Neopixel Static and Wonky color changes

[scott]

Built my first two neopixel blades this weekend and am connected to a Prism 5.5. 26 awg wire from the pins and 24 closer to the board. I tried on blade with the SMD resistor and a second with an in line resistor. In both cases, I am good with solid blue and green, but experience speaker distortion with reds and mixed colors (so likely any time red is invoked.) Also, randomly, the blade turns to rainbow color like a weird lollipop. I am using the Panasonic 3.7V cell from TCSS and all components are TCSS. I followed Madcows video explicitly while wiring both the hilt and the blade sides and was very confident in all connections. I feel confident after extensive searches that there isn’t a thread covering this issue..some similar, but hoping someone has experienced, and solved a similar issue....read several of Tom T’s posts which were helpful, but still stumped.

[Tom Tilmon]

Clean data pads, clear them of flux. You have corrupted data or bad strips. I have found that leftover flux will cause massive data corruption. You must clean off anywhere you have soldered a data line onto with PCB Solvent. Blade Adapter, Hilt adapter, Strip Data, and SOund BOard. Clean the mess out of them.

[scott]

Thanks Tom - that certainly seems reasonable - I will try cleaning the LS pad and re-soldering the data line to the board since I experienced the issue in both blades that I built (and pop your video on the monitor while I do it...very appreciative of all of the tutorial videos that you all make!)

[scott]

Clean data pads, clear them of flux. You have corrupted data or bad strips. I have found that leftover flux will cause massive data corruption. You must clean off anywhere you have soldered a data line onto with PCB Solvent. Blade Adapter, Hilt adapter, Strip Data, and SOund BOard. Clean the mess out of them.

So I took apart the blade and cleaned and resoldered all connections after cutting and retwisting all wires. I reconnected the smd resistor and tested for 330 ohms...took the data line off of the board and cleaned the pad, cut and prepped the data line and resoldered very clean. Royal blue and green are great, but any red is an issue (starts with massive static through the speaker and then turns the blade into a rainbow). I tested with two blades with two strips each (and watched Toms video while rebuilding the blades.) same behavior with both blades, so either the board (prism 5.5 brand new), a systemic issue with the blade build (doubt it as I studied Tom and Robs tutorials), general wiring (possible), or a wonky config (curious how the neopixel settings interact with flkr and similar variables.)

Hoping that there are config suggestions, otherwise, I will revert to tri cree for now and pick the neopixel back up when my patience meter is reset...4 days of this has been frustrating. Any advice is greatly appreciated, and happy to provide any additional context if it can help diagnose my issues. I have legit read every post and tried every video available without luck. Thanks in advance for the help!

[Tom Tilmon]

Where did you get your strips, and which ones are you using? Did you twist your data wire around your positive and negative in blade? DO not twist the data wire, I have seen some wonky stuff when the data wire was twisted with the positive and negative feed wires (in blade and in the hilt) DON'T TWIST DATA WIRE. Have you checked your bridge on the Prism at the Data pad? Anything along the data path needs to be checked. I'm using the standard Prism neopixel setup on my test rig, and I have never had problems like this with most strips. If you have checked and re-cleaned all your data connections, you either have config issues or bad strips. You may also have a short somewhere in the strips. Shorts between data and negative can cause these kinds of rainbow issues like you're describing. DO the pins on the connector look ok? Does the center pin on the hilt side connector spring in and out? Did you insulate joints in the strips before you put them together?

I think you should take the connector out of the equation and see if you experience the same behavior in the strips. SOme strips can be wonky. I have bought good ones and bad ones. I have a pile of bad ones. I have not yet tried TCSS strips. I'm sorry this is frustrating for you. I found neopixels generally easier and more fun to wire than tri-crees. I think its useful to have a board not in a saber to play around with them to get the processes down. Thus my prism test rig. That rig saves me time and headaches.

[scott]

Thanks Tom, I really appreciate the guidance....strips from from TCSS, and I followed Madcow's TCSS video, so the strip wires are twisted to seat them into the blade adapter...I will try wiring that so that the data wire is not twisted at all. Bridge on the Prism is good. I also have a new Prism for my next build, so will wire that up tonight and attach the strips directly, bypassing the PCB connector (which does look sound, and all pin springs appear operational.) I did apply electrical tape at the joints per the videos. Fingers cross...and thanks again!

[jbkuma]

Is your saber still open or is everything locked away? There are a few things you can do to eliminate confounding factors and make sure each part is working correctly.
- solder the data line directly. by pass the connector and don't use a resistor. The resistor is for data integrity, but the strip should almost certainly work without it. (assuming everything else is fine)
- solder the power taps directly to the power source, or use a sufficient external source with a common ground.
- connect just the trim off from the strips you used to make the blade. A bad pixel on the blade could spoil it all.

The strips will draw power through the data line if there is a bad power connection.