Author Topic: PocketBeagle based controller thoughts/ideas....  (Read 7047 times)

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
PocketBeagle based controller thoughts/ideas....
« on: January 03, 2018, 07:23:55 AM »



I want to start a discussion around peoples thoughts/needs are for PocketBeagle based controllers.   


First, the PocketBeagle itself doesn't come with any headers attached.  Thus, people either need to be comfortable soldering on 2 36pin female headers onto the bottom or it's something I can probably provide as part of any group buy.   I was thinking of offering an option of PocketBeagle with headers pre-mounted for $30.    Thus, it would be add SD card and USB Wifi/Ethernet and up and running.


Next, the PocketScroller - I have this completely designed an prototype working.   This would be a 6 output P10/P# controller capable of 72 panels.   It looks like the cost would be around $18 assembled, maybe a little less.   The reason it's a little more than an Octoscroller despite 2 less outputs is the USB.   It needs to have a USB port added along with a USB power switch/regulator.   Those regulators are difficult to find and expensive for non-SMD so I went with the SMD parts for those which adds a stencil to the cost.   Anyway, with the $30 PocketBeagle, that would be $48 ready to go.


F8-PB - I have an F8 designed for the PocketBeagle.   I'm debating various options for this.   The design I currently have is an exact equivalent of the F8B+ I did last year, with 8 local ports, 2 RJ45's for driving 8 universes of DMX/PixelNet or Pixels (via differential receiver), a 3rd RJ45 for pixels (differential receiver), and the 40pin expansion header.   That said, the expansion header adds over $2 to the cost of the board and I'm not sure many people use it.   Each differential output also adds a couple dollars to the board (differential chip and RJ45).  Thus, kind of debating offering a "light weight" version of the board that excludes those, or at least would require you to solder on those parts later if you do need them.   (current design uses SMD differential chip, so that would be harder, might adjust to through hole for these to make this option easier)  No idea on full cost of these yet.   Probably around $45-50 as it's similar to the F8B+ from last year, but does have the added USB parts.


Those are what I'm planning right now.   Do folks feel the need for anything else?    I did an F32B last year, but didn't sell very many (I still have a couple if folks are interested) so I'm not sure if going larger makes much sense.    Also thought about an F8 with 6 differential outputs instead of 3.   Again, not sure how useful that would be for folks.  (not at all useful for me, I refuse to to use Cat5 for anything)
Dan Kulp

Offline brianweb

  • Jr. Member
  • **
  • Join Date: Jan 2014
  • Location:
  • Posts: 54
  • Kudos: 0
Re: PocketBeagle based controller thoughts/ideas....
« Reply #1 on: January 03, 2018, 08:05:33 AM »
I would be in for 1 or 2 Pocket Beagle and Pocket scrollers.

Offline pixelpuppy

  • Hero Member
  • *****
  • Join Date: Aug 2015
  • Location: Dallas, TX
  • Posts: 1,350
  • Kudos: 42
Re: PocketBeagle based controller thoughts/ideas....
« Reply #2 on: January 03, 2018, 08:18:41 AM »
... I refuse to to use Cat5 for anything...

OK I'll bite.  What is the reason behind that?
-Mark

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
Re: PocketBeagle based controller thoughts/ideas....
« Reply #3 on: January 03, 2018, 08:44:40 AM »
... I refuse to to use Cat5 for anything...

OK I'll bite.  What is the reason behind that?


Being extremely colorblind I really struggle to put ends on cat5 cables that actually result in usable cables.    I find making cat5 cables extremely frustrating and it takes all the enjoyment out of setup.   I couldn't count the number of times I've had to redo ends multiple times to get it right.


I removed all cat5 from my display 3 years ago (other than a few small "cat5 -> 4pin pigtail" for DMX) and have not regretted it one bit.   


