORE the cards all went out deserves a medal.
Would have been much worse if 10,000 new customers started complaining…
After all no-one wants to see raspberry crumble!
As we say in my country, “ah seriously now”.
No sane technical person in the entire world would take delivery of the first volume batch of their first product from their brand new subcontractor, take one look at it, go “hmm, that’s vaguely the right colour”, and ship it on to customers without testing at least one of them.
And since the fault was with every single board of this run and with such a highly visible impact, a cursory functional test would have discovered the problem.
So it’s not really a surprise that the foundation, who appear to be quite sane for the most part, found the problem.
Praising Liz and Eben and James et al for discovering that networking didn’t work on the first volume Pi run before shipping them is a bit like praising Rory McIlroy for knowing that you hold those golf stick things by the black rubber end: you wouldn’t expect anything less.
Sanity is a rapidly reducing resource in Foundation territory. It’s been a long hard slog (for other people- I just waddle around here doing stuff)
My problem has been that though I checked out both sites Farnell was the one that had them to order and yet wanted me to order even more stuff because the board is/was not priced to hit their minimum order value before shipping to the US.
Therefore hubby is in ‘wet diaper’ mode because the place he found in the US to provide them (http://www.alliedelec.com/RaspberryPi/) isn’t selling any yet.
The jack with magnetics has golden color and is on the left. Just look at angle of these metal holder “wings” (I do not find the correct english word) inside the jack.
Paul M. Suckow
This kind of transparent process has GOT to be in everyone’s future! Won’t go so far as to be happy hiccups happen, but this one really helped open my eyes to the possibility of a more cooperative future! Can you imagine government, manufacturing, distribution, relationships, sales all helping out together for good and loving it?!
I suspect this falls under tragedy of the commons.
Mind you, the following is speculation, but falls inline with my rather pessimistic view of the general populous.
If everyone plays by the rules, everyone wins. Or at least the average of everyone comes out ahead.
But if you withhold one hard won lesson, maybe your competition will fall into the same trap that cost you, putting you back on even footing.
If you are concerned your competition may have not shared that one pitfall that could still catch you unawares, you might want to retaliate in advance.
If you are concerned there may be some efficiency you lack, or don’t want to employ, knowing where the pitfalls are, so you can dodge them but leave others to play catch up, using those efficiencies you don’t have, leaves you ahead.
Think of it like not warning the new guy the boss doesn’t like certain jokes….because you really want someone else to be the goat for awhile.
This is one of those areas where non-profits have an advantage.
This type of community is a sort of gift economy. Status is won by giving stuff away.
While commercial interests might dip into giving away knowledge to gain kudos, it’s what makes open source communities flourish.
He who teaches the best has the highest status.
That punch, I know exactly what it feels like, happened to me yesterday… No worries, as one of the prev posters said “shit happens”. yeah it does. moving on and insisting on what you want is the key.
Youtube/diggv3 redesigns their webpage, millions goes apeshit and demands the webmaster to be beheaded.
Raspberry Pi team reports a fault on one of their product’s ethernet port, everyone(including the buyers), remains calm.
Makes a whole lot of sense…
Something not right about this… if non magnetic connectors mean no network connection, then why do they exist in the first place???
Murphys Law – Anything that can go wrong – will go wrong. :)
I am so exited to get hands on my RP, but theese things takes the time theey nneed. You girls and guys are doing such a great job!
Left has magnetics. You can see by how the pins are straight both on the xray and the picture
Oh well, you can’t do anything that big without having an error somewhere. Don’t let it get you down though
One thing I don’t understand. RS and Farnell are independently producing the Raspberry Pi, aren’t they? Or are they both putting orders in to that same factory.
If not, how can it be that this mess up by the Chinese factory affects the production by RS and Farnell?
Android Overload: ZTE Fury Coming To Sprint March 11th, T-Mobile MobileLife Sweepstakes and More | Droid Universe
[…] Manufacturing hiccup could cause Raspberry Pi delays. [RaspberryPi] […]
Whoops… oh well, at least it was spotted in time rather than them going out to customers and then having to go back to be fixed (although I guess a few customers will have the skills to fix issues like this themselves).
Not the end of the world :-)
The one with the gold pins is the nagnetic version (hint: there is a physical give-away)
Sanguine about this? No. I for one am absolutely sick of the whole debacle at this point. As of now nothing has changed in a week – Have we got one? If you ordered from RS – nobody knows still today. We don’t know if we are even IN THE QUEUE for the first 5,000 despite them already having the timestamped info about who registered first. We don’t even know IF we’ll be one of those who can place an order or not, or when we will find out.
