Hi,
finally we got a time slot tomorrow to do the next
soldering tests for the GTA04A5. The reason why it
took so long is that these pick&place machines
sometimes run for several days with the same program
producing some 10 thousand PCBs for other customers...
So they can't easily stop and do experiments like
we need.
The plan for tomorrow is try to solder 4 PoP systems
with different preparation:
1. without (almost no) solder paste and no flux. The
idea is to see if the effect comes from too much flux
added or is inherent to the DM3730 package.
2. with solder paste in the outer ring of the DM3730
package and no paste in the middle (where the shorts
usually occur). The idea is to reduce the amount of
flux in the critical area.
3. with a tiny snippet of Kapton under the DM3730
and doing the same flux dipping as we tried before.
4. with a tiny snippet of Kapton under the DM3730
and full solder paste as one would usually do.
Puh. Quite costly to produce scrap since we will not
solder any other components and therefore we can
only do X-ray inspection and measure for short circuits.
There is no possibility to test if the DM3730 SoC would
work well and has really all pads connected.
Anyways I hope this will give a lot of insights and finally
will allow to decide how to do the real production
run. Probably we need to run 4 full boards and test
them before doing the remaining ones...
On the kernel side there is also some news.
Nothing for those readers who are strong advocates of
FLOSS and do never want to run any binary from unknown
sources. Those can skip this section...
The reason is that it is about the PVR/SGX530 3D GPU
inside the DM3730 chip.
The whole 3D system is made of four components, two
of them are binary blobs only, one is GPL open source
and the fourth unknown is the SGX hardware.
The news is about the open source kernel driver. I
have finally managed to get it compiled again on
letux-4.9-rc2. And in a way that it is a simple kernel
module. Well, if you want to call a 4 MByte .ko "simple".
modprobing on the GTA04 almost succeeds, except an
error message from the "reset framework". This was
expected, since it is one of two parts which need
adaptation to newer kernel APIs. One is the "reset
framework" and the other one is about dma/cache
interaction where the TI driver calls some very very
low level (even assembler) code directly. This has
been declared private API some time ago and so it
has to be rewritten to use some more official API.
If someone wants to dig into this topic, the
source code is here:
http://git.goldelico.com/?p=gta04-kernel.git;a=shortlog;h=refs/heads/work/hns/gpu/pvr-v2
Now back to the topic of FLOSS. This driver is of
course free. But it will need / load firmware binaries
into the GPU part of the SoC. These binaries are
on one hand some "microkernel" running on the SGX
GPU and the ARM libraries for OpenGL etc.
So thats it for today. And I hope I can bring some
news about the GTA04A5 tomorrow.
BR,
Nikolaus
_______________________________________________
Community mailing list
Community@openphoenux.org
http://lists.goldelico.com/mailman/listinfo.cgi/community
http://www.tinkerphones.org
