Intro to the Octoscroller

[dkulp]

The Octoscroller was designed to leverage the power of a BeagleBone Black to drive P10 and P5 LED panels. 

Octoscroller Capabilities:

   
• 8 HUB75 Connectors/Outputs - each can driver 12 P10 panels or 4/5 P5 panels
• The Octoscroller supports various types of panels including 1:4 and 1:2 scan "outdoor" panels, various ways panels multiplex rows, etc...

Requirements:

   
• SD card with the latest FPP image.  The image can be flashed to the eMMC on the BBB and run from there.

Supported FPP Versions:
Octoscroller 1.x support was added to the BBB LED Panels page in FPP for FPP 1.5.   All versions of FPP since 1.5 support it.   2.x pinout support was added for FPP 2.0.

Supported BeagleBones:

   
• All versions support the BeagleBone Black, BeagleBone Black Wireless, BeagleBone Green
• The gigagit adapter on the SanCloud BeagleBone Enhanced consumes pin P8-10 which is used for the green on the bottom half of port one on the v1.x pinout or the blue on the top half of port 5 on the v2.x pinout.  Thus, those ports would not work properly with the BBE.
• The BeagleBone Green Wireless WIFI adapter consumes a bunch of pins.  With the 1.x board, nothing will work.  The consumed pins include clock line and several of the address lines.  With v2.x board, ports 1 through 6 will work, 7 and 8 will not.

Notes:

   
• v1.x and v2.x boards use a completely different pinout. You need to make sure the proper version is set on the "BBB Strings" page in FPP.
• The Octoscroller only has 4 address lines and thus can only support up to 1:16 scan panels.  There are no extra pins on the BBB to be able to add support. 

Availability:
Octoscroller 1.x (and variants) are available from several vendors.   Some of the variant have added things like RTC's, better connector spacing, extra buffers, etc...

Hanson Electronics	http://www.hansonelectronics.com.au/product-tag/octoscrolla/
DIY LED Express	http://www.diyledexpress.com/index.php?main_page=product_info&cPath=75&products_id=458
Renard Plus	http://renard-shop.com/home/60-octoscroller-cape.html
Wired Watts	https://www.wiredwatts.com/octoscrolla-rtc
Kulp Lights	https://kulplights.com/product/octoplus/

The OctoPlus is the only 2.x pinout board available.

[AAH]

Thanks for putting this info up Dan.I have put up a caution in a few places. Some versions of the octoscroller have had SN74LVC245AN used as the buffer chip in them. The LVC series of 74 series IC's are for 1.6 to 3.6V supply. They aren't suitable for running off 5V which is what they do in the octoscroller. Somewhere, somehow there was a change the original 74LS245 for a 74LVC245 without knowing what they were doing or the consequences. Running a 3.6V chip off 5V is a great risk to kill the IC potentially causing it to go into flames and also potentially kill the Beaglebone Black as its outputs are fairly fragile.

https://trmm.net/Octoscroller
I have Octoscrolla's fully built, with RTC, as bare pcb's, as kits and will soon have them with a couple of extra features. I don't use 74LVC245.
http://www.hansonelectronics.com.au/product-tag/octoscrolla/

[Stormyblade]

I have a question for you Dan, regarding the Octoscroller and it's pinouts under the Channel Outputs tab. Under the "older" versions of FPP, you had the two wiring pinout selections, v1.x and v2.x -- now with FPP 2.7 there is a dropdown list:

   
• Adafruit
• Adafruit PWM
• Standard Pi1
• Classic
• Classic Pi1

Which one of these would correspond to your OctoscrollerPlus with the 2.x pinout? Also, you mention that the OctoscrollerPlus is the only board that supports 2.x pinouts -- why do I need 2.x pinouts? Would a 1.x pinout work for P5 panels (I am still waiting for mine to show up and can't recall if they are 1:16 scan and DIYLEDExpress no longer shows them along with their specifications on their website)?

[dkulp]

I have a question for you Dan, regarding the Octoscroller and it's pinouts under the Channel Outputs tab. Under the "older" versions of FPP, you had the two wiring pinout selections, v1.x and v2.x -- now with FPP 2.7 there is a dropdown list:

Which one of these would correspond to your OctoscrollerPlus with the 2.x pinout? Also, you mention that the OctoscrollerPlus is the only board that supports 2.x pinouts -- why do I need 2.x pinouts? Would a 1.x pinout work for P5 panels (I am still waiting for mine to show up and can't recall if they are 1:16 scan and DIYLEDExpress no longer shows them along with their specifications on their website)?

Umm... that's not good.   That's got to be a bug.   Those are the Pi pinouts, not the Octo's.   Let me take a look. 

[Stormyblade]

Doh! *smacks head*  It's not a bug, it's the operator.     I had both my Pi 3B+ (I ordered one, just for the hell of it) AND my BBB hooked up to my PC and got my wires crossed when I was looking at configs. I switched over to the BBB FPP and it does show just v1.x and v2.x pinouts for a BBB.


Aside from that, my other questions still remain about the v2.x pinouts using a BBB/Octo setup for a set of P5/P10 panels: do I *need* your OctoPlus with the v2.x pinouts to run P5 panels? Is there something your board gives that the other ones you referenced in your original post don't give to P5 panels? 


(I'm just checking (again) to see if I need to update any other hardware and I can't see the details/specs of the P5 panels I ordered from DIYLEDExpress during the pre-sale as they have removed the panels from their site - probably because they are gearing up for another pre-sale.)

[dkulp]

Aside from that, my other questions still remain about the v2.x pinouts using a BBB/Octo setup for a set of P5/P10 panels: do I *need* your OctoPlus with the v2.x pinouts to run P5 panels? Is there something your board gives that the other ones you referenced in your original post don't give to P5 panels? 

The v2.x pinout will give better refresh if using 5 or fewer outputs.   Once you start using 6 or more outputs, then they end up equal at this point. 


The v2.x pinout DOES put the "oe" pin on a pin that is capable of doing hardware PWM, but we don't use that yet.   At some point, I do want to work on getting it to use the hardware PWM to handle the brightness and such so that the PRU code just needs to output data.   That should boost the refresh a bit (10% or more) and reduce flicker a bit.  However, I've never gotten around to doing it and other things have taken a higher priority.  Part of the reason is that using the PWM from the PRU involves setting values in various registers, and the "chapter" in the technival documents about the PWM subsystem that documents it all is well over 350 pages long.  Very complex.   Some day maybe, but other things have become more important. 

[Stormyblade]

Aha - so in my specific case, where I am going to be connecting 1 BBB/Octo to 20 P5 panels, with each Octo output going to either 2 or 3 panels, but still using all 8 outputs, it won't matter if I use your OctoPlus board or the current Octo boards I have, correct? My other one will be connected to just 4 panels, 1 panel per output, but I bet the OctoPlus wouldn't give me much in the way of benefits there either, right?

Drat...I was looking for a reason to possibly upgrade to your OctoPlus boards...because, I gotta say...those little OLED screens look pretty cool.    Although, truth be told, those little things would be tucked away inside the boxes of my matrix and Tune To sign and you'd never really get a chance to see them, so it's not that much of a loss.

As always, great to get stuff like this explained...thanks Dan!

[dkulp]

Aha - so in my specific case, where I am going to be connecting 1 BBB/Octo to 20 P5 panels, with each Octo output going to either 2 or 3 panels, but still using all 8 outputs, it won't matter if I use your OctoPlus board or the current Octo boards I have, correct? My other one will be connected to just 4 panels, 1 panel per output, but I bet the OctoPlus wouldn't give me much in the way of benefits there either, right?

Right.  Neither one would really help.  In the second case, with one panel per output, we have to throttle it down as the data would be too fast and we'd get ghosting.   So yea, not help.

Drat...I was looking for a reason to possibly upgrade to your OctoPlus boards...because, I gotta say...those little OLED screens look pretty cool.    Although, truth be told, those little things would be tucked away inside the boxes of my matrix and Tune To sign and you'd never really get a chance to see them, so it's not that much of a loss.

Well, it also has an RTC and a temperature sensor.   You do NEED to know the temperature inside your box, right? 

[Stormyblade]

Aha - so in my specific case, where I am going to be connecting 1 BBB/Octo to 20 P5 panels, with each Octo output going to either 2 or 3 panels, but still using all 8 outputs, it won't matter if I use your OctoPlus board or the current Octo boards I have, correct? My other one will be connected to just 4 panels, 1 panel per output, but I bet the OctoPlus wouldn't give me much in the way of benefits there either, right?

Right.  Neither one would really help.  In the second case, with one panel per output, we have to throttle it down as the data would be too fast and we'd get ghosting.   So yea, not help.

Drat...I was looking for a reason to possibly upgrade to your OctoPlus boards...because, I gotta say...those little OLED screens look pretty cool.    Although, truth be told, those little things would be tucked away inside the boxes of my matrix and Tune To sign and you'd never really get a chance to see them, so it's not that much of a loss.

Well, it also has an RTC and a temperature sensor.   You do NEED to know the temperature inside your box, right? 

lol...I do...I really do!     Of course, my wife wants me to have a really good display, so... [size=78%] [/size]

[Stormyblade]

Hey Dan,


It's been a few months since you opened up this topic and now that I've got all 20 of my P5 panels hooked up to the BBB/Octo in the most efficient way I can think of, I'm wondering if this stuttering that I'm seeing is a problem or not.


Here's the deal: FPP is set to Bridge mode on the BBB/Octo, and I have a network cable going directly from my router on my network to the BBB. In xLights I turn on the 'Output to Lights' option and play sequences through (to see how much sharper they look with P5 panels vs P10 panels) the system. Now, it doesn't seem to matter if I use 5 outputs (O1 daisy-chains across 4 panels, O2 daisy-chains across 4 panels, etc), or all 8 outputs (O1-O4 go to 2 panels each, and O5-O8 go to 3 panels each). I still see stuttering and flickering across the entire matrix as it tries to render the video clip that is playing within the sequence.


Is this an issue? Will this all but disappear if I load the sequence onto my Pi, and then load the video file to the BBB, and have the Pi start the sequence with the BBB in Remote or Bridge mode? Or is there something I should be looking at to help things look better?

[JonB256]

On the BBB in Bridge Mode, are you seeing any Packet Errors on the Status Page?

I know that 20 panels of P10s don't stutter using DDP, but P5 panels are 4x the data. The Ethernet (100mbit) on the BBB may be overloaded.

It is so easy to switch the BBB over to be a Remote and test, assuming you have room for the FSEQ file.
xLights (xSequence, anyway) can be a Master.

[Stormyblade]

On the BBB in Bridge Mode, are you seeing any Packet Errors on the Status Page?

I know that 20 panels of P10s don't stutter using DDP, but P5 panels are 4x the data. The Ethernet (100mbit) on the BBB may be overloaded.

It is so easy to switch the BBB over to be a Remote and test, assuming you have room for the FSEQ file.
xLights (xSequence, anyway) can be a Master.

Yes, I am, usually on the order of a few thousand to several thousand errors, depending on how long the sequence runs.


I am planning on setting up the BBB to be a player, and was thinking of doing a setup test to have the Pi be the Master and the BBB be the Remote player. I just haven't gotten around to that point yet is all, and have been doing testing with the matrix hooked up directly to the computer I am using for sequencing via the home router. I was just curious if I should move forward with what I am seeing, or should I make some adjustments before trying again.

[dkulp]

An error rate like that suggests network issues of some sort, although running in remote mode would eliminate those.   I've had a matrix setup of 16 P5 panels (98K channels) but with the DDP output set to 120K channels running a looping butterfly and I'm seeing a less than 0.02% error rate.  (250 errors out of 1.7M packets)  And my computer is on Wifi (beagle is wired).  This is also using a 25ms sequence, not 50, so tons of data.


Some hints for using FPP in bridge mode with large matrices:


1) #1 important thing: use DDP, not e1.31.  DDP is way more efficient at getting the data there.


2) If using DDP or using an e1.31 frame sync packet, go to the advanced setting page and increase the bridge interval to 100ms (or at least higher than your sequence rate.   My sequences are all 25ms, so setting to 50ms is appropriate.


The setting in #2 is how often it sends the data out in absence of some other trigger to do so.   By default, it's 50ms.   That means if you send 25ms data in bridge mode without some sort of sync, it will pretty much not display 1/2 the frames.   DDP sends an "end of data" flag with the last packet so FPP can know to display the frame now.   The 1.31 sync universe also can do that, but not commonly used.  If you make sure that setting is higher, a slightly delayed packet may not come in after the timer and thus tearing should be reduced.    For me, I keep it at 50ms, but my sequences are 25ms so even a slight delay should keep it well under the timer. 

[Stormyblade]

Dan,


I'm using DDP. Discussions earlier this year convinced me this is the way to go with a matrix the size of mine -- I didn't do it during last years holiday season because I wasn't sure how to set it up.


Okay - Bridge Mode Interval set to 100ms, and not the default of 50.  Question, though: when running the matrix in my display, I'll run it as a Remote player, right? That way, the media file (in my case, it will be mostly movie clips) will be stored on the BBB SD card and it will get triggered when to play said file from the Master player, which, in my case, is the Pi 3B+.

[Stormyblade]

All right...yesterday evening, after uploading the movie clip file to the BBB, and the sequence to both the Pi 3B+ and BBB,  I hooked up both devices to my home network router. I set the BBB into Remote mode and hit Play and everything looks great - no tearing/clipping/etc on the matrix display. 


Thanks again for the help - I feel as though I took several steps backward in the learning curve, but it seems to have sorted itself all out.