Sorry you feel so strongly. I am not sure what the delay is with RS – I think Liz is trying to gather more information so hopefully will be able to put up a blog post when all the facts are known.
Only thing I would say is that whether you know you are in the list or not doesn’t affect where you are in that list.
richard a wenner
Liz, May I suggest that it could be wise to remove statements such as “We’ve known about this for four days now, but we haven’t been able to tell you about it” as it tends to continue this unpleasant, under-wash of superiority that is starting to pervade this site. Yes, you are being honest but it is at the expense of this ‘we know better’ attitude. It’s a little like the drip feeding of information about alarm clocks and continual delays in providing downloads. Now please don’t take this the wrong way the work you are doing is phenomenal and a credit to the UK. The real impact of this development is only likely to be really appreciated with hind site. This development is a revolution but it’s being spoilt by this emerging attitude. My suggestion does not come as any result of lack of sleep from getting up early in the morning to find crashing servers or the bitterness (I’m not bitter – no, really I’m not) of not having a Pi in my grubby mitts now. It’s probably just the frustration, felt at this end, of being treated like a dog on a leash. Bitter woof, bitter woof. I still believe the first 1000 units should have been sold at a considerable premium – the market would have accepted that as it would have chocked of the excessive demand and raised a fortune for the Foundation (where is the buy one donate one button on the FEC/RS sites)?
I’m not quite with you. I’m not sure what superiority you are talking about. The Foundation needed a few days to check the boards out and make sure there were no other issues. Once that was done, we posted about the problem. Had we gone straight out with the issue, and then found another one, and had to do another post, that would have been much worse and would have led to some very bad press indeed – one mistake post acceptable – two separate mistakes inexcusable. Consider how other companies let out information; the vast majority would not give you even this sort of information. It comes down to not washing your dirty linen in public – even for the Foundation, not everything gets released – some stuff needs to be carefully handled or the whole project could be put at risk.
Re: increased prices for the first batch. That would have inevitably lead to many people crowing about how we were unable to keep to our prices, and, in our opinion, was not worth the risk it would have put on the rest of the project. A very big part of the popularity of the device is the price point. That sort of market awareness would not have come had the price been much higher at launch.
Unpleasant superiority? Really? The fact that you infer this from a simple, honest statement says more about you than it does Liz.
richard a wenner
Nope, just that the stories regarding shipping don’t add up.
no frills, straight to the point and totally informative………………….if you could go into politics we’d be sorted in no time.
Obviously huge cost savings by manufacturing in China (even with “hiccups” that could occur anywhere). Is there any possibility of considering a Pi made in Britain? Even at three times the price I’d happily pay that to encourage electronics manufacturing in the UK as well as educating our young to learn programming.
The Foundation could of got the boards at a reduced price from farnell/RS then sold them for people who have the skills to solder and replace the defective RJ45 themselfs. This way if you really want a PI you could of bought one and you are at least adding a componant yourself. so you not only learn programming you learn how to solder. just an idea.
Sorry to hear about the hiccup. I guess everything will be fixed until the second batch. Btw, any plans yet for the next delivery date?
Sorry to say that but in the meantime I am thoroughly disappointed.
I know many people have spent voluntary work and effort but I’m pretty sure that there would have been a better way to manufacture and sell it.
Starting with a strong “why we do not accept preorders” statement I’m now left with less than a preorder from a third party (…I was just about able to express my interest in it…) and there will be only one per person. Somehow I’m loosing confidence that I will ever get the chance of ordering one at all.
I guess you could have sold millions of pieces by now.
Why wasn’t there any interest in it?
Open to suggestions on the better way, given the constraints the Foundation is under (limited funding to go it alone is a big one, esp. for big batches). Although RS are a bit behind with regard to ordering, you shoudl be able to order from Farnell (I think). They are using different mechanisms.
We couldn’t have sold millions, because we could not have made millions. And you only get money (in the EU) when you deliver a product.
does this sound familar?
Got a mail from RS saying im in the first batch. :) Exacting!
Magnets, how do they work?! =P
Raspberry Pi sufre problemas de fabricación, se esperan pequeños retrasos en la distribución
[…] día, los problemas no parecen abandonar a la pequeña fundación, que ha confirmado algunos problemas de producción en Raspberry Pi Modelo B, actualmente el único a la venta, que podrían conducir a retrasos en la distribución de nuevas […]
Why is everyone saying “yay transparency” to this post? It’s obvious that despite numerous vociferous comments to the contrary, these units were not shipped to the UK when we were told they were. Otherwise, surely someone would have said something along the lines of “well we got the first batch delivered here, and had to ship it back”. Instead, we’re told “the factory is already replacing the bad connectors”. Well at this point, how can we be sure?
First you said no preorders, then you said you’d be selling directly and that the rpi.com store was hosted by a super-awesome provider to handle the load, then we got sprung with the RS/farnell farce. Oh, and despite multiple assurances that sales wouldn’t be opened until the boards were en route, the factory in china still has possession of the boards! in fact, the production run wasn’t even in *testing* according to your own reckoning, until the day *after* they had supposedly shipped! I have absolutely no beef with any of the calls you made as to where to sell, who to use for manufacturing, etc, all the issues are perfectly understandable for a new hardware startup. What is NOT acceptable is the constant bait-and-switch. Also how on earth does it take 4 days to perform acceptance testing on these devices? Surely this remote test platform we’ve heard about can handle more than one rpi at a time?
They are saying “yay transparency” (apparently) because they have a certain amount of trust in the Foundation and they find the honesty refreshing.
Whereas you are just making up your own stories and your own little straw men, and then whacking them down. To what end I still haven’t worked out.
Huh? What? Do I really have to go through and link the posts where these statements were made? Either you haven’t been following the rpi very closely, or you just like to shout “strawman” any time someone says anything you don’t like.
How do you correlate “transparency” with picking a totally different distribution method than the one stated, WEEKS in advance (see the slashdot interview), and not telling anyone until the day of ‘release’ (which turned out not to be a real release, it was a pre-order, plain and simple, the boards weren’t even ON THE WAY from china, or even tested!). How do you correlate “transparency” with telling everyone the boards were to be shipped “monday at the latest” when they hadn’t even been tested yet? I don’t attribute any of this to bad intentions, obviously Hanlon’s Razor applies, but please don’t say the process has been transparent when it has been rife with bait-and-switch.
I am guessing that you have never been involved in the setup of a hardware company. Things do not work at all in the way you are suggesting. We received one initial small test batch of boards which did not have these jacks on them; they tested fine. The first production batch had a different jack in place and it was that batch (we test every single thing that comes from the factory) that failed testing. And as for not telling you about our licensed manufacture agreements in time (in time for what? What would you have done with the information? This is just a wild guess, but I imagine you would have decided you didn’t like it and complained about it) – well, I’m terribly sorry, but negotiations on these things, during which time everybody involved is under NDA, take weeks. Contracts were signed and finalised only a few hours before we were ready to push the button. We would love to have been able to run our business precisely to your convenience, but it’s not the way the real world works, even for solipsists.
And any more accusations of lying, bait and switching and so on are going to get you plonked. Would you be happy accusing us of these things to our faces? They’re not true, we find them insulting and offensive, *especially* since we make genuine and enormous efforts to inform you all of what we’re doing to the very best of our ability; and I’m not prepared to pay for bandwidth which you are just going to use to tell me I’m a liar.
Liz, it is a fact that things went differently to what you kinda promised.
– Here http://www.raspberrypi.org/archives/509#comment-7895 you say that the you make sure that there will be enough capacity for the launch.
– And here http://www.raspberrypi.org/archives/320 in the launch of your shop you also state that one of the reasons for the sticker sale is to test your shop system for the eventual launch.
This made people to anticipate that you are prepared for the launch and will handle it yourself with hardened servers and shop software.
But when you launched, it was through Farnell and RS, neither of which was up to task. Neither were you.
You also promised (I will search the comment if you want) that you can order RPi regardless where you are and regardless wether you are a individual or company. This didn’t work either.
The fair way to handle the launch would have been to sell the initial 10k batch yourself as you promised and then let RS and Farnell handle upcoming batches.
You also promised to keep registered people up to date. You did not.
Further, during past months, you stated many times that you have adequate funding, you won’t need money from preorders etc. Now you say that you really cannot reach your goals without using RS and Farnell. What else is this than using their financial backbone to support the project?
Please look this from the perspective of your loyal fans. More humble approach would be proper response insetad of threatening to plonk uncomfortable opinions.
