P10 Panels... Displays correctly BUT also has flickering pixels

[derektoews]

Alright, so at this point I gave up on the 5mm panels I had and I bought the 10mm Indoor panels from Ray Wu.

(https://www.aliexpress.com/store/product/P10-rgb-SMD-3in1-full-color-indoor-LED-Display-screen-unit-board-32-16pixels-320mm-160mm/701799_866772137.html)

Hooked them up and everything is almost perfect, but here's a video at what's going on...

https://www.youtube.com/watch?v=ubukU2EnegE

There's pixels lighting up in front of the one's that are supposed to light up... it's kind of odd and I'm not sure why that might be happening. I've double checked that it's the correct scan rate and correct setup on the FPP, I've restarted the BBB / Octoscroller a few times... Rebooted the computer, restarted Xlights, reconfigured it in both Xlights and in FPP. I tried running FPP in Bridge mode as well as a Standalone player just playing the .fseq and it does the same thing in both instances. It's most noticeable with just standstill text, but there's flickering with all effects and all colors as well.

Any ideas?

[dkuntz]

See this thread.
http://falconchristmas.com/forum/index.php?topic=7423.0
I have read elsewhere also that the problem may actually be that the pi 3 model b is actually too fast which is requiring the changes in the thread above to be made. If you have a pi 3 zero or a pi 2 you could try one of those to see if it fixes your issue. I was able to fix my issue by making the changes detailed the thread above.

[derektoews]

So I'm running a BBB with the latest Dan Kulp FPP version... so is the BBB possibly to fast?

From what I gather in the other thread, it's a known issue to have a FPP window open - windows are all closed and the issue is still happening.

There's a coding change that can be made in the PWM Nanosecond file... is this the change you were referring to that fixed your issue?

or

The other fix I saw was someone running a Beta FPP version...

[dkuntz]

Sorry, missed that you were using a BBB. I am not familiar with the BBB but I think the version you are using was created at least partly to address the issue you are seeing. Can't tell you how to make changes to fix your issue but I can tell you it looks exactly like mine did on a PI before I made changes to timings.

[JeffL98]

If it’s happening with an fseq playing, are you sure the starting channel in fpp matches what you have in xLights?

[derektoews]

Yep, definitely sure. I've even changed the starting channels a few times just to make sure it wasn't something weird going on.

Since I'm running the Dan Kulp latest version... I'm trying to do the solution I've seen on a couple forums where the change is made:

cd /opt/fpp/external/rpi-rgb-led-matrix/lib
sudo nano framebuffer.cc

modify the following line "static const long kBaseTimeNanos = 130;" and change 130 to 300
Save the file and exit

sudo make
cd /opt/fpp/src
sudo rm fppd
sudo make

However, that file does not seem to exist in the Dan Kulp version... should I revert back to the regular FPP version? or am I missing something here? I'm not very familiar with linux, but I login(fpp) pswd(falcon) and then type this in: cd /opt/fpp/external/rpi-rgb-led-matrix/lib      and then it says directory not found so I can't get to the framebuffer.cc file to change anything

[Gilrock]

I'm betting this turns out to be hardware and not a software issue.  That's why its always good to have 2 of everything.  If you had another Octoscroller I'd swap it out as a test first.

[derektoews]

I'll gladly order another Octoscroller to see if that's what it is. I just bought this one 3 weeks ago from pixel sequencing because that was the only place I could find one... I'll order another one from there unless anyone knows of another place that sells them...?

I guess this could be why my 5mm panels were messed up too, huh? Potentially

[Gilrock]

For the record I wasn't able to watch your video from where I'm at so I'm not seeing what the flickering looks like.  I just know there are high frequency signals going out all those Octoscroller pins and its easy for a solder joint to cause issues.  I bought mine before people were selling them so I ordered raw boards and parts and soldered up my own.  I had issue with one where I had to reflow all the solder joints and it fixed a problem.  You might just inspect it first under a magnifier and see if all the solder looks good.  Most of that board is mechanical so the chip that could go bad is that one buffer chip.

[derektoews]

Well Gilrock, you were right. I just got another Octoscroller in, hooked it up and the "ghosting" pixels or whatever you want to call it are gone. The P10 panels work perfect. Haven't tried the 5mm panels yet but at least I know it was a hardware issue.

Thanks for the advice!