2 New Blog Posts: 98% Success Rate! + Lightpipes


2 boards may sound like a small amount, but 4,000 is 38 times greater of a number than 105.

(2 / 105) * 4000 = 76 faulty boards.

Either each board needs to be tested individually, the manufacturing processes need to be improved to reduce the failure rate, or are there are a sufficient number of spares to send a replacement unit to those affected?

I can't see how we can get this thing before Christmas. 5 weeks today! Even if they were being assembled as we speak...
 
MWeston said:
http://i198.photobucket.com/albums/aa99/MWeston_2007/Pandora_assembly.jpg

103 out of 105 passed. The other two may come up after another reflow but sometimes they are just dead and get tossed aside. This whole time has been spent working on the ones that have been functional and building the test jig with pins that probe the PCB. We've also been focusing efforts on adding more features to the software so that an automated application can be run on top of the OS to check all of the hardware functionality with as little human interaction as possible.
From this I take no board will go without passing this test.
 
Last edited by a moderator:
http://www.gp32x.de/board/index.php?/topic/50450-2-new-blog-posts-98-success-rate-lightpipes/page__view__findpost__p__774086__fromsearch__1

on normal stages (with no Q improvement or Q lowering), while on production they will be rounded near those numbers.

of course i also believe they wouldnt send out any board faulty within a preliminar test, i expect not all boards be 100% tested, but at least they will check the boards after firmware mounted.

it all depends of what range of safety OP Team wants to deliver over the produced boards,
since 100% cant be ever reached even if they did 100% testing over all the boards,
it also depends on how many boards they expect to be delivered back on warranty after x amount of time
 
Last edited by a moderator:
Not that it makes much difference to the argument, and pretty much off topic, but...

The Japanese don't generally write Right to Left.
Traditionally it is written vertically with the lines going right to left, but this is very much traditional, used only in books and old-fashioned signs and such.
In most modern contexts Japanese is written from Left to Right just the same as English.

There are of course languages like Arabic which are straight up RTL so the point is moot, but I just wanted to mention it. :p
 
To add a point on the board testing.
I would hope that, unlike the first 105 boards, that the mass production boards have the nubs included in the surface mount soldering process.
Testing the 105 boards before reheating them to add the nubs creates a possible positive bias in the test.
It is likely that one or more additional failures will be found in the 105 boards when tested after soldering the nubs in place.
Any time you heat up a multilayer, precision lead pitch board, you are going to create potential failures.

Does anyone know if the nubs will be attached to the mass production boards prior to the test rig testing mentioned above?
 
j.pickens said:
Does anyone know if the nubs will be attached to the mass production boards prior to the test rig testing mentioned above?
They've mentioned that there is a test to be done after everything is completely assembled. It's probably like "execute a whole bunch of stuff, fill ram with 1s and 0s, turn the screen different colours, activate the s-video, run the audio up and down a scale, install the firmware". So it hits all the important stuff so they'll know it won't fail outright. That'd be enough of a test for me, but others might disagree.
 
Last edited by a moderator:
WizardStan said:
j.pickens said:
Does anyone know if the nubs will be attached to the mass production boards prior to the test rig testing mentioned above?
They've mentioned that there is a test to be done after everything is completely assembled. It's probably like "execute a whole bunch of stuff, fill ram with 1s and 0s, turn the screen different colours, activate the s-video, run the audio up and down a scale, install the firmware". So it hits all the important stuff so they'll know it won't fail outright. That'd be enough of a test for me, but others might disagree.

That test is a pretty much full system test by the sounds of it and would be fine.
But it does sound like the step of testing the board should be done with the nubs on by what j.pickens is saying. Otherwise building the full machine just to find out soldering the numbs on caused a failure, seems like a waste of time. etc
 
Last edited by a moderator:
Na-Noo said:
WizardStan said:
j.pickens said:
Does anyone know if the nubs will be attached to the mass production boards prior to the test rig testing mentioned above?
They've mentioned that there is a test to be done after everything is completely assembled. It's probably like "execute a whole bunch of stuff, fill ram with 1s and 0s, turn the screen different colours, activate the s-video, run the audio up and down a scale, install the firmware". So it hits all the important stuff so they'll know it won't fail outright. That'd be enough of a test for me, but others might disagree.

That test is a pretty much full system test by the sounds of it and would be fine.
But it does sound like the step of testing the board should be done with the nubs on by what j.pickens is saying. Otherwise building the full machine just to find out soldering the numbs on caused a failure, seems like a waste of time. etc

You managed to miss the first line of his reply. From what I can tell from ED's/MWeston's posts, there seems to be a test pre-nubs and a test post-nubs.
 
Last edited by a moderator:
Vorporeal said:
Na-Noo said:
WizardStan said:
j.pickens said:
Does anyone know if the nubs will be attached to the mass production boards prior to the test rig testing mentioned above?
They've mentioned that there is a test to be done after everything is completely assembled. It's probably like "execute a whole bunch of stuff, fill ram with 1s and 0s, turn the screen different colours, activate the s-video, run the audio up and down a scale, install the firmware". So it hits all the important stuff so they'll know it won't fail outright. That'd be enough of a test for me, but others might disagree.

That test is a pretty much full system test by the sounds of it and would be fine.
But it does sound like the step of testing the board should be done with the nubs on by what j.pickens is saying. Otherwise building the full machine just to find out soldering the numbs on caused a failure, seems like a waste of time. etc

You managed to miss the first line of his reply. From what I can tell from ED's/MWeston's posts, there seems to be a test pre-nubs and a test post-nubs.

No I never missed that and even mentioned it unless he's talking about a different test. I thought he was talking about the one that TI helped with.

Edit:
Currently.
1 Board test (possibly pre nub)
2 Assembled system test (post nub)

pickens suggestion
1 Board tested with nubs soldered on.
2 Assembled system test

Edit Edit: To clarify my first edit lol
 
Last edited by a moderator:
I guess my question about nub installation should be more like:
Is there any reason not to attach the nubs at the time everything else is soldered to the boards?
It would seem that a process to attach the nubs separately from the big multicomponent populating event would entail lots of extra time and effort.
That's all I'm asking.
Thanks,
j.
 
I think the nubs are soldered differently so that they don't melt. They are made of plastic after all.
 
mali said:
I think the nubs are soldered differently so that they don't melt. They are made of plastic after all.
they might be, but the audio jack and volume wheel is also plastic
 
Last edited by a moderator:
I'd expect (hope) the analog nubs are a much softer plastic than the cheap plastic used for the volume wheel.
 
Gary13579 said:
I'd expect (hope) the analog nubs are a much softer plastic than the cheap plastic used for the volume wheel.
Because your tender thumbs are abraded so by that cruelly ridged volume wheel of devilishly unyielding composition?
 
Last edited by a moderator:
As long as the sort of plastics is not known it's just speculation. Most plastics wouldn't survive a reflow oven, some would.
 
Back
Top