Pixels Strings going all white

[Wygantmb]

I asked this question on Facebook and after trying all the replies am still having this issue. Here's the issue. Most of my sequences will cause my mega tree to go all white once I put any kind of  large demand on the controller such as fast flashes or all lights on. I've tried jockeying with the power supply voltage, I've made large changes to the sequences with less demands and still no luck. A little back ground. The controller is a Pixie 16 V4. Its powered by Twin 12v x 30amp power supplies. I'm using a USB485HS adaptor and the controller is on its own 500k enhanced network. 15 strings with 90 pixels each. Nothing else on the controller. When I consulted LOR they said it was weird and to just make sure that my HU setting were the same as the other 3 Pixie16 controllers in my show that run perfectly. I have a two spare controllers and I'm about to try a swap out but I'm concerned it might be something else. I'm going to try the controller swap tomorrow but until then if anyone has any ideas, I'm all ears.  Thank you

[TheDucks]

What is the distance from the pixie to the first node on each string?  What are you using for wire?

Power and CAT5: How far apart are these?

Ah! your power supply is flailing.. 90px = 1.8*3= 5.4A(90px) * 8ports = 43.2A (all white)  (I estimate 3A per 50px)

You need to seriously power inject  (another 2 supplies)

 

[k6ccc]

I'm guessing a voltage problem.  You have 1350 pixels.  At full white, they are likely drawing about 60mA each which comes out to 81 amps.  That is a lot more than your two 30 amp supplies can deliver.  If the Pixie controllers have the ability to dim all pixels by a percentage, try dropping it to around 50 % and see what happens.

I ran into the same issue (on the engineering side before construction) with my new 24 x 100 pixel tree.  I have eight strings on each of three power supplies and have to run 50% to keep the peak load withing the limits of the supplies.

TheDucks just beat me to it...

 

[TheDucks]

Jim just confirmed my educated guess

[Orville]

Also can't the Pixie 16 run at 1,000K, so why 500K?  Seems the faster speed would also help in some of that possibly.

[Wygantmb]

I understand the power issue that is brought up. Rather then injecting I think I'll break the tree in half and use two controllers and two full sets of power supplies. It'll be much faster then adding 15 more sets of wires. My demand was never to go full white. I'm not sure how or why even with a power drop would they go full white and remain full white until I physically disrupt the power to the controller. That doesn't make any sense to me..

[TheDucks]

37 minutes ago, Orville said:

Also can't the Pixie 16 run at 1,000K, so why 500K?  Seems the faster speed would also help in some of that possibly.

IMHO it would make it worse. The issue is rapid changes in power draw, not giving the Caps in the PSU's time to maximum charge (there was the math back in ET school that proved that a Full charge take an infinite time)

[TheDucks]

2 minutes ago, Wygantmb said:

I understand the power issue that is brought up. Rather then injecting I think I'll break the tree in half and use two controllers and two full sets of power supplies. It'll be much faster then adding 15 more sets of wires. My demand was never to go full white. I'm not sure how or why even with a power drop would they go full white and remain full white until I physically disrupt the power to the controller. That doesn't make any sense to me..

2 more PSU's and wire is a lot cheaper than a Pixie16

[Wygantmb]

1 minute ago, TheDucks said:

2 more PSU's and wire is a lot cheaper than a Pixie16

This is a very true statement. I only say that's my plan because I have 3 extra setups ready to go. Its faster and the reprogramming is easy and can all be done in the warmth of my living room. Well minus the controller. Thats outside ;-)

[TheDucks]

4 minutes ago, Wygantmb said:

This is a very true statement. I only say that's my plan because I have 3 extra setups ready to go. Its faster and the reprogramming is easy and can all be done in the warmth of my living room. Well minus the controller. Thats outside 😉

There is no programing to inject. You take a PSU to a fuse block (1 per leg) , run a pair of wires (Bk, fused-Red) and attach them to the string  between px50 and 51 (since the PSU is not the same, cut ONLY the +12 from px 50 and feed px51 from the new PSU. The common and data STAY CONNECTED.

[Orville]

1 hour ago, TheDucks said:

IMHO it would make it worse. The issue is rapid changes in power draw, not giving the Caps in the PSU's time to maximum charge (there was the math back in ET school that proved that a Full charge take an infinite time)

I remember that, but sometimes something might sound like it'd work {in theory I guess}, but when put to practical use failed miserably.   I know been there done that too. 🤣

[Wygantmb]

You guys are 1000% correct. I've only run one song on reduced power and it ran flawlessly. Its been a busy day so I haven't had the chance to run my power injection yet. And Yes I've decided to inject and save my other controllers for something else. I didn't want to spend the time outside to cut and splice my strings. I have tons of 3 pin connectors so I just made some inline connecters that inject the power after the first string of lights. For anyone that has ever added a plug and play trailer harness its kind of the same concept. Grounds are tied together but the POS+ is separate. I'll get all that installed this weekend, at lest for now I know I can start my show tomorrow night just at reduced power. Thanks again everyone. Happy Thanksgiving.