[Gta04-owner] GTA04A5: good news and bad news
H. Nikolaus Schaller
hns at goldelico.com
Tue Feb 28 08:53:41 CET 2017
> Am 28.02.2017 um 08:24 schrieb tgrauss <tgrauss at f-gp.eu>:
> If you are confident that all of them can be repaired,
unfortunately I am only confident that *some* can be repaired.
At best 50% of the damaged ones.
> maybe we can then manufacture the other boards without the CPU/RAM and solder these 2 chips with the repair procedure? This way we don't loose more components and limit the production problem.
No. The repair procedure is not better in results than trying initially. Maybe even worse but I have no statistics.
How much improvement can be done is just pure statistics...
If soldering (independently of method) has yield of 30%, removing defective chips and trying again
30% + 60%*30% = 48%
If we try again, we get 30%*(1-48%) more = 63,3%
So we get good boards:
48% trying repair once
63,3% trying repair twice
74,31% trying repair three times
But this does not take into account that boards may become damaged beyond repair.
And that trying to repair three times needs the chips planned for 4 boards.
Since each repair cycle takes est. 1 hour per board, we have nobody who has that much spare time.
> Best regards
> Thierry GRAUSS
> On 27/02/2017 19:30, H. Nikolaus Schaller wrote:
>> Hi Marek,
>>> Am 27.02.2017 um 11:46 schrieb Belisko Marek <marek.belisko at gmail.com>:
>>>> -1.: Only 12 out of 36 (*) GTA04A5 boards we
>>>> have produced, are booting to Linux. The others either have a
>>>> short circuit under the OMAP (which is the general PoP soldering
>>>> problem we already know since the first generations of the
>>>> GTA04) or the OMAP shows no response or does not recognise the
>>>> µSD card for unknown reasons.
>>> Hmm this is bad. Doesn't something like that happen in first verification run? But maybe I'm wrong.
>>> Did you try to get some info from chip producer what could be a problem or which technology should be used to have it solder properly.
>> Yes, we asked TI in details some years ago when we built the GTA04A3 and A4.
>> From that we probably know everything what can be known about this
>> DM3730 soldering issue...
>> And we though that this knowledge is enough to avoid to fall back into the same
>> old problems we had back then when using a new (and modern) production line.
>> Back then we had to experiment several months to raise yield from ca.
>> 30% to 90%.
>> Then, the production batch was completed. But we had to do a lot of repairs back
>> then as well.
>> So all this means the production process of the DM3730 + PoP is very
>> very critical
>> to parameters we can't repeat several years later. And therefore we had to start
>> optimization almost from scratch. But do not have the means to continue to reach
>> 90% again.
>>> 12 from 36 is exactly third which is too high. I'm sorry to hear such bad news.
>> I am currently making a plan in which sequence we try to repair what
>> (desolder chips
>> and replace). So I am confident that we can come a little closer to 36 of 36.
>> Gta04-owner mailing list
>> Gta04-owner at goldelico.com
> Gta04-owner mailing list
> Gta04-owner at goldelico.com
More information about the Gta04-owner