Strange Green Flashes

[lboucher]

Hi All

I have no idea if this is a Zeus issue, but figured its a good place to start.
This year I replaced my mega tree, getting the IP68 nodes, and switched over to two 8 controller zeus instead of SSCs.
Everything in general is working fine.
Show is over 30 minutes and the tree looks fine at all times except this one slow fade/modulation to voice at the end of a short sequence. At this particular point in the sequence it should only be blue, but half the tree seems to randomly flicker green.... The front half of the tree is on one zeus and the back end is on the other. It looks like its just the back zeus controlled strands, but maybe not exactly.

I'm using that other hardware to actually send the control data, but I did try Vixen from my laptop and had the same problem.
I guess my next thing to do would be swap the zeus and see if the problem follows the zeus, but I have zero want to touch the show now that its all set up and running and were a few days till Christmas.

Any other ideas?
Pretty sure the sequence is fine, and it is attached for review.

Here is video of it
You are not allowed to view links. Register or Login

As I look at it more, maybe there is an intermittent short or the power supply crapping out.....
But I swear every time I watch the show it seems fine everywhere else...
I will have to watch closer.

Thanks

[lboucher]

Ok So now I am really really confused.

Monitored the voltage on the zeus, didn't vary by more than a tenth of a volt.
Tried a different power supply anyways, no difference.
Swapped RS485 chip with working zeus, no change. (Issue did not follow the RS45 chip)
Swapped one of the pics with the working zeus, no change. (Issue did not follow PIC chip)
Swapped Ethernet cord and plugged into a totally different boards pixelnet out, no changed.
Tried swapping Pixel Strings, issue did not follow the string.
Half removed the board from the enclosure, enough to take an alcohol brush to the back of the board to clean it off, no change.

The only thing that made a difference, was to create a brand new sequence, and copy the data from the original into the new one, this got rid of the green flashes oddly enough....
The only difference between this sequence and the original is that the rest of the sequence besides the mega tree is blank.
New sequence is attached.

The only thing left is to try replacing the crystal oscillator on the board... The timing being off would make some sense.

I sat and watched the tree very closely for a while, and I can say the back half of my tree on that zeus lags the rest of the show sometime, but not all the time. There are very complex patterns that work fine. But half of the time when the entire tree turns on/off there is a noticeable lag on the back half of the tree. Not all the time, just sometimes....

This has to be some odd timing issue with reading in the data. Unfortunately I don't have another oscillator to test with.

Thoughts?

[tbone321]

Everything is telling you that the issue IS with the sequence, not the controller.  If the crystal was the issue, none of the channels would work. 

[lboucher]

Everything except the fact that I have been using the same sequence for years.
Its true I added channels, but the mega tree has been around since 2011, see here:
You are not allowed to view links. Register or Login

At that time it was running on SSCs.

I do have 6 half built SSCs downstairs I could go try to finish, but tomorrow is Christmas Eve, not worth the effort at this point.

For fun I just created a brand new Vixen File and did a select all from the old file, and copied into the new.
Converted to fseq and uploaded.
Will see in half an hour when it plays that.
Other than that, I'm not sure how to fix a sequence file that doesn't look broken...?

[tbone321]

The minute that you added channels, you changed the sequence.  If you still have a copy of the original one, try running that one sometime after the holiday and see what happens.

[lboucher]

Finally caught the "new" sequence, it still has the issue.While I agree the sequence is technically changed, its a digital setup. I should be able to wireshark the data and see the same exact data on the ljne, just with added channels. In any case, if adding channels actually breaks things like this, then I have no way to fix it, and there is an inherit flaw in the data stream that needs fixed...

Thanks for the thoughts though.

[tbone321]

While it may be "digital", it really depends on what you are using to create the show.  Some software will recreate the show file when you make a change and your change may have altered some other effect that is causing these green flashes.  What software are you using to create the show?

[lboucher]

Vixen 2.1.4