Comments

• Chris August 5, 2015 at 9:22am

  @Mirko

  Ah, so the image has been updated too (I should have noticed), I was using one from a week or so back, that explains it.  I got confused which sd card I was using and overwrote the previous weeks work.  Since I'm starting from scratch again I'll download the new image and try again.

  I'll throw it on the scope, so if it is 3.3V then I just do as the readme.md says (bridge R4 and leave R5 alone).

  No probs, I found out after I cut the header pins.  Was funny as I tried to fit it at first and double checked the listing.  Was easily fixed and I'll use the 2 pin ones somewhere in the future.

  Thanks

  Chris

  

  App Showcase
• 

  Mirko Denecke August 5, 2015 at 8:41am

  @Chris

  Fixed P7 and P10, thanks.

  I don't know the R815X, maybe you can measurer the voltage with a scope.

  If you use the Image from the README.md additional repos are not necessary . Make sure you execute sudo apt-get update before try try to install the Kernel.

  

  App Showcase
• Chris August 5, 2015 at 6:37am

  Hi,

  Next question, do I need to add an additional repos for the latest BBBMini README.md?

  Just that on step 11, apt-get cant seem to locate the RT kernel....

  Also on step 13 it cant locate ti-pru-cgt-installer

  Are you able to advise which repo's I should add.

  Regards,

  Chris

  

  App Showcase
• Chris August 4, 2015 at 5:32pm

  Hi,
  
  This is for anyone who has assembled one of these or Mirko, I'm in the middle of doing one up.
  
  For the resisters on the RCIN line, in the spec document it puts the 2k and 1k to drop the 5V to 3.3V. If I'm using an orangeRX R815X, I believe the signal is already 3.3V so do I just bridge R4 with some solder or wire, this info came from: https://github.com/mirkix/BBBMINI-PCB/readme.md
  
  Also I believe P7 and P10 should be 1 x 5 header row.
  
  Chris
• 

  Víctor Mayoral July 14, 2015 at 3:41am

  Thanks Mirko,

  Will have a look at your Device Tree definitions.
• 

  Mirko Denecke July 13, 2015 at 12:37pm

  @Victor

  It is not possible to use the BBBMINI Debian Image with the PXF because there are some difference, i.e. the SPI interface of the BBBMINI use Kernel CS lines, PXF use custom CS lines. So it is necessary to modify the Device Tree if you want to use the BBBMINI Debian Image with a PXF.
• robert bouwens July 13, 2015 at 6:19am

  why switching the forum?

  redhut spend a large amount of money to get preempt_rt running.

  osadl

  a simple test from your side shows the opposite.

  to me a surprise.

  and pls. show the command line for the cyclictest.

  

  Latency plot of system in rack #7, slot #6: OSADL - Open Source Automation Development Lab eG

  OSADL promotes and supports the use of Open Source software in the automation and machine industry.
• 

  Víctor Mayoral July 13, 2015 at 2:35am

  I'm happy discussing this matter but let's bring it out of this group dedicated to BBBMini. Can you explain your reasoning at forum.erlerobotics.com?

  Thanks,
• robert bouwens July 13, 2015 at 2:32am

  hmm, questionalble.

  how many test cycles where used to 'verify' the latency?

  we always use the PREEMPT option for the vanilla kernel.

  but from what i read  i assume you are wrong when looking at the latency.
• 

  Víctor Mayoral July 13, 2015 at 1:22am

  @Robert,

  All our kernels include preemptible capabilities and we've tested/compared PREEMPT vs PREEMPT_RT several times. For now we are releasing kernels with the PREEMPT option since our tests keep actually showing that they outperform PREEMPT_RT patches. Refer to https://github.com/BeaglePilot/beaglepilot/blob/master/files/APM_Li... for a more detailed description.