Board versions/revisions, or different assembly lines?


68 posts   Page 2 of 3   1, 2, 3
by RaTTuS » Fri Jun 29, 2012 8:36 am
using a hand lens here are what I have
1st one - this came from farnell
FN120402384
RG2 is blank
C2 is n2 220c UD
RAM is hnmix h9tknnn2c0mp landm 1494
Lan is SMSC lan9512-jzx b1134 a1b17 8r14 77698 sta-sg

2nd one from RS
E2112rs v1.0b1.1
RG2 is Se8117t33 12-13 cf
C2 is 2e 220 16h
RAM is hnmix h9trnnn2g0mp lrndm 217a
lan is SMSC lan9512-jzy b1215-a1b17 ase-tw

both are pretty stable
well - the farnell one has randomly crashed and has not managed to stay up for more than 5 days - however I have been running a lot of stuff on it and experimenting with cases etc - we'll see
the RS one had crashed once but this could of been the position where it is .

both had been running of the same system of power i.e.
http://www.amazon.co.uk/gp/product/B005NETBQM

though now I'm powering the RS one via a USB from
http://www.amazon.co.uk/gp/product/B0002AFZVM
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 5233
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by marcoalexcampos » Fri Jun 29, 2012 8:40 am
therealeasterbunny wrote:[...] what setting to I put my digital multimeter on for ohms and what do I put my red/blue pokey sticks on in terms of Pi bits to give you the readings to update the Wiki for B1213? Happy to help if I know what to set / poke. Can anyone point me in the right setting?


Place on the minimum value of the "omega" symbol (resistance), and it shouldn't make a difference the polarization (red vs back cable).
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by marcoalexcampos » Fri Jun 29, 2012 1:37 pm
@RaTTuS: I've just updated the table with your boards.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by marcoalexcampos » Fri Jun 29, 2012 1:40 pm
So far we just seen model B boards. Does anyone has model A boards? :)
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by RaTTuS » Fri Jun 29, 2012 1:43 pm
marcoalexcampos wrote:So far we just seen model B boards. Does anyone has model A boards? :)

there are no model A boards made yet
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 5233
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by marcoalexcampos » Fri Jun 29, 2012 2:47 pm
Hi everyone!

Thank you for your feedback, and continue to have those photos and infos comming. :D

Since the list has grown quite a bit, I've moved it to it's own page here: http://elinux.org/RaspberryPi_Boards
On the page "R-Pi_Troubleshooting" I've reduced the section to it's base info and created a link to the new page, so it doesn't get cluttered.

Please keep your feedback coming in, and feel free to add info yourselves (if you don't know how, email me and I'll post for you).

Thank you. :)
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by RaTTuS » Mon Jul 02, 2012 8:48 am
got a 3rd one
from farnell
FN 120604378
C3 220 16V EE5
RG2 Se8 117t33 1213-lf
RAM Samsung 216 k4p 26324ED agc1
NIC SMSC lan9512-jzx b1212 a1b17 br149884b cti-tw

for some reason I cannot login to edit the wiki - my end fault
so posting here [thus bumping thread] ;-)
can you change my previous RS one from pretty stable to stable good runner overclocked to 900mhz
leave the existing farnell as pretty stable - It's not crashed in 3days now
this current one has not been stress tested yet

can you put a link to this thread on the wiki aslo
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 5233
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by marcoalexcampos » Tue Jul 03, 2012 10:08 am
RaTTuS wrote:got a 3rd one

Just added. Can you send some clear pictures to post?
Also made the changes you suggested.

If anyone has items that think might be relevant to put on the table (rows), please feel free to debate them here.
Also don't forget to give feedback on your boards, even the ones already on the wiki page.

Cheers.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by marcoalexcampos » Tue Jul 03, 2012 11:37 am
antijn wrote:I got two cards showing the same problems. One from RS, build week 1218 which has stability problems and one from Farnell, also build week 1218, but works in the same setup.


Can you send some pictures and other data so I can add your cards to the wiki?

Thanks.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by AndrewS » Tue Jul 03, 2012 1:40 pm
See this thread also: viewtopic.php?f=63&t=5889
and there's lots of pictures of different peoples' boards in this thread viewtopic.php?f=63&t=5559

Might be worth adding the long number (I have no idea what it indicates) from the white sticker to the wiki-table too?

Edit: And I suspect the "b1134" etc. marking on the USB chip may also be a build-week, i.e. the 34th week of 2011 ?
User avatar
Posts: 3626
Joined: Sun Apr 22, 2012 4:50 pm
Location: Cambridge, UK
by marcoalexcampos » Tue Jul 03, 2012 2:38 pm
AndrewS wrote:See this thread also:

I'll have a look. Thank you!

AndrewS wrote:Might be worth adding the long number (I have no idea what it indicates) from the white sticker to the wiki-table too?

I think that's the serial number, where the Fernell boards start with "FN" and the RS have a quite different nomenclature.

AndrewS wrote:Edit: And I suspect the "b1134" etc. marking on the USB chip may also be a build-week, i.e. the 34th week of 2011 ?

Well pointed. This makes even more important to have that info on the table since that's the chip that has been associated to more problems.

Thank you for your input.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by marcoalexcampos » Wed Jul 04, 2012 4:44 pm
18 boards have already been entered!
Keeping this rate we should be able to have a good base to work with in a week or so.

