Smart string programming ghosts - figured out - well aparently not - now I did!

[lonewolf41]

Will re-download and flash again tonight.  As a side note, I hooked up two more strings at lunch.  One V3 and one V1.  Both were programmed to 1973 the same as the house4 string.  Both new strings behaved identical to the existing house4 string.  Turning on channel 6144, turned on both 6144 & 6145.  So I doubt it is a SSC issue as it is probably a little TOO coincidental that both V1 & V3 would behave the same way.  I'll redownload and reflash the ED tonight and see what happens.

I also tried to hook up my pixelnet dongle and set the channels to the second universe (4097-6400) but it kept throwing a COM4 error and would never turn on any lights.  I had everything set correctly so don't know why it wouldn't work.  May try again, but don't know what is wrong.  I used this dongle to run my show last year.

Thanks again,
-Keith

[jnealand]

Found my problem, but it turned out to be a bad node.  Once I replaced it all was well.  Still have a couple of other bad nodes, but they are not killing half a string so will work on those after I get everything else cleaned up.

[lonewolf41]

I downloaded the firmware in the wiki.  Erased my ED.  Flashed it with the just downloaded firmware.  Nothing changed.  Until I can get my profile restructured, I just inserted 2 blank channels and have one "dead" node when the string is in pixel mode.  It works in string mode OK.  I can live with that until I can restructure my profile to move that string past 6144.  Still wonder why I am the only one seeing this, but at least I can work around it.

Thanks,
-Keith

[lonewolf41]

OK, now this is getting uber-nuts.  I added 80 extra channels to my profile and put them before my house4 string.  That moved them to start @ 6150 in Vixen (programmed to 2054 via SSC utility) now over the 2048 mark.  Now the string still works fine in string mode, but it stops at 6399.  Channels 6400 and up do not come on by themselves.  So node 83 blue through the end of the string do not light except in string mode.

And some other odd thing with this string, when I program it, the last 4 nodes do not flash nor do they work during the show; however, sometimes, they just come on by themselves.  I woke up this morning and the last 4 nodes on that string were sitting there blue.  I think they were red last night.  I am just going to leave the one node @ 2048 dark and go on with the show.  Hopefully it will not be too noticeable.  I would just like to understand what is going on.

[zwiller]

OK.  Now this sounds EXACTLY like the ED firmware issue to me and Chris agrees.  Keith, are you ABSOLUTELY SURE you are using the file from the wiki when programming?  Maybe delete the previous file before your download again. 

It just sounds strikingly similar to the issue I had which was solved with the correct firmware.  I totally understand your frustration.  I was there a few weeks ago myself... 

[lonewolf41]

I think it was Monday that I downloaded the file circled in the attached picture.  It was EtherDingle2.hex and I picked the just downloaded file for programming.  Is this the correct file?  I can download on a completely different computer and try it.  It would be nice if there was a checksum on these, but they all come up 0000...which I assume is not the actual/correct checksum.  I believe the file I downloaded was 1,424KB if I remember.  I'll try downloading on a different computer and put it on a thumb drive and see if that makes a difference.  Don't know what else to do/try.

Thanks,
-Keith

[rm357]

These systems of system can be tricky to diagnose...

The problem could be in vixen or the EtD firmware.

Can someone test this with xlights and/or LSP to see if they have an issue at address 2048?

I used LSP last year and don't remember seeing anything like this with either EtD firmware. The boundary pixel issue gave me one bad pixel where nodes were split across two 512 address blocks, but the pixels on both sides of the bad node still behaved correctly.


That leads me to believe the problem is in vixen.

[lonewolf41]

I have x-lights on a computer @ home that I can test with.  Can someone give me a crash course in how to set it up to test this part of my display?  Even just using the test function should tell me whether or not it is Vixen.  I used it to test my smart strings for burn-in, but am not sure how to set it up for the higher channel counts.  The string in question starts @ ~6069 in Vixen.  The controller is set to channel 1973 on universe 2.

Thanks,
-Keith

[chrisatpsu]

you'd need to set it up to run a sequence, since i believe the test stops at a "boundary" anyways.


I'm soon to start my show, sequenced in vixen, and i have the adress you're having trouble with in the middle of one of my strings on my mega tree.

[zwiller]

Just got my conductor fixed and plan to run show with it tonight for first time ever .

Now that I have some time fool around, I will test as well. 

[lonewolf41]

Well let me know what you guys find.  If i can get my last sequence finished this weekend, I may play around with it some more as well.  Try out xlights running the show.  Run it with conductor, etc.

Thanks,
-Keith

[lonewolf41]

OK, I have finally narrowed this down to ED firmware.  The exact same profile and sequence that produced the wrong colors with the ED look fine when running from the conductor.  So the exact same profile and sequence that looked wrong when running in Vixen, looked fine after running through the xlights conversion and running on the conductor.  Conversely, the profile that I used to run my show that skipped the channels @ 6143 & 6144 and looked good..except for the one node that would not light...looked wrong when running on the conductor.

So it looks like there is some kind of a bug in the ED firmware at channel 6143/6144 (2048 of the second universe).

Thanks,
-Keith