Notifications
Clear all

E682 issues

33 Posts
9 Users
0 Reactions
10.7 K Views
ebrady
Posts: 670
Topic starter
(@ebrady)
Prominent Member
Joined: 17 years ago

Thanks Jon,

I don't believe the timing of the data stream causes this issue, it is due to how power is initially getting applied to the pixels. One thing I noticed in my "old" setup, was a number of pixels would immediately jump to a random color once I switched on the power supply on. Once the E682 booted, some of the pixels would correctly reset, but others would stay stuck. I am guessing this slow power ramp followed by the overshoot "glitch" caused some of the pixels to stop responding completely. I know GE recently changed the power supplies on their newer GECE pixels, I wonder if something in the design may have changed also to make it more sensitive to this type of power transient.

To correct things on my end, I made a simple change. I moved my On/Off switch from the 120V side of the power supply to the 5V side. This means that the power supply will come on immediately when I plug it in, but the 5V feed to the pixels can be switched on after the supply has come on and had a chance to stabilize. Works like a charm! 🙂


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

This is an odd one. I wouldn't expect a power swing like that to cause these problems. The stock GECE power supplies are 6v so I wouldn't think the .3 of overshoot would hurt you.
That rise time doesn't seem that drastic either. How does your scope trace look on a stock GECE power supply?

Are you noticing the stuck pixels start at a certain length down the string? I'm thinking after some wire length has further distorted the signal/voltage. I'd wonder what the trace looks like at that point.
Also, was that trace with or without the pixel load?
What kind of power supply is it? How far from the first pixels? What kind of wire in between?

--Jon Chuchla--

Sent from my iPhone using Tapatalk


Reply
ebrady
Posts: 670
Topic starter
(@ebrady)
Prominent Member
Joined: 17 years ago

The GE power supply is 6V, I think it drops to 5.2 or 5.3V when it exits the stock controller, but I can't remember. Regardless, I don't think the .3V overshoot is causing the problem. I am guessing the "slow" rise time from 0 to 5V is.

The scope trace was with one string of 25 pixels, connected directly to the E682. On the bench, I was able to get 1 pixel to lock up frequently, which means it would come on when power was applied to the power supply, and would not go off after the E682 booted. As more load is added (in my case 16 strings of 49) I would expect this rise time of the power supply voltage to get worse and stretch out further. The behavior I saw on the tree correlates with this, up to 16 pixels would take on a random color when the power was switched on, and of this number 75 - 80% of them would return to normal operation after the E682 booted, the rest would remain "stuck". There has never been an instance out of 40 or so power cycles were none were stuck on. When I get a chance, I may drag the scope out to the tree and take a measurement with all 32 strings attached.

On the cable run lengths, they vary from 5 feet to 20 feet as they go around the tree. Interestingly enough, the stuck pixels always happen on the strings with the SHORTER run lengths (5 to 12 feet). I have never seen it happen on the runs of >12 feet. Thinking out loud, I am not sure that if it occurs on the shorter lengths, it proves or disproves the slow rise time theory (maybe rise time is to "fast" instead). Cable runs for power, ground, and data are done with 16 guage stranded wire I found it to be much easier and cheaper to buy a couple of 100 foot, 3 conductor extension cords and use it for the string to controller connections. 16 guage would minimize voltage drop to boot. These cables terminate to waterproof connector pigtails on both ends, all joints are done with waterproof but splices that are then grouped together and heat shrunk.

The tree is configured with 16 strings of 49 pixels (two strings of 25 spliced together). Stuck pixels always occur between pixels 1 and 15 and have never occurred on 16 - 49. This seems to also correlate with the run length behavior mentioned earlier.

Attached is a picture of the power supply - 5 Volt 60 Amp. I can't remember who I purchased it from as it has been sitting on a shelf for 1 1/2 years now.

power supply

Steady state output is 5.05V
As mentioned earlier, if I let the power supply come up first, before applying 5V to the GECE pixels and E682, it all works fine. No stuck pixels in at least 25 power cycles now.

Ed

One other note, when a string has stuck pixels, the other pixels still appear to be addressed properly and not shifted.


Reply
ukewarrior
Posts: 4674
(@ukewarrior)
Illustrious Member
Joined: 18 years ago

Wow.
this is a good thread on some GECE trouble shooting.
Thanks for documenting it Ed.


Reply
ebrady
Posts: 670
Topic starter
(@ebrady)
Prominent Member
Joined: 17 years ago

Thanks David
Since this appears to be power supply related, my thoughts are that I may not be the only one who has or will experienced this. Hopefully this thread proves to be of benefit.


Reply
Page 4 / 7
Share: