Author Topic: Strange output on my P10 panels  (Read 501 times)

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Strange output on my P10 panels
« on: November 21, 2018, 06:22:28 PM »
I will preface this with the fact that I am a noob to P10 and FPP. My setup is as follows:
3x3 P10 1/4 scan panels (P10(3535)16x32-4S-M2.1)
Rasberry Pi 3B
PiHat from CFOL
FPP 2.4


Pictures of panel and FPP setup attached.


Here is my dilemma, when I run test from FPP the output does not seem to be what I would expect. An example is when I run the RGB chase, I would expect to see continuous vertical lines of RGB, what I get is every fourth column, it becomes a vertical line but those prior almost seem like they are driving more than the number of pixels that they should and they become scrambled. So, I tried just doing an output of 64 channels and I get two blocks of RGB output (see picture) and the block isn't even a full block. I would expect that if I ran a test using 96 channels, I would get the first line of pixels across the bottom displaying an RGB chase. When I do the same test in a sequence sent from Vixen, I get the exact same result so that is why I am suspecting that this is an issue in my FPP configuration.


I have also read something about needing to configure what scan the panels are as there are timing issues and I cannot seem to find where that setting is.

Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,137
  • Kudos: 119
    • Granbury Christmas Lights
Re: Strange output on my P10 panels (1/4 scan)
« Reply #1 on: November 21, 2018, 06:45:44 PM »
I see from your text and pictures that your P10's are 1/4 scan instead of the much more common 1/8 scan.
Are they the waterproof models?

I am not positive that the PiMatrix supports the 1/4 scan.

It is specifically called out for use on a BeagleBone with Octoscroller, though.

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Re: Strange output on my P10 panels
« Reply #2 on: November 21, 2018, 06:55:31 PM »
Based on the circuit board on the back, I am going to say that the panels are not waterproof. Could the 1/4 scan be the root of why things just don't seem to be addressed/displayed correctly?


If I were to go with the BeagleBone Black and Octoscroller, would the BeagleBone Black Wireless work or do I need to get a normal BeagleBone Black and some type of wireless USB dongle? Also, I ordered a P10 Scroller from DIY LED Express, is that the same as the Octoscroller you mention?


I have scratched my head over this for the past 3 days, scoured the forums and have seen mention about a setting for the scan rate. I am going to guess that the setting is on the BBB install of FPP and not the Pi install.

Online dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,430
  • Kudos: 78
Re: Strange output on my P10 panels
« Reply #3 on: November 21, 2018, 07:01:53 PM »



The BeagleBone Black Wireless should work.  The BeagleBone Green Wireless will not.  Avoid that.   \


The P10 scroller is an octoscroller. 



Dan Kulp

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Re: Strange output on my P10 panels
« Reply #4 on: November 21, 2018, 07:23:18 PM »
Ok, so I have ordered a BBB Wireless from Amazon and already ordered the P10 Scroller from DIYLEDExpress. Hopefully this will do the trick.


Am I correct then in assuming that the most likely issue is the 1/4 Scan panels vs. the standard which is 1/8 scan? I would further assume it has to do with the refresh rate to the board and/or how the data is sent to a 1/4 vs. an 1/8? Since I am learning and somebody else might learn from this, are there any other Gotcha's that I should be aware of? I didn't even realize that there were 1/4 scan, 1/8 scan, even 1/16 scan panels until I went down a deep and dark rabbit hole in one of the many forums I was searching through. The only thing anybody said was that the 1/4 scan boards are a lot brighter (which these are at 100%, in the picture they are at 25% to help save my retinas)


Maybe I should start writing a noobs point of view document for a wiki somewhere.

Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,137
  • Kudos: 119
    • Granbury Christmas Lights
Re: Strange output on my P10 panels
« Reply #5 on: November 22, 2018, 07:45:01 AM »
Dan is more knowledgeable about this (since he's done most of the improvements to the BBB LED Panel Support) but here goes. (and Dan will correct anything I get wrong. :) )

1/8 Scan - since the original and most P10 panels sold are 1/8 scan, I'll start here. There are 16 "rows" of LEDs, and 32 LEDs per row.
So with 1/8 scan, it means on each Refresh cycle, only 1/8 of the rows are lit. so it takes 8 cycles to light the panel(s) but the cycle is fast enough that you don't see it. One thing it means is lower power usage (fewer LEDs lit at one time) but also lower total brightness.

1/4 Scan - Now it takes 4 cycles for a whole panel. Makes it brighter, less potential flicker and a little higher power consumption

See the attached picture for all the currently supported Panels supported by v2.4.2 FPP and the Beaglebone

Now, this may get me some "hate mail" but I don't understand why people continue to use the Pi and PiMatrix combination to control LED Panels.
I still have one panel driven by a Pi 2 and a Pi Matrix hat. It is on the "upgrade" list, though. :)


Online dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,430
  • Kudos: 78
Re: Strange output on my P10 panels
« Reply #6 on: November 22, 2018, 07:55:03 AM »



Close enough.   :)


More technical details:   The Hub75 interface (the ribbon cable) has 2 sets of RGB data lines (the rest are ground and control).   Thus, it can only clock out data for 2 rows at once.   For 16 row, 1/8 scan, that's perfect.   It does rows 1 and 9, then 2 and 10, etc.. Likewise for the 32row 1/16 scan.   These are the simple case.   It clocks out the row, turns it on, goes to the next.


