Very bright view

[GoogleCodeExporter]

Dears,
I've got an issue when I use Mavlink-OSD board. View is too bright/white. I've 
removed OSD board and view is ok. Please see attachment- print screen of view 
with OSD. If target is black then it looks almost OK but in case of brighter 
backgroud everything is white.  Could you please help me?

Regards,
Darek

Original issue reported on code.google.com by dario...@gmail.com on 18 Feb 2015 at 3:30

Attachments:

• 20150218_161232.jpg

[GoogleCodeExporter]

I also have this problem

Original comment by dannie.s...@gmail.com on 3 Mar 2015 at 6:25

[GoogleCodeExporter]

any solution? , I have the same problem too.

Original comment by da...@dtatecnologia.com.br on 27 Mar 2015 at 11:21

[GoogleCodeExporter]

I have some ideas, seems like initialisation of black level has been forgotten. 
But in some cases it's working even without initialisation. I'll try to insert 
it into init and will let you know if it helps.

Original comment by elf...@gmail.com on 29 Mar 2015 at 2:24

[GoogleCodeExporter]

for those who think they connected something wrong when OSD is too bright...

I have 2 units, 
first from rctimer 2 years old, works ok,
but I got a new, unused one in a bundle(apm 2.6, gps, osd, amp/volt sensor... 
around 110$)
and OSD is acting strange, too bright,

I plugged in both in my plane and recorded it,
so that's my solution for now, glad I had the old one laying around,
hope the fix will come up and make all units working(newer one from ebay seem 
affected :( )

video: https://www.youtube.com/watch?v=KFpBk_VeUvA

Original comment by T.Resko...@gmail.com on 29 Mar 2015 at 11:23

[GoogleCodeExporter]

Ok, After very long investigation I've figured out this issue has nothing to do 
with firmware.
There is electrical issue with the board. I'm not sure what exactly it is, but 
something is definitely wrong in feedback circuit of onboard voltage regulator. 
AVCC is not stable and jump around 3.2V while according Max7456's datasheet it 
suppose to be at least 4.75V
Rising AVCC to 5V is fixing this issue.
My best guess, someone sold a batch of faulty OSD, which failed to pass QC.

Here is a sample of the signal. Bottom is input. Top signal's amplitude is very 
low. It's about 0.1-0.2V, while it suppose to be 1V, so it's x5 times scaled 
up. All levels are wrong. Well, IC is dramatically underpowered - no surprise.

Original comment by elf...@gmail.com on 7 May 2015 at 7:38

Attachments:

• IMG_20150507_145732.jpg

[GoogleCodeExporter]

Yep. Resistors in feedback circuit have wrong values. Voltage regulator is 
holding 0.83V and they've put 30K and 10K resistors making 1/4 divider. That 
give's 3.2V on regulated output. Pretty standard for common 3.3V electronic, 
but wrong in this case.
This is definitely non-single-board mistake, so I'm thinking they've made 
couple thousand boards like this and they're on the market.

Original comment by elf...@gmail.com on 8 May 2015 at 5:27

[GoogleCodeExporter]

So, to FIX IT, you have following options:
 1) Make a traditional 5V-only hack, when you power the board with 5V from digital side. There are tutorials and how-to all over the internet.
 2) Change resistors to get proper 1/6 division. 

I would recommend to go with the first option for anybody who has no SMD 
soldering skills, cause changing 0604 size components without experience will 
not be a picnic. However, the easiest way is to stack second  15K resistor on 
top of 10K (it's in the very corner of the board). Overall resistance will 
become 6K. Alternatively, you can replace 30K with 50K. 

Original comment by elf...@gmail.com on 8 May 2015 at 6:48

Attachments:

• IMG_20150508_144329.jpg