Post by thread:[OT]Raspberry Pi has an isolated hiccup or lack th

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’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. /> 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 inspections. 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 <spam_OUTcolinTakeThisOuTbtech-online.co.uk> wrote: > The foundation have just discovered the factory substituted the network > jack from the specified phy inclusive to one without magenetics. The > complete batch needs to be reworked and all three parties are trying to > find an immediate source of 10K jacks with integral gubbings. Shipping now > delayed by whatever time it takes. > > Reference plus some nice x-ray pictures > http://www.raspberrypi.org/archives/781#comments

Yes... even a basic 1st article check would have caught this. Unless the mfgr would do it correctly for the 1st one or 100 and then switched parts, which would be clearly fraud. I would never ever ever do mfgr like this without 1st article checks even at the place in San Jose I've used for years. A breathtaking what were they thinking fail on Rasp Pi's part. J RussellMc wrote: {Quote hidden}> 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’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. > />

On Thu, 08 Mar 2012 17:30:45 -0800, Jesse Lackey wrote: :: Yes... even a basic 1st article check would have caught this I used to work for a manufacturer of some well heard of medical devices. Whilst each production run required the checking off of components against the customers' BOM (the customer normally supplied the parts), if components had been mixed up or the production was under time pressure, then the production manager would insist everything went ahead, knowing that authorisation for substitution or procedure would add days to production start and of course cost money, so it would go ahead. Two weeks later the customer having performed compliance checks on delivery would send all boards back to be scrapped and to be started again. Things were only done correctly when the supplier made an audit visit which of course was known about a week in advance. My point being, this sort if things happen in non Asian countries and oversight at a distance (this case NSW to Queensland) by a stockmarket listed medical equipment company. Now how many Pi's does it take to suffer indigestion? Answer 10,000 Colin -- cdb, .....colinKILLspam@spam@btech-online.co.uk on 9/03/2012 Web presence: http://www.btech-online.co.uk Hosted by: http://www.justhost.com.au This email is to be considered private if addressed to a named individual or Personnel Department, and public if addressed to a blog, forum or news article.

> Unless the mfgr would do it correctly for the 1st one or 100 and then switched > parts, which would be clearly fraud. For 10,000 run the jacks will come in trays which may come in boxes. Expect 100's per outer carton so not too many cartons. Putiing best face on it a distributor may have mixed similar looking/names/located parts. Or may have not had stock of all required and somebody helpfully substituted an "identical" item. Fraud or lack of knowledge or accident could all be explanations. A few years back there was a major 'lead paint on toys' furore n the US. Major waves among major sellers. I subsequently visited the large (to me) factory complex where the problem occurred. A Canadian consultant I was working with who had significant involvement with the companies concerned (but no involvement in the lead paint debacle) told me how it had happened as he understood and believed it. If what he had been told as true, and it may have been, it was one person being helpful and creative that innocently caused the whole mess. Can happen. I do not say WHY you have to be there - just that you must :-). Russel

> I used to work for a manufacturer of some well heard of medical devices. > Whilst each production run required the checking off of components against the > customers' BOM (the customer normally supplied the parts), if components had been > mixed up or the production was under time pressure, then the production manager > would insist everything went ahead, knowing that authorisation for substitution or > procedure would add days to production start and of course cost money, so it would > go ahead. Two weeks later the customer having performed compliance checks on > delivery would send all boards back to be scrapped and to be started again. Yep, the old 'we don't have time to do it properly the first time, but we do have time to do it again' ... Seen it happen rather too often ... -- Scanned by iCritical.

On Mon, 12 Mar 2012 10:51:55 +0000, .....alan.b.pearceKILLspam.....stfc.ac.uk wrote: > > I used to work for a manufacturer of some well heard of medical devices.. > > Whilst each production run required the checking off of components against the > > customers' BOM (the customer normally supplied the parts), if components had been > > mixed up or the production was under time pressure, then the production manager > > would insist everything went ahead, knowing that authorisation for substitution or > > procedure would add days to production start and of course cost money, so it would > > go ahead. Two weeks later the customer having performed compliance checks on > > delivery would send all boards back to be scrapped and to be started again. > > Yep, the old 'we don't have time to do it properly the first time, but we do have time to do it again' ... ....which is a close relative of "I don't want it good, I want it Tuesday!" :-) Why is it that so many people in a position to control the way things are done, seem to excel at Missing The Point? Cheers, Howard Winter St.Albans, England