I was just about to post on the same topic as Colin as I just found comment on it elsewhere. What I read was: (from here: http://www.raspberrypi.org/archives/781) (Ah, almost identical ref to Colin's) > It=92s inevitable, isn=92t it =96 you=92re freewheeling along perfectly happily and then you get a puncture. As you=92ll have noticed, there=92s 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=92d specified jacks with integrated magnetics in the BOM and schematics, the factory soldered in non-magnetic jacks. No magnetics means no network connection. We=92ve known about this for four days now, but we haven=92t 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. /> Hopefully the following will not be taken as racist, or anything other than the intended factual comment. It's an excellent lesson on how to do things right when manufacturing in China. Or not. While this is suggested as a simple mistake it is, alas, entirely typical of what is all too common in Chinese manufacturing. In this case it may have been a genuine mistake. You'd have hoped that the specification was as iron clad as it should have been for a product of this degree of complexity where one thing like this essentially cripples the device, and if this was a volume run (I think 100,000 was mentioned but even a 10,000 run would have warranted careful attention) you'd have hoped that there had been independent factory verification, either by an RP representative, or by a qualified independent inspector. The latter can be had at very reasonable prices. You can also pay more and get rubbish services, so some experience in who does a good job and who doesn't is needed. That none of these "you'd have hoped"s appear to have occurred suggests a lack of competence in manufacturing that I'd not have expected from people who it appears have strong connections with such a large manufacturer as Broadcom. That's not meant as slanging or personal criticism - more a statement of surprise that they seem to have been caught out by something which was not especially unexpected - not this specific problem but that there was every chance that there would be a problem of this sort if every possible effort was not made to prevent it. It at least suggests that the manufacturer was less than acceptably cabale - whether by choice or in a lack of incoming acceptance testing or vendor qualification or some mix. My number one rule for manufacturing in China is "You have to be there" - where "you" is you or a representative who is competent, authorised to act on your behalf and solely your person. They can be a Chinese national or whatever else suits - ethnicity or nationality is not at all what matters - it's being YOUR person on site that counts. FWIW, if anyone is interested I can recommend an Indonesian whose small inspection team does a good job in Chinese factory =A0inspections. FWIW my number 0 rules for manufacturing in China, formulated after seeing rule 1 repeatedly vindicated in practice is "Dont". That can be understaood as "If you are not willing or able to follow rule 1, don't even start". So far I have found almost no exceptions that suggest that Rule 1 is bad advice. Russell McMahon On 9 March 2012 13:11, cdb wrote: > =A0The foundation have just discovered the factory substituted the networ= k > jack from the specified phy inclusive to one without magenetics. =A0The > complete batch needs to be reworked and all three parties are trying to > find an immediate source of 10K jacks with integral gubbings. =A0Shipping= now > delayed by whatever time it takes. > > Reference plus some nice x-ray pictures > http://www.raspberrypi.org/archives/781#comments --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .