Per usual, I’m well underway in getting ready for the holidays. October through mid-January is a big deal for me. All of my bulk-buy pre-orders of components have been received and dozens of Amazon orders have been placed. My weekends between now and early 2020 will be consumed by prop building, sequencing, cabling, and testing. It’s not easy, but it’s so much fun.


For Halloween this year, I’m upping my game. Last year was the first time in a while I’d really spent more than an hour or two decorating. The entryway was covered in faux spiderwebs, and I used a fog machine to up the spooky effect.

The entryway display from Halloween 2018.
The entryway display from Halloween 2018.

I also added some cheap 10-watt LED RGB floods for lighting. These weren’t the usual E1.31/DMX floods I have used for Christmas displays. Just cheap non-networked, remote-controlled versions. Finally, we gave away glow-stick necklaces and bracelets, but we ran out of these quickly.1Don’t worry kids, we hand out a lot of candy, too, and we rarely run out of that! Pictures were requested by a lot of parents, and the kids really enjoyed the setup.

The immediate feedback convinced me that 2019 needed to go even bigger. I will be adding black plastic backdrops to conceal the brick and door. I may cover the floor as well, but I am undecided there. I am planning to roll out one of the matrix boards that run a static sequence of GIFs/videos and some periodic text(Like Happy Halloween, BEwArE, Keep Out!!, etc.).

I am going to replace the LED floods I used last year with E1.31 floods from the Christmas display. I am not sure I will sequence them to music. I may just have them set to a specific color or just cycle in a loop.

Finally, I have already ordered about 500 glow-stick necklaces, bracelets, and a few other glowing goodies to hand out. Hopefully, this will be enough this year 😉


As I mentioned, everything I pre-ordered for this year’s display has been received. All pixels have run through burn-in tests with only two nodes being flagged as defective. I recently purchased 12, gently used MeanWell RSP-320-5 from a seller in the forums, and they came in great condition.

All pixel nodes added after 2018 will be 5-volt. Prior to that, I favored 12-volt to decrease the amount of power injection points. However, I’m coming to prefer 5-volt for its efficiency.2Efficiency means that I can power more than twice the number of 5-volt nodes vs. 12-volt nodes for a given wattage. However, the change introduces the need for more power injection power injection points due to a faster voltage drop over distance I ordered a few dozen power injection tees, and I have built three new enclosures for 5-volt, fused power only.

This change from 5-volt to 12-volt also nudged me to try differential receivers. Given the loss of voltage, pixels need to be closer to their power source. The differential receivers allow a separation of up to 250 feet from their source via a standard Cat5e cable. Each receiver has four string ports. For this year, I am adding a differential expansion board to output to three differential receivers.

Next, we’re adding a 10-foot mega tree. I’m adding another Falcon F16v3 dedicated for this. The smaller size allows me to get my feet wet with our first mega tree – in terms of sequencing and build – without overwhelming me. Since we have the area capable of supporting a very large mega tree, we’re going to eventually end up there. For now, however, we’re starting small.

Finally, I am converting one of the matrix signs from P10 panels to P5 panels. I am considering using this upgraded one for the tune-in/information board. I may disassemble the second P10 board to create a “super matrix” with 20 panels, but there may not be time for it. This P5 conversion may require a change from the BeagleBone Black-OctoScroller controller to a more capable controller setup, like ColorLight. That, however, requires some software added to the process, and I am not comfortable adding more than I already have for this year.

Show Website

Another project in the works is a dedicated website for the Christmas lights. I’ve had the domain for a while, but I haven’t had time to actually build a site. Finally, I got a head start on that this spring. Initially, it’ll be a single-page site using React displaying information about the show, such as the schedule, map, details about the technology used, and viewing suggestions.

The first dynamic integration I want is show status. Using the Falcon Player APIs, I’d like to query the show state(active/inactive), song/artist, run time, “next show starts in XX minutes”, etc. I’ve researched and talked to the developers, and there are mechanisms to retrieve this data already in the code. There are also intentions to add a full-blown REST API to fpp.

The challenge will be proxying HTTP requests from the website server3I’m not sure about the website infrastructure yet. If there’s no back end, I may just host on AWS S3 as a static site. If there is a Node/PHP back end, I’ll host it on a traditional Ubuntu VM somewhere. through my home network and on to the show network. I should be able to do this by using a specific port then port-forwarding that request to a bridged connection on the Raspberry Pi serving as the fpp “master”.


Web Developer

Add Your Thoughts

Your email address will not be published. Required fields are marked *