The things you mention were part of the plan BEFORE half a million people (or more) decided that they wanted a Raspberry Pi. Things changed radically, so the plans changed radically. That’s how business works. At one time the RasPi was going to be USB format; it was going to be made in the UK; Model A was going to have 128Mb RAM.Things change.
Things have also gone wrong. When they have the RPF has been open about it.So I really don’t understand why you are saying that things “went differently to what you kinda promised”. Because the Foundation know this. This isn’t “lying” or “breaking promises”.
But still a small percentage of posters just come to say: “you lied to us” … “but you promised” … “you should have done it this way”. But that’s all, nothing constructive, just poking holes and complaining. These posts are very different in tone and content from welcome constructive criticism or genuine concerns and queries.
It’s hard to stay humble when you’ve read such comments for the 1000th time, especially when the poster is often far from humble themselves (usually quite self-important and demanding in fact). Liz does a sterling job dealing with such posts. On most boards they would simply be disappeared as they add nothing to the debate, are often offensive and yet take a disproportionate amount of time to answer.
So don’t confuse “humility” with agreeing with all and sundry that they in fact know best. Don’t confuse “superiority” with having to deal quite abruptly with some of the posts here. And don’t confuse “uncomfortable opinions” with whingeing for the sake of it. The Foundation has better things to do than deal with these posts, like trying to get tens of thousands of RasPis out as soon as possible.
Any further posts – from anyone – that say nothing more that “you lied to us” or “you are incompetent” or “I could have done it better” will be deleted. I hope that it transparent and open enough (me – I’m not so humble ;)).
Raspberry Pi $35 mini-computer suffers more setback and delays due to a production mistake | Daily Tech News
[…] and attempted to scold the complaint as fast as possible. It also went on to contend the emanate is minor, and factories are roughly accomplished with replacing the improper tools to functioning ones for […]
So, we’re in a ‘Raspberry Jam’ ?
Yup – it’s just like the launch of the BBC Micro! Manufacturing glitch delays initial orders. It’s kind of predestiny for all the mistakes and bugs that will hit Pi owners as they experiment with their little devices – you learn by making mistakes.
It is not clear what happened other than the wrong part was used.
Did the manufacturer of the Raspberry Pi computer make a part substitution of the ethernet connector?
Was the manufacturer shipped a wrong or mis-labled ethernet connector?
Was there a mistake in the documentation as to what part to use?
Or none of the above, it was ….
How can this problem translate into a four-month delay for delivery?
This problem is not why you’re being quoted a four-month delay (and it seems to have been a part substitution) – the fact that we’ve had a simply incredible number of orders (beyond what anyone expected) is the reason for that, but unfortunately this has slowed down the release of the first batch and may cause a bump in later manufacture if we can’t source the correct jacks fast enough. We hope that those times will be brought down, but they’re the way they are because our manufacturing partners have to build up enough stock to satisfy everybody – and then some, because in a few months time we need to be selling in bulk so we can satisfy schools and businesses, as well as the people at home who want to own more than one, too.
We don’t have pockets as deep as RS and Element 14 do. If we’d stuck to making them and distributing them ourselves, that date would likely be four years in the future, not four months. Look at the history of the Open Pandora project for an example of how hard this can be if you’re doing it on our sort of funding without manufacturing partners.
from Bogota ,Colombia.
I need one Rasperry Pi B.
Where in Bogota
I do not have a problem with this. Delays happen. However, I signed up for notices about a year ago and have NEVER received anything. When I tried to re-enter my email address, it said that I was already entered. I did not receive this information from you. I saw it on: http://goo.gl/hTi90
How do I get the current notfications?
There have been problems with the emailing due to a very big mailing list. I suggest keeping an eye on the twitter feed, or just popping back to the forum home page occasionally just to be sure.
Worse mistakes could have happened but what a shame that the remote testing of the early boards did not pick this up. I am intrigued to understand what is meant by a magnetic connector and why this means no connection. All the best and good luck with the re-soldering. Still waiting to be able to order via RS.
Just commendable that you put this up and explained it so clearly….instead of what we’re all used to seeing from our hardware/software vendors: “Something went wrong, Sorry, we’ll know something soonish”….kudos for that, and for the project in general.
Why are integrated magnetics required ? What are we talking about here ?
#1 ferrite beads ?
#2 inductors ?
#3 transformers ?