Remember to update your feedback (important) and any missing information, so we can correlate possible problems with board versions.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by RaTTuS » Thu Jul 05, 2012 10:39 am
My phone has just deid so I cannot do any more images ;-(
it's been one of those weeks
I've also lost the charger for the camera ;_(
I'll update when I can
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 5233
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by beatmag » Thu Jul 05, 2012 10:47 am
this is interesting.

i was curious to understand what kind of testing gets performed at the factory. because surely a few people have got pi's that are faulty. and has the fault occurred during shipping? the fault occurred by itself? or straight from the factory its faulty...?

currently it seems most of the pi's are stable, but a few people have been getting faulty pi's.
can we say RS are better? or Farnell are better?
Posts: 25
Joined: Sun Apr 22, 2012 5:28 am
by Montala » Thu Jul 05, 2012 11:06 am
My board from Farnell/Element14 which I received on 12th June has a serial number starting with FN1205 and has a Samsung memory chip.

So far it seems stable on 'wheezy' with about 4.89v across TP1 and TP2.

I am expecting to receive a board from RS next week so will be interested in doing a comparison.
User avatar
Posts: 638
Joined: Mon Mar 05, 2012 11:54 pm
Location: Herefordshire (U.K.)
by Mortimer » Thu Jul 05, 2012 11:19 am
The white stickers on mine only differ by one digit. I don't think this is a serial number:

RPi1 : E2112RSV1.0B1.1
RPi2 : E2212RSV1.0B1.1

I suspect that these stickers are added when prior to shipping to the retailer in this case RS.

I reckon the first two digits after the 'E' are week numbers. So my boards were on the shelf at the factory for 3 and 4 weeks respectively prior to being packaged up for RS.
User avatar
Posts: 715
Joined: Sun Jun 10, 2012 3:57 pm
by marcoalexcampos » Thu Jul 05, 2012 11:23 am
@RaTTuS: Sorry to "ear" that. It looks that your Friday 13th came a bit earlier. :(

@beatmag: Both RS and Farnell seem to have a good service (from my personal experience so far), and both have shipped one or another faulty card along side thousand of good ones.
I believe that the assembly lines, or the manufacturing process on each line when using slightly different components, may account for those best or worst experience in some cards.
The purpose of this thread is exactly gather evidence and make it public, to help both the engineers behind the project and us users, understand and work-around/fix these issues, and eventually avoid them in new batches. :)

@Montala/beatmag: Care to join the wiki page with the data, feedback and pictures from your board(s) :D
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by marcoalexcampos » Thu Jul 05, 2012 11:27 am
@Mortimer: You might be right there...

What do you all think, should I add a row for that sticker to visualize better?
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by Mortimer » Thu Jul 05, 2012 11:33 am
The stickers appear to have version information, which may not capture minor differences that have a greater impact for users. No harm in having it, but as my stickers are the effectively the same, and don't indicate that my boards have clearly come from different production lines, I am not too sure how useful the information will be at this stage.
User avatar
Posts: 715
Joined: Sun Jun 10, 2012 3:57 pm
by Montala » Thu Jul 05, 2012 11:55 am
Is the 'Board Version' the number on the white sticky label, or is it screen printed somewhere on the board itself, although I do accept that it might not always be in the same place, depending on the distributor/manufacturer?
User avatar
Posts: 638
Joined: Mon Mar 05, 2012 11:54 pm
Location: Herefordshire (U.K.)
by marcoalexcampos » Thu Jul 05, 2012 1:15 pm
The row "Board ver." has a reference that we made up based on the board model (B), the type of RAM (Samsung or Hinyx), and the manufacturer date that is usually printed on the board like 1218 in calculator like characters.

The sticker that Mortimer was talking about we don't know if it is some sort of serial number (at least for Farnell), or shipping reference (from RPi to the distributors).

What we know is that the boards sold by Farnell seem to be like "FNxxxxxxxxx" (except for one older board so far) and the RS use something like "Exx12RSV1.0B1.1"

I think I'll email them both and ask for the nomenclature of those stickers. Probably I won't have much luck, but it doesn't hurt to ask. :)
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by Mortimer » Thu Jul 05, 2012 1:26 pm
I reckon the Farnell sticker is a serial number starting with the date the board is assign to a retailer. So for FN120402384, this could be April 2012, board number 2384. This would fit with the RS sticker scenario as the example I have chosen has a manufacture date code, 1213, which is the last week of March, so a delay would put it into April.

The RS stickers are not serialised, and not unique, there are a couple of boards in wiki list wih exactly the same stickers as mine.
User avatar
Posts: 715
Joined: Sun Jun 10, 2012 3:57 pm
by marcoalexcampos » Thu Jul 05, 2012 3:53 pm
So far this is the information I have, though not confirmed:

[...]the FN is to distinquish the ones made under our contract as opposed the the other source of them, RS.

YYMM seem to be the next but the last would only limit us to having 100K a month - which would make sense given how fast we can get them into and out of the warehouse.
User avatar
Posts: 37
Joined: Fri Jun 22, 2012 2:45 pm
by Montala » Thu Jul 05, 2012 5:49 pm
Surely it is YYWW, being the year followed by the week (rather than the month number) as I don't think we have 18 or more months in a year... yet! ;)
User avatar
Posts: 638
Joined: Mon Mar 05, 2012 11:54 pm
Location: Herefordshire (U.K.)
by Mortimer » Thu Jul 05, 2012 6:35 pm
No it can't be, if you look at my real example.
Board code: 1213
Sticker code: 1204

So the board code is YYWW and sticker YYMM.
User avatar
Posts: 715
Joined: Sun Jun 10, 2012 3:57 pm