Weird data signal c...
 
Notifications
Clear all

Weird data signal corruption issue...

9 Posts
3 Users
0 Reactions
2,324 Views
Posts: 108
Topic starter
(@irishale)
Estimable Member
Joined: 12 years ago
[#44262]

I've got a strange data corruption issue. I've made a pixel star, 8 radial arms pvc, and am using ws2812 30/30 ribbons 18 pixels each, sequential, so the tip of an arm is wired to the base of the next. All has been well, until today... the first pixel of the second arm started acting up, which of course gives me really interesting color patterns.

In diagnosing it, I broke the data connection and reconnected it. It will work properly so long as I am holding the bare wire splice in my fingers. The instant I let go of it it goes back to a misfire.

What in the world is going on??


8 Replies
Posts: 451
(@utimgr)
Honorable Member
Joined: 14 years ago

Your changing the characteristics of the signal when you fingers are touching the wire(s). Is it only that data wire, or the power and ground your holding? I'd start with a cap across the power and ground, at last a 10uF (voltage applicable to the system). See how that might help. Putting a cap (smaller...say .1uF) on the dataline, would use the impedance of the cable to slow down the rise time


Reply
Posts: 236
(@afremont)
Reputable Member
Joined: 10 years ago

You didn't say how big your star is or what shape it is, but I take it that your data wires are longish. Maybe picking up RF noise like an antenna from something in the vicinity and holding it detunes it from it's resonant frequency. Is it a unique length, or do you have others the same length?


Reply
Posts: 108
Topic starter
(@irishale)
Estimable Member
Joined: 12 years ago

I'm only touching the signal wire... I'll try that in a bit.


Reply
Posts: 108
Topic starter
(@irishale)
Estimable Member
Joined: 12 years ago

star has 18 pixels per arm of a 30per m WS2812 ribbon... plus a little in the center for the cross... it's 4'-ish from tip to tip. It's basically an 8-pointed asterisk. no data wire is longer than 2' without hitting a pixel, and this is the only arm doing it. It's consistently "wrong" after the first junction.


Reply
Page 1 / 2
Share: