Author Topic: Random trash data outputted  (Read 23553 times)

Offline drlucas

  • Developer
  • ******
  • Join Date: Sep 2013
  • Location: Whitby, Ontario - Canada
  • Posts: 1,422
  • Kudos: 3
Re: Random trash data outputted
« Reply #15 on: December 10, 2014, 06:54:25 PM »
I tried the fpdskip branch and made no difference to me. maybe your milage will vary.
Ryan

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #16 on: December 10, 2014, 07:10:16 PM »
On my schedule for tomorrow morning then.

Will advise after that.

Offline Steve Gase

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Georgetown, TX (near Austin)
  • Posts: 1,037
  • Kudos: 5
    • WinterLightShow in Georgetown, TX
Re: Random trash data outputted
« Reply #17 on: December 10, 2014, 07:21:43 PM »
my apologies... i've not been very helpful since i first tried the fpd firmware fix1 and fix2. 
work, finishing sequences, redoing my web site, fixing my radio transmission, etc.

i will try to reproduce this tomorrow in a smaller setup that doesn't impact my show.
http://WinterLightShow.com  |  110K channels, 50K lights  |  Nutcracker, Falcon, DLA, HolidayCoro

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #18 on: December 10, 2014, 07:31:23 PM »
What, no guts, no glory. LOL

The only way to see some of these glitches I believe will be under the full load of a show.

I just need to get up early enough to be able to see the lights before the sun washes them out.

I sure hope the PI code upgrade at least slows down the random glitching. People may start to think that that is the way it's supposed to be.

I also am having audio problems with my transmitter, I am getting severe distortion of my audio stream. I don't think its with the PI because I am monitoring the PI output through my mixer and the audio sounds good there, it is just the sound through a receiver that does not sound so good. I hope it is something as easy as a bad cable and not my transmitter.

Offline Steve Gase

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Georgetown, TX (near Austin)
  • Posts: 1,037
  • Kudos: 5
    • WinterLightShow in Georgetown, TX
Re: Random trash data outputted
« Reply #19 on: December 10, 2014, 08:30:51 PM »
load the whole show, yes.  but i should only need a string or small matrix listening on the same problem channels on the back side of a universe to see the issue... if we understand this correctly.

if I can recreate under that condition, i will create a sequence (with my full channel range) that has no sequence activity (no effects selected).  I had done that by mistake earlier... and I saw the random pixels dance when nothing in the show should have been dancing.

If I have have no sequence data, i expect the file will contain a lot of 0's -- I should be able to view the data in the file and confirm there are only zeros after the header  (to prove its not a NC problem).  If I play it -- and still get the randomness to proves its a problem with something other than NC.

There are other tests to confirm Pi, rule out Falcon16/8, ...narrow to FPD, rule out E1.31...

...i know, talk is cheap... get off my #$@ and prove it! :)

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #20 on: December 10, 2014, 08:47:19 PM »
My glitching is not limited to either Pixelnet equipment i.e. Falcon 8/16s, uSSCs on Falcon hubs but is also showing up on my DMX Lynx Express and Lynx SR4DMX stuff but appears all over the system.

I believe that this pretty much points to either the PI or FPD code.

Now to look back this same PI and FPD hardware with the old code worked without any glitching last year. Now I don't want to go back because that would be even more of a nightmare.

I made some changes just before my show tonight and didn't get a chance to reboot either the PI or FPD so my testing tonight my not have been complete.

I will check the forum first thing in the morning in case there is any code changes for testing.


Offline CaptainMurdoch

  • Administrator
  • *****
  • Join Date: Sep 2013
  • Location: Washington
  • Posts: 9,181
  • Kudos: 180
Re: Random trash data outputted
« Reply #21 on: December 10, 2014, 09:34:33 PM »
Are any of you running a model B and want to be a guinea pig to test reverting the kernel to the previous version we used prior to the audio firmware fix?  On a B, we were running an older firmware/kernel combination which didn't have the bug.  That older kernel also did SPI differently, so that is something else I want to test.

It's a simple 1-line change in a script to tell the Pi that it needs a different kernel, then reboot and it will automatically copy the older kernel into place.   If the problem doesn't go away then it's a simple reversion of that 1-line change and reboot and you are back where you started at with the v1.0 SD image.
-
Chris

Offline Steve Gase

  • Supporting Member
  • ******
  • Join Date: Mar 2013
  • Location: Georgetown, TX (near Austin)
  • Posts: 1,037
  • Kudos: 5
    • WinterLightShow in Georgetown, TX
Re: Random trash data outputted
« Reply #22 on: December 10, 2014, 09:38:56 PM »
I plan to swap out my model B for my B+ in the show...  with the B i was going to try reproducing the issue... once I have the issue reproduced, i can try your procedure, Chris...  please provide details.

others:  please don't rely on me to test this... work may get in the way of my playtime. :)

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #23 on: December 10, 2014, 09:44:36 PM »
I couldn't stand it, I just couldn't wait.