Offline travcam

  • Newbie
  • *
  • Join Date: Feb 2017
  • Location: Huntington, WV
  • Posts: 33
  • Kudos: -1
Re: PocketBeagle based controller thoughts/ideas....
« Reply #4 on: January 03, 2018, 11:12:13 AM »
... I refuse to to use Cat5 for anything...

OK I'll bite.  What is the reason behind that?


Being extremely colorblind I really struggle to put ends on cat5 cables that actually result in usable cables.    I find making cat5 cables extremely frustrating and it takes all the enjoyment out of setup.   I couldn't count the number of times I've had to redo ends multiple times to get it right.


I removed all cat5 from my display 3 years ago (other than a few small "cat5 -> 4pin pigtail" for DMX) and have not regretted it one bit.


I worked with an IT guy years ago that was completely color blind.  YES, you read that right, an IT guy!  I'm still not sure how he did it, but he did cables all the time.  He once asked me if something was slate...  that's when I asked if he was color blind and he fessed up.  But he did it for years.  Hard to say how much rework he had to do.

Offline JonB256

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Granbury, Texas
  • Posts: 5,128
  • Kudos: 117
    • Granbury Christmas Lights
Re: PocketBeagle based controller thoughts/ideas....
« Reply #5 on: January 03, 2018, 11:37:10 AM »
I was in the Navy and colorblindness would have kept me out of the Electronics Technician (and Nuclear) fields.
At almost 65, my vision is more of a problem but can still see color. I have a pile of CAT5 that needs new ends because the stupid tabs break off. Just need very good lighting and clean glasses.

Offline Emuney18

  • Hero Member
  • *****
  • Join Date: Nov 2015
  • Location: Apex, NC
  • Posts: 551
  • Kudos: 7
Re: PocketBeagle based controller thoughts/ideas....
« Reply #6 on: January 04, 2018, 05:30:05 AM »
I would be in for 1-2 PocketBeagles and scroller.  I am new to Beagles. Is the F8 just a pixel controller like a PiHat?  Is there a wiki or something where I can read about it?


Sent from my iPhone using Tapatalk

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
Re: PocketBeagle based controller thoughts/ideas....
« Reply #7 on: January 04, 2018, 06:48:47 AM »
I would be in for 1-2 PocketBeagles and scroller.  I am new to Beagles. Is the F8 just a pixel controller like a PiHat?  Is there a wiki or something where I can read about it?



Much of the information can be found from the links in the group buy from last year:


http://falconchristmas.com/forum/index.php/topic,7341.0.html


Basically, it's much like the PiHat, but for the Beagles.    The Beagles, due to the 2 dedicated PRU's (basically, 2 embedded real time processors) and tons of GPIO pins can output a LOT more pixels than Pi.    The Beaglebone Black has 55 usable GPIO pins so, in theory, it could drive 55 strings of 800 pixels using 25ms sequences.   That said, the largest capes I've sean only expose 48.   A lot of that is due to the board size (the F32B is 14.9cm x 14.9cm SPECIFICALLY because once you go above 15cm, the cost of manufacturing the boards jumps quite a bit), but also not many people need/want 48 strings in one place.   Thus the expansion ports are nice as those that DO need it, add an expansion board.   Also, the differential ports allow a cat5 cable to send pixel data for 4 strings over a long distance.


The PocketBeagle is much smaller and only has 44 usable GPIO pins.   One of my early possible layouts for an F8B on the PB is shown in:


http://falconchristmas.com/forum/index.php/topic,7851.msg75907.html#msg75907


I've since made it a little wider to hold an additional RJ45 and will likely make it a little wider again to hold an expansion port.


In any case, if you used a PiHit, using the Beaglebone controllers is really not much different.   The "bigger" bones have on-board flash that CAN (optionally) be used to hold the OS and then use the SD card for storage.   The PocketBeagle doesn't have the onboard flash so you have to use the SD card for both OS and storage (the USB port will be needed for wifi or ethernet, unless you use a USB hub)  Configuring the strings on the beagle based capes is a LOT easier and more powerful than on the Pi's.   Each output port can have it's own starting channel, it's own RGB order, it's own count of null pixels, own brightness/gamma, etc...


