Lawn of Light current status. Any last min advice?

Uncategorized
1

Lawn of Light: 735 ground lights (soon ~1,500) on one Pixelblaze. Advice welcome before I double it

This is a preliminary layout to make sure everything works before the real event. We’re running it next Friday as part of the celebration for the inauguration of Boston College’s next president. The statue is St. Ignatius, founder of the Jesuits (BC is a Jesuit university), and the lights ripple out from him in arcs across the lawn.

What’s running in the photos

  • 735 lights, all on a single Pixelblaze V3 output. The final build will be about 1,500.
  • Lights: Wally’s Lights 12V Ground Path Lights. RGB, WS2811, Ray Wu connectors, 15 lights per strand at roughly 1 ft spacing. That’s 49 strands so far, 100 for the full build.
  • Pixelblaze settings: WS2812/NeoPixel type, 800 kbps, color order RBG, brightness limit 50%. 50% is plenty bright at night.
  • Power: one Mean Well LRS-350-12 (12V, 29A), which is far more than this half needs. It feeds a 12-circuit fuse block with a 5A fuse per feed, then 12/2 landscape wire out to Wally’s 3-2-3 Ray Wu power injection tees. A GFCI cord connects it all to an outdoor extension cord.
  • Layout: one long serpentine laid as arcs about 3.5 ft apart, growing from about 65 ft to 110 ft long. Every arc turns on the right side, near the power box, so I inject power at the start and at each right-side turn. Right now that is 5 injection points for 735 lights.

A test that might be useful to others

To see how little injection I could get away with, I fed 180 lights (12 strands) from one end only, all white. At the 50% limit I measured 12.0V at the supply and 9.6V at the 120-light point, with no visible color difference between the first and last light. At 100% it dropped to 8.6V with a slight fade at the far end.

Questions for the pros

  1. Power. For the second half I plan a second LRS-350-12 feeding the remaining right-side turns, with +12V broken where the two supplies’ sections meet (data and ground carried through). Anything you’d change before I build that?
  2. One output for 1,500 pixels. Will frame rate be acceptable, or should I split it across an output expander? I have one on hand but haven’t wired it.
  3. Data lead length. I’d like the Pixelblaze behind the wall, about 28 ft from the first pixel. Is that asking for trouble?
  4. Laying out the arcs. Any tricks for getting clean, evenly spaced arcs on a lawn? I thought about swinging a string from the wall, but the wall isn’t a true semicircle, so that won’t measure reliably.
  5. Mapping. My first-pass map models each arc as an evenly curved row and assumes the turns fall where I counted them. Is there a better way to map what I end up with, other than counting the lights in each arc?

Thanks for any suggestions!

2 Likes
2

Nice!

  1. Power: seems reasonable to me
  2. The data for each WS281x takes ~30uS. 30us * 1500 = 45000us or 45ms. 1000 / 45 = ~22 FPS, which is pretty noticeable. Splitting those up evenly across two outputs would double your FPS to 44, four outputs to 88 FPS, etc. That’s all assuming the Pixelblaze can actually render your patterns at that speed, which is unlikely. On my Fibonacci1024, with the 1024 LEDs split across five output expander channels I think I usually see ~40 FPS max.
  3. 28 ft to the first pixel seems pretty long and highly susceptible to noise. You can try it, and might get away with it, but you might see flickering. If so, you could try splicing in a few pixels along that run. You can also use CAT6 ethernet cable, but might need a 100 ohm impedance-matching resistor on the data line, like the OctoWS2811: OctoWS2811 LED Library, Driving Hundreds to Thousands of WS2811 LEDs with Teensy 3.0
  4. If you want true circular arcs / chords, can you find the center easily, then place a stake and string there?
  5. Mapping: any chance of using a drone from directly above, take a picture (or grab a still from video), then use an image-based mapper, where you click on each pixel in sequence? It’d work best in the dark, with all pixels lit. There are lots of them available, including mine here: LED Mapper
1 Like
3

For laying out the arcs, since the desired radius is large, and there are some obstacles to placing a center pin, getting everything perfect with the string method might be difficult. If you have any budget for this at all, here’s what I’d try:

  • mark/stake a linear grid of edge positions and a baseline center for each arc. Looks like you can measure the edges using the sidewalk, or use a laser level to get everything as regular as possible.
  • figure out how much you’ll need to displace the center to get the radius you want. (should be the same for each arc if you want offset arcs of equal radius.)
  • Stake the displaced centers.
  • Now the good part: go to home depot and buy enough 1/2" or 3/4" PVC pipe and connectors to span the field in the arc radius you want.
  • Run the PVC from the stakes at one edge to the other, using the displaced center stake to make an arc. PVC has enough flex to do this, and is stiff enough to keep a smooth, regular arc.
  • Set up the LEDs along the shape described by the PVC, or mark the arc with chalk and then run the LEDs.
  • repeat until you’ve got all the arcs!
1 Like
4

Thanks, all three of you. This is exactly the kind of help I was hoping for.

@JasonCoon: your frame rate math is super helpful. For this event I’m staying on one output at ~22 FPS. The patterns are mostly slow ripples and fades, and with the show on Friday I’d rather not add an expander and more data runs. I’ll also move the Pixelblaze right next to the first pixel, so the 28 ft lead isn’t needed. The drone photo plus LED Mapper is a great idea. I’m working on getting a drone overhead at Wednesday night’s rehearsal and will share the photo. I’ve not used any mappers before, but it’s great you’ve got a tool for this. Excited to try it! I’d also welcome any thoughts you may have on cool patterns for something like this. Your work is always so well done.

@zranger1: you’re right about the center. By my estimate it sits about 66 ft behind the first arc, past the statue, so a stake is out. The plan is to lay the first arc by eye along the wall, then run five guide strings out from it, each marked every 3.5 ft, so every later arc is pinned to marks and doesn’t drift. I like the PVC idea for shaping that first arc.

I’ll report back with video after Friday. Thanks again!

1 Like