The other scans are more complicated.  To do 4 rows at once over only two sets of data lines, it has to multiplex two rows of data together.  Unfortunately, there are SEVERAL different ways to do it.  The original 1/4 scan panel that we saw years ago would clock out all 32 pixels for rows 1/9 and then all 32 for 5/13.  However, newer panels have done things differently.   Some do 16 pixels at a time, others do 8.   Some that do 8 do the second row in a backwords direction, etc... Basically, whomever laid out the circuit board was free to do whatever they want figuring it's the software persons job to get it to work. :(


Another interesting note about the 1/4 scan (and 1/2 scan):  there are 5 lines in the ribbon for the row address.   With 1/4 scan and 1/2 scan, there are two ways to use them.  The traditional "each is a bit so you can represent numbers 0-31" which is mostly used.  The other option is each line represents a row.   Want to address row 1, toggle that line.   Etc... Again, it becomes a software persons problem.




Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,137
  • Kudos: 119
    • Granbury Christmas Lights
Re: Strange output on my P10 panels
« Reply #7 on: November 22, 2018, 07:57:52 AM »
See, I told you! And it doesn't even hurt my feelings, much. Have a great Thanksgiving.

Don't you have a Turkey to cook or something?

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Re: Strange output on my P10 panels
« Reply #8 on: November 28, 2018, 04:43:47 PM »
Ok, so I now have a BBBW and an Octoscroller. It is hooked up and operating, but it still has issues. i don't have the flicker and the other issues but I now have a new issue. I have not moved the panels around nor have I changed out the cables on the initial test and what I am getting is that the second row of panels are not displaying red or green, only blue and white. See attached images.


Setup
BBBW with DIYLEDExpress Octoscroller
FPP v2.5 BTRFS



Troubleshooting steps:
1) Tried new cables from output 2 to that row of panels.
2) Tried output to just one panel.
3) Tried steps 1 & 2 on a different row. Same results on that row
4) Tried a different output from the Octoscroller.
5) Tried the output from step 4 on a different row. Same results on that row
6) Tried 16X32 1/8 scan.
7) Tried a shot of whiskey.


I would try other panels, but I only have these 9.


I have only changed the date, network, and Output Settings > LED Panel from the default install.


It is probably a NOOB issue.
I am very frustrated with this and could really use some help. Help me Obi Wan...
« Last Edit: November 28, 2018, 04:49:34 PM by dkulp »

Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,137
  • Kudos: 119
    • Granbury Christmas Lights
Re: Strange output on my P10 panels
« Reply #9 on: November 28, 2018, 05:18:37 PM »
Can you add pictures of the back of the panels showing HUB75 cables and Octoscroller?

Your LED Panel page looks EXACTLY like mine except for the 1/4 Scan.

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Re: Strange output on my P10 panels
« Reply #10 on: November 28, 2018, 06:11:48 PM »
Attached...

Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,137
  • Kudos: 119
    • Granbury Christmas Lights
Re: Strange output on my P10 panels
« Reply #11 on: November 28, 2018, 06:38:09 PM »
I see nothing wrong with your setup except your ribbon cables are very long. Do you know how to carefully undo one and shorten the ribbon? The shorter the better.

Offline CaptainMurdoch

  • Administrator
  • *****
  • Join Date: Sep 2013
  • Location: Washington
  • Posts: 9,748
  • Kudos: 205
Re: Strange output on my P10 panels
« Reply #12 on: November 28, 2018, 07:02:12 PM »
Did you try the other 1/4 setting?  From the RGB sample pic you posted, it looks like your panels are one of those oddball ones where they break the panel up into regions and send the data in a non-contiguous pattern.  The colors in some regions are reversed as if their data was out of order or reversed.

For the issue with some rows not lighting up, if you tried that row of panels on a different output and different cable, you may want to try swapping the order of the panels in the row.  It could be that the first panel is bad and therefore the second and third on the row aren't getting good data.
-
Chris

Offline mwestling

  • Newbie
  • *
  • Join Date: Sep 2016
  • Location:
  • Posts: 12
  • Kudos: 0
Re: Strange output on my P10 panels
« Reply #13 on: November 28, 2018, 07:04:45 PM »
Ok, I have tried more troubleshooting steps including swapping panels around and putting them into the other two rows, etc. with no luck. I tried daisy chaining the middle row off of the first output so O1 had 6 panels and O2 had three. Same result with row 2. I only have one BBBW so I cannot swap that out, I have swapped outputs and Octoscrollers and that didn't work. I am at my wits end with this thing. I don't know what to do other than maybe try an earlier revision of FPP or something. Cables are not the issue, Panels are not the issue, Octoscrollers are not the issue. The only two options are my BBBW and my setup in FPP.


Captain Murdoch, I have tried the 1/4 ABCD as well and same issue with row 2. I have gone directly to the other panels in the row and panels in other rows. Anything addressing the 2nd row shows the same result, no matter what.

Offline CaptainMurdoch

  • Administrator
  • *****
  • Join Date: Sep 2013
  • Location: Washington
  • Posts: 9,748
  • Kudos: 205
Re: Strange output on my P10 panels
« Reply #14 on: November 28, 2018, 07:07:25 PM »
Captain Murdoch, I have tried the 1/4 ABCD as well and same issue with row 2. I have gone directly to the other panels in the row and panels in other rows. Anything addressing the 2nd row shows the same result, no matter what.

What happens if you connect the second row of panels to the octoscroller output that the first row is currently plugged into?  does the problem follow the output on the Octoscroller or is it always the second row?  If it follows the output on the octoscroller then the output or BBB is potentially bad.  If it is always the second row, that indicates the second row cable or first panel in the second row is bad.

 

Back to top