I went up and tried both the old and new FPD code with the latest version of the PI download.

Sorry to say didn't fix it.

I am running a plain ole B.

Anything to get things flowing smoothly again. Does that mean that I have to convert all audio over to OOG?

Oh yea, now I am going to go hit the sack.

Offline CaptainMurdoch

  • Administrator
  • *****
  • Join Date: Sep 2013
  • Location: Washington
  • Posts: 9,181
  • Kudos: 180
Re: Random trash data outputted
« Reply #24 on: December 10, 2014, 09:54:24 PM »
Anything to get things flowing smoothly again. Does that mean that I have to convert all audio over to OOG?

No, this is just a change to change the model B to use the kernel/firmware version that was included with FPP v0.4 instead of the updated kernel/firmware in FPP v1.0.

If you want to test, you just need to login to the Pi and edit /etc/init.d/checkkernel.  Here's a simple walk-through.

sudo nano /etc/init.d/checkkernel

# Scroll down until you find the following section:

Code: [Select]
        # On 'B' model because we have 26 pins on the header
        PIVERSION="B"
        NEEDKERNEL="3.12.29"

Make sure that you are the model B section, not the B+ which is above it.
Change that 'NEEDKERNEL="3.12.29"' line to this:

Code: [Select]
        NEEDKERNEL="3.10.33"

Hit 'CTRL-O' to save and then ENTER to accept the filename.  Then hit CTRL-X to exit.

Reboot the Pi and the checkkernel script will detect that it needs to be using a different kernel and it will copy the original firmware/kernel into place and will automatically reboot again.  After the second reboot, the Pi will now be running the 3.10.33 kernel and older firmware.  You can now test to see if the issue is still happening.  The FPP code is still the same, you're still running v1.0 or the fpdskip branch if you were on that, only the firmware and kernel have changed.  This will help narrow down to eliminate one more variable.

To revert, follow the above procedure, but look for 'NEEDKERNEL="3.10.33"' and change the 3.10.33 back to 3.12.29 to match the above line in the B+ section.  Reboot and it will update back to the new kernel and reboot automatically and then you'll be back where you started at.

Offline drlucas

  • Developer
  • ******
  • Join Date: Sep 2013
  • Location: Whitby, Ontario - Canada
  • Posts: 1,422
  • Kudos: 3
Re: Random trash data outputted
« Reply #25 on: December 11, 2014, 03:06:55 AM »
I've moved to a b+ so I can't easily test.  I do notice though that I sometimes see a glitch with no show running. I have that always output to channel option checked so maybe it's interfering ?? Is there anyway that is pushing garbage when the show is running? I may try to uncheck that box and see if there is any difference.

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #26 on: December 11, 2014, 05:31:42 AM »
Sorry that did not make any difference, the random pixels and blinking still occurred.

If you want to talk direct I can PM you my phone or you can get it from David.

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #27 on: December 11, 2014, 12:00:16 PM »
Hope the Capitan didn't over sleep today. LOL  ;)

Offline CaptainMurdoch

  • Administrator
  • *****
  • Join Date: Sep 2013
  • Location: Washington
  • Posts: 9,181
  • Kudos: 180
Re: Random trash data outputted
« Reply #28 on: December 11, 2014, 12:16:59 PM »
OK, I have two more things to try since the kernel revert did not fix the issue:

1) I changed the skip threshold to 1ms instead of 2ms in the fpdskip branch.  If you are on that branch, can you pull in the latest changes and then turn on 'Debug' level logging with the 'Channel Outputs' mask checkbox checked and see if the problem still occurs or if it occurs less than it previously did?  If we are sending data too frequently when trying to speed up, then this will cause us to skip more which might make the issue better.

2) I want to make sure that data is not being corrupted somewhere inside the player before it is written out.  I am going to refine a debug patch and commit it to the fpdskip branch to help test.  The patch looks at the data read in from the sequence file and compares it to the data we are sending out to the FPD to see if there are any differences.  This will tell us if there is an issue in the player code overwriting data somewhere.  I ran this patch for 16+ hours on a Pi and did not see any cases where it logged that data was being corrupted, but I didn't try a bunch of options and also wasn't testing with an actual FPD.  I will push this patch in tonight and post here so that you guys can test and see if there is an issue internal to the player.

Offline MrChristmas2000

  • Falcon Beta Team
  • **
  • Join Date: Apr 2013
  • Location:
  • Posts: 481
  • Kudos: 4
Re: Random trash data outputted
« Reply #29 on: December 11, 2014, 12:33:18 PM »
Educate me a little.

I am not clear on what you mean by being on the fpdskip branch. Where is it and how do I get there?

My unix is more rusty than a Detroit steel mill. :)

 

Back to top