Another pixel contr...
 
Notifications
Clear all

Another pixel controller

292 Posts
18 Users
0 Reactions
81.5 K Views
Posts: 313
Topic starter
(@ee351)
Reputable Member
Joined: 13 years ago
Bob Moody wrote:okay... who came by my house today and reworked my set up??? why do I ask ??
Becasue everything is working perfectly... So ... what did I change?

8604 is running 36 GECE's
EE_GECE_V3 is running 36 GECE's
EE_2812_v1 is running 40 WS2811's (only 36 defined in the sequence... )

I also modifed the LOR .05 sequence to make only 1 pass before it loops.
Running in LOR ... 0 missed pixels, 0 dropped/missed colors.

HOWEVER ... If i run the color change that is timed to the metronome ... there is dropped pixels and missed colors all over the place (technical term for all 3 controllers are affected)

I'm going to modify the color change beat sequence and resave it ...
I'm wondering a couple of things...
1. Did I finally stumble on the right power up sequence to get everything to enumerate properly?
2. Is LOR occasionally writing garbage (written as an XML file) when creating/saving the sequences?

Bob

I'm not sure about LOR, but this is how I do it:

1. Power on the controller
2. Power on the GECE pixels
3. Start sequence from xlights

I did the above steps every night for about 5 weeks and had zero enumeration problems. The code I wrote holds the outputs low until there is data to be outputted to the GECEs.

However, I don't think this will work with the SanDevices controller since it is always outputting data on the pixel outputs. The solution here is to add some pull down resistors, there's a thread somewhere on this forum about it.


Reply
Posts: 217
(@bob-moody)
Reputable Member
Joined: 18 years ago

I recall that discussion........somewhere...lol


Reply
Posts: 313
Topic starter
(@ee351)
Reputable Member
Joined: 13 years ago
Bob Moody wrote:I'm really thinking its LOR ...
Instead of watching the lights I ran the sequence and watched the colors blink on the screen.
I noticed that at the 12 second mark there was supposed to a WHITE event. Not only did it not engage the lights .. BUT .. it didnt blink on the screen !!!
It just flat out missed it ...

If I'm going to have be a ghost buster .. I want one of those cool plasma guns mounted to the backback ...

Bob

Is there any way for you to test your metronome sequence in another sequencer, like vixen or his?


Reply
Posts: 217
(@bob-moody)
Reputable Member
Joined: 18 years ago

I do have the Vixen software .. fairly current I think..
I'll have to load it up, configure etc etc .. oh yeah .. and learn how to create sequence .. LOL..
But .. yeah .. I suppose I could ..

Bob


Reply
Posts: 217
(@bob-moody)
Reputable Member
Joined: 18 years ago

Going to need a little help setting up the strings in Vixen.
I "THINK" I figured out the controller set up .. I have 1 SD6804 with 4 outputs, 1GECEv3 with only 1 output defined, and 1 WS2812v1 with only 1 output defined.
How do I define the 36 pixel GECE string then patch it in?
If there is a How-to, video or directions somewhere I will read on my own .. but if someone has a quick answer .......

Bob


Reply
Posts: 217
(@bob-moody)
Reputable Member
Joined: 18 years ago

Got the SanDevice Working ...
Figuring it out .. disregard previous post ..


Reply
jchuchla
Posts: 8698
(@jchuchla)
Illustrious Member
Joined: 15 years ago

There is a sample config write up specifically for the sandevices controllers on the vixen website in the documentation section.

--Jon Chuchla--

Sent from my iPhone using Tapatalk


Reply
jchuchla
Posts: 8698
(@jchuchla)
Illustrious Member
Joined: 15 years ago

And I don't mean to pawn it off. But it's one of those things I've typed so many times it's just best to refer to the polished version.

--Jon Chuchla--

Sent from my iPhone using Tapatalk


Reply
Posts: 217
(@bob-moody)
Reputable Member
Joined: 18 years ago

Okay ... I have the SD-6804 and the EEGECEv3 boards configured and running in Vixen 3.1u2 (64bit)

I ran chases down both .. and while the SanDevice seems to be MAYBE... a half a pixel ahead, I cant say that it really is because I'm still stumbling around with how to put things on the grid.. I know they both started at 00.00 and ended at 3.2000
However when I dropped a nutcracker bar effect on the grid it looks pretty darn right on the money ...

By the way .. it wasnt until I dropped the 36 pixel chase to about 1.5 seconds that I started noticing drops and yes, they occured in both strings ..
From about 2 seconds on up .. rock solid (except it for as I mentioned.. the SD seemed about 1/2 a pixel ahead... if you werent staring at it .. you may not even see it..

Maybe Jon could send me a test sequence in Vixen ??? (since he IS THE VIXEN GURU...)

But first impression.. the EEGECEv3 controller is right there with my 6804 as far as keeping up.

Bob


Reply
jchuchla
Posts: 8698
(@jchuchla)
Illustrious Member
Joined: 15 years ago

Vixen isn't going to give you frame accurate tests. That's kinda the point of vixen.
I could go into detail if you like. But the basic premise is that the sequence is decoupled from the controllers. There's an execution engine that receives input from various contexts. The sequence editor being one of them. These instructions from the contexts are in the form of linear intents. E.g fade from color x to color y over z time. The execution engine is freewheeling with respect to the contexts. The engine samples the intent objects in memory at an interval that's somewhat variable. The controller modules actually drive the update at the rate they want to see. But the engine has a rate limiter to keep it all in check. With default settings the engine will sample the intents at an interval of roughly 50ms (+- 1ms typically).
The bottom line is that the v3 engine will update faster than actually matters and adapts to what the system can handle. But it's not really suitable to use as a test tool.

--Jon Chuchla--

Sent from my iPhone using Tapatalk


Reply
Page 28 / 30
Share: