P10 Panel Setup

[ndutton]

OK banging my head against a wall here. I am sure I am missing something very obvious but I can't see it.

I have some P10 panels and I just cant seem to get them working correctly. 

• This is the S5 test seq I made: https://www.dropbox.com/s/zuzpbgnu9sbwrfb/IMG_0932.MOV?dl=0
• This is the result: https://www.dropbox.com/s/a44hrxmckchi9se/IMG_0933.MOV?dl=0

These are the P10 Specs: https://www.dropbox.com/s/jb0djly37ia3vio/2019 03 P10 Specs.png?dl=0

 

Here are my configs:

• LOR Prop Definition: https://www.dropbox.com/s/sq3gvnuw2carqrs/Prop Definition.png?dl=0
• E131 Channel Inputs for Bridge Mode: https://www.dropbox.com/s/1funxkbg9y4e1k4/E131 Channel Bridge Mode Inputs.png?dl=0
• E131 Channel Outputs: https://www.dropbox.com/s/10inkotorbhs7n7/E131 Channel Outputs.png?dl=0
• LED Panel Configs: https://www.dropbox.com/s/n2516j21revc2ax/LED Panel Configs.png?dl=0

 

Can anyone see what I am missing. I tried to limit it to a single panel for the time being so I can limit the variables but I have swapped out the panel incase it was a hardware issue. Same result on all.

 

 

Edited March 8, 2019 by ndutton

[k6ccc]

Well for starters, remove the stuff on the E1.31 / ArcNet / DDP tab of the Channel outputs section.  It does not need to be there.  Confirm that LOR network configuration is set for MultiCast.  You really should use Unicast unless there is a real pressing issue requiring multicast.

Can you control the lights properly from the Channel test page on FPP?  If that does not work, you know the problem is entirely within the FPP world, but if that works properly, then it's likely a config issue between FPP and LOR.

I can get on the phone with you later.

[ndutton]

Jim,

Unfortunately, none of those suggestions worked. I will try it with an RPi instead of a BBB, maybe there is a problem with the BBB build of FPP, or maybe its some wierd thing about these particular panels. I will try the RPi idea, failing that I'll get a new panel and see if that is the issue.

 

N

[k6ccc]

I am currently running my 4W x 3H P10 panel just fine from a BeagleBone Black in Bridge mode at the moment.

My offer of getting you on the phone still stands, and if you have TeamViewer, either you can look at my setup or I can look at yours.

 

[dkulp]

These are 1/4 scan panels.    None of the RPi hats will work with them.   You would need a colorlight card to get it to work with a Pi.

It looks like these are "1/4 scan ABCD" with what appears to be a 32 pixel interleave.  I would give that a try first. 

[k6ccc]

At least initially he was using a BBB.  He will be calling my shortly.

 

[ndutton]

@dkulp

Dan, Thanks for the feedback. Seems that these panels may have some weird combination of settings that currently isn't supported by either the FPP to the BBB Octoscrolla. To prove the theory I just ordered some different panels from CFOlights.

At some point hopefully the original panels will be supported and I can make a different prop with them. These original panels appear to be quite common out of China on eBay from a few different vendors.

[dkulp]

If you send one or two to me, I can get them working.      But did you try the "1/4 Scan ABCD" setting?    I'd like to see a video with it reset to that setting.

 

[ndutton]

OK I can send you one. PM me with an address and I'll send it out today.

[ndutton]

@dkulp

Dan,

Here is the information you requested: (when 1/4 ABCD set to 32 pixels interleave)

• https://www.dropbox.com/s/iiy8o91jld69b4f/IMG_4109.MOV?dl=0
• https://www.dropbox.com/s/oasqyh84gmyiz0t/IMG_4108.jpg?dl=0
• https://www.dropbox.com/s/dfkew0ckxs8xvkl/IMG_4107.jpg?dl=0

and when set to interleave off:

• https://www.dropbox.com/s/8nsprigg4shs2u2/IMG_4111.MOV?dl=0
• https://www.dropbox.com/s/ik50vanxuumc2hb/IMG_4110.jpg?dl=0

 

FYI:

This is a BBB with the Octoscrolla v1.5c from Hansen electronics.

