[Elphel-support] 353 Problem Mainboards

Fri Aug 17 12:39:50 PDT 2018

Thanks Oleg.

000E64081C8B
Here is the requested diagnostics:
https://docs.google.com/document/d/1WfGlWGk0A7zLgAl1EptMntTKoKLJLodron0P0zYcpB4/edit?usp=sharing

000E64081CA1
I will send that one back to you. Will wait until we know if 000E64081C8B
needs to go back as well or if I can resolve the issue here.

The one with the color tinting is behaving much better.

Thanks,

Linden Mills-Connery
Zone4 Team <https://zone4.ca/contact/>
205-820 Main Street
Canmore, AB, T1W2B7
403-401-7215

On 17 August 2018 at 12:50, Oleg <support-list at support.elphel.com> wrote:

> Hello Linden,
>
> 000E64081C8B
>> Boots up fine and I can access it on the network fine. The problem is
>> with the interface to the image sensor. All that is available on bimg is a
>> 1px x 1px white image. I have tried connecting this mainboard to a working
>> sensor and there was no change in behavior, still a 1x1 bimg. I have also
>> connected a working mainboard to this sensor and the sensor is working
>> fine. So this points to an issue with the sensor interface on the 000E64081C8B
>> mainboard.
>>
>
>  telnet or http://192.168.0.9/phpshell.php, then
> 1. ~$ dmesg
>
>> ...
>
> The sensor clock set to 20000000
>> removing MRST from the sensor
>> trying MT9P001
>> Found MT9P001 2592x1944 sensor, chip ID=1801
>> trying MT9X001
>> Resetting MT9X001 sensor
>> Reading sensor registers to the shadows:
>> Initializing MT9X001 registers with default values:
>> arch/cris/arch-v32/drivers/elphel/quantization_tables.c:174:init_qtable_head_cache
>>
>
>
> 2. ~$ fpcf -i2cr16 4800 100
> It will just read sensor's registers:
>
>> 004800: 1801   32    c  793  a23  1c2  284    2    0   15    0    0    0
>>    0    0    0
>> 004810:   50 6404    0    0   36   10    0    0    0    0    0    0    0
>>    0 4006    0
>> 004820:   40    0    0    0    2    0    0    b    0  480 1086   10   18
>>   16    f    0
>> 004830:    0    0    0    0    0    8    0    0    0    0    0    0 171d
>>    5 8041    4
>> 004840:    7    0    3    3  203 1010 1010 1010   10   a8   10   28   10
>> 120f 1010   14
>> 004850: 8000    7 8000    7   16    0   20    4 8000    7    4    1   5a
>> 2d13 41ff 231d
>> 004860:   13   13    0   1b   1b    0    0    0    0   cf   ef   ed   ea
>>   f1    0    0
>> 004870:   ac a700 a700  c00  600 5617 6b57 6b57 a500 ab00 a904 a700 a700
>> ff00 a900    0
>> 004880:   22 1f04    0 1b06 1d08    0 1806 1a08    0    0    0    0    0
>>    0    0    0
>> 004890:  7d0    0    1    0    0    0    0    0    0    0    0    0    0
>>    0    0    0
>> 0048a0:    0    0    0    0    0    0    0    0    0    0    0  794  a24
>> 2809   b3 49f5
>> 0048b0: d6f0 ba9c ace6 fcf4 d8b6 fe6e fa7e  f1f  4fd  ac1  dff  c87  fcf
>>  ff7  57d  dea
>> 0048c0:  cff  dfa  3ca  b7a  adb  e97  bfa dcef e7ff dffe 187f 4957 eef3
>> 7f79 6fad 37ff
>> 0048d0: 68bc 7fad fedd fd3e fabb f263 2f7a 95de 3fcb faff ff33 9e9f f779
>> cbf7 8b0d fe6d
>> 0048e0: d9a6 ff3a fd7a 855d e85f a9fc dd4f 6e7f 2ebf fe71 7fef fffe 5cb7
>> 71fd ecfc 11fd
>> 0048f0:    0    0    0    0    0    0    0    0    1    0 e0b0 48cf 8068
>> e118    0 1801
>
>
> 3. Check if write pointer is changing:
> http://192.168.0.9:8081/pointers
>
> 4. Check if frame number is increasing:
> http://192.168.0.9:8081/meta
>
> 5. http://192.168.0.9/parsedit.php?images=9:3:.2
>
> If 1 & 2 do not fail. Try removing autocampars.xml? Sometimes it gets
> corrupted for some reason.
>
> 000E64081CA1
>> Won't boot at all, no LEDs on Ethernet port or switch it is connected to.
>> I have done a visual inspection and cleaned the Ethernet port pins but no
>> luck. I have also connected a working mainboard to this sensor and the
>> sensor is working fine. So this points to an issue with the 000E64081CA1
>> mainboard.
>>
>
> Hard to say. Need to have a look. I would check the serial console output
> but it requires 10389.
>
> I also have another camera that is producing color tinted images:
>> https://drive.google.com/drive/folders/1bA6VOga8DPdiyhQXq7PWLPg5b7u-
>> 4L03?usp=sharing
>> I had a look and the ZIF connector between the mainboard and sensor *may*
>> have been loose. I am running some more tests on it but wondering if a
>> loose connected would result in those tinted sample images.
>>
>
> It could have, yes.
>
>
>> How should I proceed with the bad mainboards?
>>
>
> You can send them to us - we will inspect them and see what we can do.
>
> Best regards,
> Oleg Dzhimiev
> Electronics Engineer
> phone: +1 801 783 5555 x124
> Elphel, Inc.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://support.elphel.com/pipermail/support-list_support.elphel.com/attachments/20180817/e7e783cb/attachment.html>