My team wanted to put LEDs on our robot this year (first time doing this). We got the AndyMark LED kit and added them at our event last weekend. When we finally turned the robot on, the 20A circuit breaker kept tripping. We then switched to a larger gauge wire and a 40A breaker, and then the small bit of 18 AWG wire after the first split started to smoke. We then removed the wires from the PDP to the microcontroller and power converter and went on with the tournament. So, where did we go wrong?
Attached is a diagram of the wiring. Terminal blocks were used to split the wire. The wire that started smoking is labeled in red. This happened before any code was put on the microcontroller. Also, the reason that the pins are connected to DIO ports is because that seemed easiest for an inexperienced programmer like myself.
The next time we will have access to the robot is at our week 6 tournament.
Well, saying you said “AndyMark LED Kit”, I’m going to assume you mean AM-3010. If not, please ignore the following piece of advice. I (as someone who has only linked around with electronics and LED strips) don’t see anything wrong with your power wiring. I do see one thing wrong with your signal wiring. These are WS8212B LED strips, which means that they are addressable per LED “pixel” and therefore you shouldn’t need multiple digital outputs to control all of these. I do however see a problem with how your LEDs are wired. The LED strips have both a signal in and a signal out, and my rudimentary understanding of the communication protocol used by the WS8212B chip is that they do signal altering stuff (technical term) inside each chip, so you then have to use the signal out to the next set of strips. You can’t have multiple signals coming from the same wire. This doesn’t however seem to solve the issue of it smoking. Have you tried taking off the power converter and testing just it with a multimeter and power supply?
We have been using those LED’s, from another company… Which is probably who supplies Andy Mark…
Last year we had the same thing happen after a few matches. While we are in the pit we run the LED system as attention grabbers…
Good thing we did as it started smoking!! Bad Robot!
Turns out that the adhesive that is on the back of the strip was off center. It allowed the contact / cut & solder points that is normally under the glue to contact the frame. This short was drawing a lot of current and started to melt the wire insulation.
Easiest thing to do was purchase some double sided foam tape that was wider than the LED strip and mount the LED’s to it, then mount to the robot.
This is something we did on this years robot too. It also makes installing easier…
Due note that most adhesives on these LED strips is just sprayed on and very thin… The LED strip does not do sharp turns well either… SO if you plan of making sharp turns you are better off cutting the strip at one of the cut points and soldering on some wires. once you have these new wires connected use some hot glue to reseal the strip. This will help a lot with strain relief. The solder pads on the LED are not very robust.
Also it is a good habit to zip tie the LED strip down every so often as the adheavsives on the LED and the double sided foam tape do not work very well on metal that isn’t super clean…
Note: do not upgrade a fuse/breaker that is tripping to a higher rating if you are having problems… If it did not trip the 20 amp breaker during this smoking incident, allowing the short to be able to draw twice as much before tripping is asking for the smoke to become fire… ::safety::
Yes, it is am-3010. WS2812B are Neopixel LEDs. According to this, connecting LEDs to separate signals is ok, and according to this, connecting multiple strips to one signal is ok, they will just run in unison, which was our intention.
The LEDs did not seem to have any adhesive, so we attached them with VHB tape from the KOP.
I constructed a similar Neopixel RGB LED light assembly controlled by an Arduino compatible controller. It is very similar to the AM-3010 kit they sell at AndyMark. I bought my parts from Adafruit, and used their Neopixel Uberguide to connect them.
We wanted 4 short strings of 7 LEDs on each of the four corners of the robot. This was accomplished using a single 30 LED RGB Neopixel strand https://www.adafruit.com/products/1376 , cut into 4 sections of 7 LEDs with extension wire in between. I used a Teensy 3.2 (Arduino compatible) to control it with a single signal wire. Further, I did not need an additional power supply, since the current drain can be controlled by varying the brightness & color. I am strobing the LEDs and colors so that the current drain is less than 300ma total. This is easily powered by any garden-variety USB port, and allows quick software testing using just a laptop alone.
With that as background, I noticed some things during the assembly that could cause your shorting issues. When you cut the strand, you are also cutting through the flex circuit containing 5V / Ground / and the clear rectangular silicone insulating sleeve (with solder pads exposed at the end on the top and bottom of the flex circuit). Therefore you must insulate the connections! I slid a small piece of heatshrink over the middle signal wire, covered the solder joints with RTV silicone (not shown in the first photo), and then covered the ends with large heatshrink.
I attached an in-process photo, and a picture of the near final version 1 prototype. The photo does not show much of the Teensy 3.2, since it is enclosed in very large heatshrink. The version 1 prototype shown includes external 5V/GND power connection capability that is no longer needed.
It uses three DIO ports (not shown) of the roboRIO to indicate what LED pattern and color to display. Scrolling, flashing, steady, etc. For example, pattern “111” is when the roboRIO is in reset or booting, and displays a single dim red LED on the middle of each section. Pattern “000” is for climbing, and strobes white from the bottom to the top of each section simultaneously.
I’m moving the controller and DIO connections onto an MXP protoboard https://www.andymark.com/product-p/am-2994.htm to simplify the installation and wiring. I’m not done, so no photo yet.
I just realized something, the connectors we used have the wires completely rearranged. The signal wire connects to power, the power wire connects to ground and the ground wire connects to signal. Boy am I glad these are WS2812B strips. We’ll have to fix that.
Could that have an effect beyond them just not turning on?
The individual LEDs internal to the WS2812B devices should not be damaged by the incorrect wiring (since they are diodes and conduct current 1-way), but they also include serial shift register logic for color choice / intensity / buffer to the next Neopixel in the string. I don’t know how the internal transistors of the shift registers are configured, but I would not be surprised if they were damaged. If you had a resistor in series with the signal wire as recommended, it might increase your chances.