 

N

 

[dkulp]

Interleave off should be removed in this case.  It's impossible to have as 1:4 scan panels have to have some sort of interleave.   This LOOKS like a 16 interleave of some sort as the 32 interleave it putting things in groups of 16 pixels.  Try the two different 16 pixel interleaves.    On a positive note:  you definitely showed it's the ABCD style addressing.   That's a good start. 

This is why 1:4 scan "outdoor" panels are such a pain and we generally recommend the 1:8 scan panels.  The hub75 interface has two sets of rgb data lines.  For 1:8 scan 16 row panels, it can fully output 2 rows per scan using the two lines.  For 1:4 scan, it has to output 4 rows of information on 2 lines.  Thus the interleave.  However, each manufacture outputs the row data in different order depending on however the person who laid out the PCB panel decided to route the signals.    Some panels require one entire row and then the entire second row (32 interleave).   Others do 16 pixels on row 1, then 16 on row 2, then back to 16 on row 1 then 2 (16 interleave).  Some do 8 or 4.     And some, when doing 4/8/16 will do the second row backwords (the zig zags).   The main issue is that there is no indication on the boards themselves what it should be configured as.   It's really trial and error.   You have to try each until you find what works.   And if none works, then some new PCB designer decided to do something completely new that I haven't seen yet.

 

[ndutton]

@dkulp

LOL - well I *believe* I have tried every combination at this point from v2.6 FPP interleave option pull down, so I do agree this is a magical unknown.
The 16 pixel options are below for you to see.

• 16 pixel only : https://www.dropbox.com/s/r62bz2x5cb7qxtp/IMG_4113.MOV?dl=0
• 16 pixel ZigZag : https://www.dropbox.com/s/cgf8kvh3i6320vs/IMG_4114.MOV?dl=0

I am still open to send a panel to you if you wish, I can't believe I will be the last person to run into this panel design down the road?

N

[ndutton]

as expected, it didn't work on the RPi either but it was interesting to see that the RPi and BBB screens are quite different on the LCD panel page of version 2.6

[ndutton]

OK here is the latest. I changed the panels to a 1/8 scan and everything works. I can out put to the panels.

I have a new problem. I created a test sequence that scrolls through the panels displaying each panel name (Panel 1, panel 2, etc.) and doing a quick RGB colour flash so I can see I can get to each panel and test the LEDS. Nothing too complex.

Here is the weird bit.

There are 12 panels. If I hang all panels off one BBB octascroller port then everything works fine (Doesn't matter which port). If I split them so 6 are on O1, and 6 are on O2. there is NO output observed at all. I assume I am missing something in setup?

https://www.dropbox.com/s/t5rxwcy8b7smece/2019 04 30 MultiOutput.png?dl=0

 

Ideas?

[k6ccc]

Looking at your layout, I see a couple questions.  Your model layout shows a starting location of lower left, but your outputs are starting on the top.  Not that this should make it fail to light something.  Also, curious about the individual panels.  I'm doing part of this by memory, but as I recall, if the arrows are Up, the panels are oriented horizontally.  That would mean based on the layout shown that you have your inter panel cables running diagonally.  This will work, but the cables that generally come with the panels are too short to do it that way.

 

[ndutton]

by the time i am finished there will 3 columns of 6 panels (we have a deck in the front of the house where the roof is held up by 3 columns, the props will be mounted to the 3 supports)

aka: https://www.dropbox.com/s/2aqq66rdk2de6t4/3columns.JPG?dl=0

I want to run each column on its own output port (O1, O2, O3) so the cable length is minimised. As I say, If a single output is connected up it all works just fine but when I use multiple outputs nothing works. I assume I am missing a fundamental setting.

(for testing, I am just using 2 panels. either hanging them off their individual port (O1, O2) which doesn't appear to work, or one panel hanging off the second (which does work).

Edited April 30, 2019 by ndutton

[ebrown1972]

On 3/8/2019 at 6:11 PM, k6ccc said:

Well for starters, remove the stuff on the E1.31 / ArcNet / DDP tab of the Channel outputs section.  It does not need to be there.  Confirm that LOR network configuration is set for MultiCast.  You really should use Unicast unless there is a real pressing issue requiring multicast.