#4 other ?
I don’t need Ethernet, I’ll gladly take a defective board :)
I hope everyone …. RBPI , Farnell, RS & ??? start doing better in the area of manufacturing, logistics and fullfillment…..
Typical english service
YAY! After sooo looong of wait I was able to Pre-Order my RaspberryPi!
I had to “Register My Interest” with the Element 14 Website.
I received an E-Mail back stating “We’ve received your request and we are holding your place in line…”
Then I received another E-Mail later with a personal Order Link for Pre-Order.
My cost $38.96CDN+Tax+Freight
Delivery expected late July… I hope that we do not have any more trouble with the Manufacturing and have them on our doorsteps soon!
FYI Element 14… I am a “Person” not a Company.
Hey I used to work in a CSIRO electronic development workshop, after some of our classic “woopsies” this is just a minor hiccup I actually had not hear the term “integrated magnetics” so it had me read all the comments. The delay wont bother me as its going to take weeks to get to grips with linux CLI and QEMU (I might actually know what I’m doing by the time the RPI arrives) Keep up the good work.
Ouch, that’s not very good. I’m looking to get one as a ‘low end’ type server once it becomes easier to obtain in Canada.
No worries guys – a tiny error in the scheme of things.
I am sure the majority of people will be happy to wait excitedly for a little while longer.
Uk only, sorry.
Good find! As an IS Project Manager involved in procuring systems for my employer, I recognise a great deal of what they were saying :)
yeah I’d be interested in a broken board. Also has anyone, yet address the concept of case for the board?
Many thanks, I am excited!!!
Hope we’re ‘playing’ soon!!
Best regards & appreciation
How long would it take to get a Rasberry Pi device due to this hiccup?
which brings us an interesting question… Did you not prepare a bunch of test machines for your Chinese partners’ QA? Connect all wires, hit the blue button and wait 30 seconds for green light?
Raspberry Pi — Cambridge innovation yet again! Add this to the long list of achievements over the last 30+ years.
Well done Raspberry Pi team. OK, so some of us are a bit disappointed right now but at least we know what is going on and why instead of being kept in the dark. You can have it now or you can have it working — I know which I would choose. Can’t wait to get my hands on one.
Thanks for sharing all this info. The positive and informative responses from many experts has been well worth reading, I have learnt a lot. The shared info on this and the circuit board issue are something that may well save me making a related mistake in the future.
The gold plated one is obviously the correct one. Quality over cheap and cheerful — which would most likely be the correct one?
How many clients does this manufacturer have? How many of those clients have seen that the volume production run had an unannounced substitute AFTER the initial finished (and tested) QA supply. How many of those clients are now also worried about the quality of their products? Not a bright move by the plonker that made the substitute decision.
To be perfectly honest, I am disappointed but not surprised. Put me down for the first model B made in the UK, where there are more quality checks (and more enthusiasm for doing it correct than doing it cheap). Further justified by the fact that it was the UK Raspberry Pi team that spotted the mistake (after the factory supply QA failed to notice that the components to be fitted were not as specified by the designers). I would happily pay more for a Raspberry Pi made in the UK.
Some questions (sorry, not sure if they have already been answered):
When is the UK Government going to drop import tax on components for charitable foundations (and other justifiable reasons)?
Why didn’t ARM and Broadcom loan RPF the funds (at 0%) to get going? £100K (or more) each would be pocket-money and they probably spend more than this on customer samples and development support etc. The Raspberry Pi is the ultimate enthusiast demo board that is going to get everyone working on VideoCore and ARM11, so it would have been a sensible marketing move for both companies. I realise that a charity may not want to accept actual funding but surely a loan would not be breaking any rules or ethics. In the highly unlikely event that the Foundation would not be able to repay the loan, the loan could just be written off by companies of this size (put the expense down to R&D).
Why is it called Raspberry Pi and not Raspberry Py (since the name comes from Python)?
Of course, out of respect, Raspberry Pi should only be abbreviated when absolutely necessary but which is the correct abbreviation? Please could the RPF team state the official version (and advertise it on the home page etc.). The following seem to be popular but which gets the official vote?
RPi, R-Pi, RasPi, Raspi, …
I am eagerly waiting to see Raspberry Python’s Flying Gigapixels (a new Monty Python tribute show from the BBC).
I’m finally over not getting the very first batch of RasPi. Finally abled to order.