Anyway, the F8-PB will be slightly larger than a PiHat (but will definitely remain below 10cm x 10cm due to cost of manufacturing), but mostly because 8 pixel connectors and fuses take space no matter what you do.  However, with the differential outputs and expansion, it can drive over 28K pixels.   Quite a bit more than a Pi.  :)




Offline Emuney18

  • Hero Member
  • *****
  • Join Date: Nov 2015
  • Location: Apex, NC
  • Posts: 551
  • Kudos: 7
PocketBeagle based controller thoughts/ideas....
« Reply #8 on: January 04, 2018, 06:16:42 PM »
Thanks Dan. That helps a bunch. So I read that correct that I can use the Falcon receivers as the beagle differential receivers?   This would have helped this year.  How far can the first pixel be from the F8 boards?  If I could get 10' that would be awesome for what I want.


Sent from my iPhone using Tapatalk

Offline Bshaver

  • Developer
  • ******
  • Join Date: Aug 2014
  • Location: Denver, CO
  • Posts: 1,426
  • Kudos: 22
Re: PocketBeagle based controller thoughts/ideas....
« Reply #9 on: January 05, 2018, 07:47:53 AM »
Dan,
Is it possible to have the beagle be on TOP of the board instead of on the bottom? I accomplished it by using standoffs from the CG1500 - and that works ok i'm just concerned about pressure points when connecting the euro-connectors into the board at times. I think the F4 is perfect being on top of the beagle, but the larger boards if the processor is accessable from the top (even if it is upside down). Now not bending the pins is another thing taking it on and off. Maybe someone could tinker up a 3d beagle extraction tool that pulls it straight up and off the boards!
Denver Colorado
45,000 blinky blinks mix of pixels and 130 AC Channels (powered by Minleon & Lynx) ALL Powered by FPP and Falcon!

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
Re: PocketBeagle based controller thoughts/ideas....
« Reply #10 on: January 05, 2018, 08:17:15 AM »
Is it possible to have the beagle be on TOP of the board instead of on the bottom? I accomplished it by using standoffs from the CG1500 - and that works ok i'm just concerned about pressure points when connecting the euro-connectors into the board at times. I think the F4 is perfect being on top of the beagle, but the larger boards if the processor is accessable from the top (even if it is upside down). Now not bending the pins is another thing taking it on and off. Maybe someone could tinker up a 3d beagle extraction tool that pulls it straight up and off the boards!


The PocketBeagle based boards will all have the beagle on top.  The PB doesn't come with headers so we can do whatever we want with them.  However, it's actually HARD to get headers installed on top of the PB as the chip is too close to the headers so you need to find special (and expensive) headers to put them on top.   


I've attached 3D screen shots of rough ideas for the PocketScroller and F8-PB.  You can see the PB will need female headers mounted on the bottom of it so that it can sit on top of the boards.   I can utilize the space under the PB for some things, but obviously not for pixel connectors and fuses and such.   Also, there isn't enough space between the headers for most through hole chips to be used unless its parallel with the PB (like in the PocketScroller screen shot), which that really only fits one.  Thus, it needs to use the smaller surface mount things OR the board needs to be much bigger to move those chips out from under it. 

Offline Bshaver

  • Developer
  • ******
  • Join Date: Aug 2014
  • Location: Denver, CO
  • Posts: 1,426
  • Kudos: 22
Re: PocketBeagle based controller thoughts/ideas....
« Reply #11 on: January 05, 2018, 09:52:13 AM »
Hope this isnt a silly statement. Is there benefit to have a daughter board that accepts either the BBB, or PB and that connects onto the board. I know there are various GPIO restrictions, but is that possible One board - two possibilities with the proper adapter? Or is it make it much more complex. Heck for that matter, could not a PI flipped upside down or like the FPD be used. A pi could control 8 ports of a B-16 board. But when you need more, you put a Beagle on it....? Doable? Switchable with dip siwtches or jumpers based on the processor that is connected?

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
Re: PocketBeagle based controller thoughts/ideas....
« Reply #12 on: January 05, 2018, 10:12:41 AM »
Hope this isnt a silly statement. Is there benefit to have a daughter board that accepts either the BBB, or PB and that connects onto the board. I know there are various GPIO restrictions, but is that possible One board - two possibilities with the proper adapter? Or is it make it much more complex. Heck for that matter, could not a PI flipped upside down or like the FPD be used. A pi could control 8 ports of a B-16 board. But when you need more, you put a Beagle on it....? Doable? Switchable with dip siwtches or jumpers based on the processor that is connected?


The Pi cannot control 8.  I can just do two.  That's why the PiHat only has two outputs.   The ws2811 timing is fairly exact and cannot be "bit banged" out the Pi GPIO's reliably as the OS may interrupt things and disrupt the timing.   I believe the code on the Pi uses the two PWM pins to send out the data.    The Beaglebones use the PRU's to drive the timing and thus are immune (more or less) to the OS running on the main CPU.


An adapter card could be made to make the PB fit to a board that takes a regular beagle for our use case would be fairly easy.  However, it's actually not a cheap board to have assembled (would be ok to sell as a kit, assemble yourself).   Two 18x2 headers (for the PB), 2 23x2 headers (to mimic the beagle), a few parts for the USB.  The big kicker is all the through whole pins.   This is really the major downside of the Beaglebones over the Pi.   Elecrow charges 4 cents per through hole to solder.   So the four headers themselves would total $6.50 just to solder.  Add in the USB parts, the actual price for the boards, shipping, etc.. and it's likely over $10.   

Offline Bwinter

  • Sr. Member
  • ****
  • Join Date: Jul 2016
  • Location:
  • Posts: 434
  • Kudos: 10
    • First Show 2016
Re: PocketBeagle based controller thoughts/ideas....
« Reply #13 on: January 05, 2018, 10:49:41 AM »
Maybe I'm missing something here, but from what I read, a BB "Hat/Cap" would be able to provide 8 outputs (versus a PiHat/PiCap of just 2).


The cost is "likely over $10."  Yet, the PiCap is already $35.  Are there other reasons why a BB "Hat/Cap" hasn't come along yet (or has it, and I'm just not seeing it)?

Offline dkulp

  • Moderator
  • *****
  • Join Date: Sep 2013
  • Location: Framingham, MA
  • Posts: 1,418
  • Kudos: 77
Re: PocketBeagle based controller thoughts/ideas....
« Reply #14 on: January 05, 2018, 11:33:57 AM »
Maybe I'm missing something here, but from what I read, a BB "Hat/Cap" would be able to provide 8 outputs (versus a PiHat/PiCap of just 2).

The cost is "likely over $10."  Yet, the PiCap is already $35.  Are there other reasons why a BB "Hat/Cap" hasn't come along yet (or has it, and I'm just not seeing it)?


There have been several different Capes for the BB driving between 4 and 48 strings.   I *personally* have designed and ran group buys for capes that do 8 and 32 local with various options for differential expansions.   Previously, there were group buys for a 4 port cape and a 16 port cape.     So yea,  you missed them.   :)    That said, all the existing capes are for the full beaglebones.   What this thread is about is capes for the new PocketBeagle which is a much smaller/cheaper beaglebone.   The question above was what would it cost to create an adapter board to allow the smaller PocketBeagle to mount onto the existing capes designed for the larger beagles.   


I think one of the big issues is that there really isn't a store that carries any of the beaglebone capes (other than the Octo).   Thus, you need to get them via a group buy.   I wish a store would pick them up...  Quantity could drive the prices down a bit. If one of the store "owners" would like to carry them, have them contact me.   All my designs are public in my DropBox, although they now use the nightly builds of KiCad 5.

 

Back to top