Can you control the lights properly from the Channel test page on FPP?  If that does not work, you know the problem is entirely within the FPP world, but if that works properly, then it's likely a config issue between FPP and LOR.

I can get on the phone with you later.

Jim, I have a P10 matrix that I have used the last few years in standalone mode. I am using a BBB to control it. I have considered using the P10 as part of my display. I am by no means a network guy and find that I am lucky that I even get all of my controllers to work in E131 lol. How do I set up the P10 to run in unicast? Also how does that relate to the network configuration in LOR? In multicast the only thing I really need to know is what universe I am on. I just need a little guidance to get it set up in unicast. Thanks,

Eric

[k6ccc]

11 hours ago, ebrown1972 said:

Jim, I have a P10 matrix that I have used the last few years in standalone mode. I am using a BBB to control it. I have considered using the P10 as part of my display. I am by no means a network guy and find that I am lucky that I even get all of my controllers to work in E131 lol. How do I set up the P10 to run in unicast? Also how does that relate to the network configuration in LOR? In multicast the only thing I really need to know is what universe I am on. I just need a little guidance to get it set up in unicast. Thanks,

First of all, you CAN run your FPP installation in Multi-cast, but I recommend against it.  Let's see if I can walk through all the steps required to drive your P10 matrix from LOR.  The screen captures I am going to include here are from LOR 5.3.10, and FPP 3.4.1 running on  Pocket Beagle with a Pocket Scroller.  The matrix is 12 P10 panels in 4W x 3H configuration - or 48 pixels high by 128 pixels wide.  Although you likely know the output page parts in FPP, I am going to include them here because it may benefit someone else - and besides, it's too hot for me to go outside and finish the roof peak lights (84 outside and likely close to 150 for the composition roofing that I have to lay on for part of it).

Let's start in a preview for the matrix:

And then the channel layout:

Make sure to select 512 channel Mas (the green arrow) and Auto-number (the red arrow).  For this matrix, there are 36 universes, but this screen capture only shows a little over half of them.  When set for auto-number, the software will fill them all in so you only enter the first universe.

Next up is Network Preferences.  Select DMX and the Advanced settings.

When you select a universe, you get something like this:

Select "Use E1.31" (the orange arrow).  To use Unicast, select "Specify" and enter the IP address of the FPP (the blue arrow).  You can enter the data for the first universe and have it create up to 36 universes by selecting a number in the dropdown (the brown arrow).

Now in FPP, go to the Input/Output setup and select Channel Inputs.  I am going to show you two different ways to fill this section out.  In the first method, set one line for each universe - 36 in my case (the red arrow), then fill out the first line and select "Save".  Make sure to set the Universe Size to 512 (the orange arrow) and Universe Type to "E1.31 - Unicast" (the yellow arrow).  After saving the first line, select the first line and then the "Clone" button and enter the number of universes minus 1 to create the additional universes.

  There is an easier way however.  Just use one lie for Input count (the red arrow), and set like the previous example, EXCEPT set Universe Count to the number of universes needed - 36 in my case (the brown arrow)

Next up is the Input/Output Setup and select Channel Output and the LED Panels tab.  Select the dropdown for Panel Layout - 4x3 in my case (the red arrow).  Set Single Panel Size - what is shown is fairly common (the green arrow).  The Model Start corner must match the model in the LOR Preview (the light blue arrow).  The connection presumably will be "Hat/Cap/Cape" (the purple arrow).  The Wiring Pinout needs to match what you are using (the brown arrow).  The LED Panel Layout needs to be set for your actual layout and wiring arrangement.  For each panel, set the Output, Panel number and Color order (O, P, and C dropdowns).  In my case Output 1 of the PocketScroller drives the top row of panels in a left to right arrangements (as viewed from the back side).

Last step is to set the operating mode for FPP into Bridge mode on the Status/Control tab and the Status Page to Bridge (the red arrow).