A classic case of bad design, especially in a product designed for education. Rule #1 is never use components that cannot be replaced with the simplest part available. The passives in the specified socket should have been included as discrete parts of the board. No brainer. Great shame.
Thanks for your vote of confidence. There is nothing wrong with the design. It works. Having the passives as discrete components would have increased the price of the board, which as you can imagine, given the amount of time REDUCING the cost of the board, wasn’t an acceptable option. So going for the chosen option WAS the no brainer route.
I am impressed with the way RS have kept me notified. I enjoyed the quiz “spot the magnetic socket”. Someone in the Customer Relations dept deserves promotion.
I’ll take a “straight jack” board too – we have plenty of Magjacks at work, I could fix it in 10 minutes ;^)
Perhaps you could, but unfortunately, the opportunity won’t be offered to you! All the incorrect jacks are being replaced – if they haven’t already been.
I have just gotten off the phone with Farnell. I understand the delay, but I had a delivery date of the 13th of March, and now they tell me it’s July!!
Please reassure me that this is just them being either cluless, or doing some serious ass-covering.
I understand that this is both a low-resource project, and subject to launch glitches, but July?? Seriously??
Initial batch 10k, initial demand >200k. The initial batch is late due to a manufacturing fault, subsequent batches also delayed to re-source faulty component. It may be quicker than July, but no promises.
Hi James… thanks for the reply.
I get all of that. However, my order was placed the morning of the release, and I had an email that morning from Farnell stating that stock had been allocated, and that a delivery date of the 13th was estimated.
I understand that the turn around has been pushed out, but by 4 months??
I am hopeful that this is just a case of Farnell having a blanket answer for all callers, since giving an accurate date would be impossible.
I’m afraid I don;t know! Fingers crossed for you also.
Hurry up will ya!
hehe jusk kidding, keep up the good work!
PS:Spectrums are better than Commadores
What? Burn him!
After working 4 decades as a developer in electronics I know such a situation very well. It had been not a smart idea of the manufacturer of the socket with magnetics to choose the same footprint as of this without. Interchangeability makes no sense, so this only causes errors.
Don’t worry for the delay, you have done good work.
I ordered from Farnell (on 29/2) and was given a delivery date in April. I was content with that. I got an email yesterday stating:
“A few days ago, the Raspberry Pi Foundation announced a small manufacturing delay. A component in the Pi is being reinstalled. In the grand scheme it’s small potatoes – but we know it’s frustrating for Farnell element14 fans and customers.
Congratulations – Farnell element14 will deliver your Pi to you around mid May. ”
I don’t understand – if Farnell are making from scratch why are they being delayed by this problem with the first batch?
on the pic which one is which
Late to the party…
Kindly pardon my ignorance, if that’s teh case, but are X-ray machines stanadrd kit in electronics operations? Or perhaps it was was this one of those magical moments of bindication when someone says, “You all laughed at me when I picked up this up at surplus. Go on, now: don one of these stylish lead aprons while I warm up the tube!”
Good question: I believe so, at least in Norcott’s sort of setup. It’s actually a microscope as well; you can also move whatever’s around in there in three dimensions too so you can check board layers, BGA soldering and so on. I’ll have to see if we can get some video of the thing working; it’s rather brilliant.
Wow. You must have been drinking like me to make sense of my misspellings. I really must go easy on the Guinness at lunch. But what’s a guy to do but drink while waiting for his Pi?
Thanks for the info. Video would be brilliant. Please post if they’ll allow it.
Wow! Honesty that’s something you don’t get from computer companies these days.
I would prefer to get my PI late and working rather than on time and not!
Keep up the good work.
In a world of shiny, pretty little gadgets that quickly lose the magic they never really had… you guys are bringing back the true geeky beauty of computing with the RaspberryPi … so you know whaT? all’s well that ends well.. great job!
Is there a chance that any of these slipped through the screening process? Are there any details on the identifying part numbers of the incorrect jack? I’ve spent a couple of days trying to bring up my new Rasberry Pi (ordered through Element 14) using the default debian image, but am not having any success. I can get the board to boot to lmogin, but neither ethernet, nor USB seems to work.
If ethernet and USB don’t work, then that isn’t an ethernet jack problem, but maybe something else. Is you power supply powerful enough? There have been reports that some lower powered PS’s cannot get the ethernet/USB going.
Try posting in the troubleshooting section of the forum. You will be more visibility there, and more help.