When in Bridge mode, there will be a listing on the Status Page that shows the stats for each universe.  This page can be set to update in almost real time by selecting the "Live Update Stats" (the red arrow), or be static by not selecting that checkbox.  If you do not have Live update selected, you can update the display by selecting the "Update" button (the green arrow).  Some people report that leaving Live update enabled takes a lot of CPU cycles.  I leave mine with Live update enabled.

Did that give you enough?

 

 

 

[ebrown1972]

10 hours ago, k6ccc said:

First of all, you CAN run your FPP installation in Multi-cast, but I recommend against it.  Let's see if I can walk through all the steps required to drive your P10 matrix from LOR.  The screen captures I am going to include here are from LOR 5.3.10, and FPP 3.4.1 running on  Pocket Beagle with a Pocket Scroller.  The matrix is 12 P10 panels in 4W x 3H configuration - or 48 pixels high by 128 pixels wide.  Although you likely know the output page parts in FPP, I am going to include them here because it may benefit someone else - and besides, it's too hot for me to go outside and finish the roof peak lights (84 outside and likely close to 150 for the composition roofing that I have to lay on for part of it).

Let's start in a preview for the matrix:

And then the channel layout:

Make sure to select 512 channel Mas (the green arrow) and Auto-number (the red arrow).  For this matrix, there are 36 universes, but this screen capture only shows a little over half of them.  When set for auto-number, the software will fill them all in so you only enter the first universe.

Next up is Network Preferences.  Select DMX and the Advanced settings.

When you select a universe, you get something like this:

Select "Use E1.31" (the orange arrow).  To use Unicast, select "Specify" and enter the IP address of the FPP (the blue arrow).  You can enter the data for the first universe and have it create up to 36 universes by selecting a number in the dropdown (the brown arrow).

Now in FPP, go to the Input/Output setup and select Channel Inputs.  I am going to show you two different ways to fill this section out.  In the first method, set one line for each universe - 36 in my case (the red arrow), then fill out the first line and select "Save".  Make sure to set the Universe Size to 512 (the orange arrow) and Universe Type to "E1.31 - Unicast" (the yellow arrow).  After saving the first line, select the first line and then the "Clone" button and enter the number of universes minus 1 to create the additional universes.

  There is an easier way however.  Just use one lie for Input count (the red arrow), and set like the previous example, EXCEPT set Universe Count to the number of universes needed - 36 in my case (the brown arrow)

Next up is the Input/Output Setup and select Channel Output and the LED Panels tab.  Select the dropdown for Panel Layout - 4x3 in my case (the red arrow).  Set Single Panel Size - what is shown is fairly common (the green arrow).  The Model Start corner must match the model in the LOR Preview (the light blue arrow).  The connection presumably will be "Hat/Cap/Cape" (the purple arrow).  The Wiring Pinout needs to match what you are using (the brown arrow).  The LED Panel Layout needs to be set for your actual layout and wiring arrangement.  For each panel, set the Output, Panel number and Color order (O, P, and C dropdowns).  In my case Output 1 of the PocketScroller drives the top row of panels in a left to right arrangements (as viewed from the back side).

Last step is to set the operating mode for FPP into Bridge mode on the Status/Control tab and the Status Page to Bridge (the red arrow).

When in Bridge mode, there will be a listing on the Status Page that shows the stats for each universe.  This page can be set to update in almost real time by selecting the "Live Update Stats" (the red arrow), or be static by not selecting that checkbox.  If you do not have Live update selected, you can update the display by selecting the "Update" button (the green arrow).  Some people report that leaving Live update enabled takes a lot of CPU cycles.  I leave mine with Live update enabled.

Did that give you enough?

 

 

 

Jim, Thank you for taking the time to show me everything that needed to be done. I appreciate that very much. It looks like my mistake was in the E131 settings in LOR. I was not using the ip address for the fpp. When I find time I will give that a shot and see if it works. Thanks again.

[ebrown1972]

I just tested it out Jim, and the one setting I had in the E131 tab in LOR was wrong. I changed the ip to the ip of my fpp and all is playing just fine now. I had originally tried running my p10 in multicast but it would flood my whole display and make everything flicker. Now that it is set up for unicast I have no flicker and it appears that I will be able to use my P10 as part of my display rather than using it in standalone mode. Your help and time is extremely appreciated. Thank you so much,

Eric