Building a Synchronised Christmas Lights to Music Display - Part 2
Woah! Never expected that Christmas Lights would take up as much time as it has so far. I wasn't under any illusion that it wouldn't take time - but as I write this on 2nd October I'm reminded that I'm just over a month away from hanging stuff up on the house and running my first synchronised light show.
First - the props. In my last post, I left you with a picture of my 3D printed singing Christmas Tree and a view to make some new props. I started working on a similar layout to the singing Santa from Boscoyo albeit at a suitable size for the space it'll be mounted in. The process was similar to the Christmas tree - eyeball the layout from xLights, reduce the size to the space I wanted to occupy (around 60-70cm square) and draw in new 12mm holes for the bulbs to sit in.
The Singing Santa
I must admit, he looked a little creepy in this one. Some things I learned from the previous print is to try and keep bulbs at notable points. While you won't see the plastic once printed and lights are up, not having the lights right in the corners and edges means the light up will look slightly off. Also - Santa needed a red hat and white outlines for this setup so while I didn't end up printing out the plastic in white or red, these would roughly serve which lights will light up those colours.
This one was a little more challenging, especially with the mouse piece just breaching 200mm (although using some angling, eventually got this to print. The downside of this was the lack of supports around the beard. There's a lot of weight in the top half reasonably well secured. I also noted that for at least one set of lights, I'll need to solder on a joining wire between parts to ensure we've got enough distance to wire him up.
These props will be mounted to a PVC pipe frame so I'm not ultra worried about it - but if I were to design this one again, there'd be two supporting brackets either side of the beard, up to the side of the mouth, and a couple of smaller chunks between the mouth and nose.
You'll note that the success of the 'train track' join style, I've used that again here as it's been a good way to join plastic together with a solid fit without removing much plastic.
in the lead up to Santa finishing printing, my delivery from Amazon arrived - my 1,000 lights that I'd need to cut, splice and add pigtails to that I'd ordered from Light it Up LEDs. Now I must admit I wish I discovered Light it Up LEDs much earlier than the Amazon order - the bulbs were a little cheaper, and he had LEDs with the connectors already soldered on. This would have been a tremendous time save, but none-the-less I had these bulbs and I had to cut off odd numbers of bulbs. For this guy, he is 203 bulbs so ... 4 x 50 + 3 from another strand. Another learning here to try and get multiples of 50. For reference, the Christmas Trees are 156 bulbs. If I could reduce (or increase) the density, it would have made wiring easier.
I don't have a lot of pictures of me soldering wires away, but it's fair to say I've soldered a good couple of hundred pieces of wire together. I've got pretty good at it too. But there's one thing I kept missing... the bloody heat shrink. I really need to print this out and frame it, frankly.
Anyway... with pigtails attached, wires spliced with extension wires and some waterproofing done (silicon around the exposed connectors), it was time to turn Santa on. And wow - it looks great.
Now the eagle eyed of you may note there's a hole with a bulb missing! Yep - I missed one bulb. And yet... I didn't have any leftover bulbs. I'm not even sure how both errors lined up. In any case, always remember to count your bulbs. Sorry Santa, but you'll probably be replaced next year.
The Candy Cane
Ok class, time for a bit of Math. If I have 1,150 bulbs, Santa consumes 203 and each singing tree takes 156 bulbs, how many do I have left? Well, of course the answer is 635 left that I need to distribute between 5 mini trees + stars and 4 candy canes. Each mini tree would be around 90 bulbs (70 for spokes, 20 for the star) so that leaves around 185 bulbs left to use. That leaves around 40 bulbs per Candy Cane left, with some spares for repairs I might need to do along the way.
So I designed a 39 bulb Candy Cane. It'd have been nice I think to add another strip in the middle of them but ... these props will probably be replaced next year anyway. So out comes good ol' Sketchup and I start working on a design.
A pretty simple design that in future I'd probably have put some holes in to allow for cable ties to go in the middle rather than around the whole prop. But for this year it should be fine. To make them look a little decorative, I'd alternate white with red ABS. I certainly have plenty of filament so might as well make use of these colours.
After a day of printing the parts out, I have a good idea on how this is going to look. After wiring a few bulbs in we have some success.
Honestly, having put the pixel tester in and saw it light up - I was a bit underwhelmed. Really needed that middle strip I think but we'll go with the two lines and see what happens. I'm sure I can make some decent effects on it and you'll be seeing it from the footpath anyway. So I digress...
With this one out of the way, it was time to set up a bulk print order...
Seeing these fully assembled though - I'm pretty happy with how they look. And most importantly, my son loves them. There's something in the reflectiveness of the white that is intriguing to him.
The Mini Tree (with Star)
So the star of the show (ha ha) would be some of the mini trees .They'll carry most of the beats and the more interesting displays but they'd also be the most difficult to print. This is perhaps where a local supplier of Coro material to do it would have been a little more effective but I figured I could print a base with the star to mount to, and have holes for 'spokes' to sit in. I'd use tent pegs to hold the legs in place which gives me a bit of an opportunity to perfect the placement of these bulbs at least for the first year. It's fair to say these'll probably be replaced as well in future.
The base is fairly solid. I'd discovered that a 25% Gyroid Infill was a good balance between a rigid structure and weighty enough. But I probably should have printed some sort of base for a PVC Pipe to hold it up. But that's OK... nothing a bit of Loctite won't fix... after all if it's good enough for my brake calliper bolts, it's good enough for my PVC pipe. The plastic cement stuff anyway.
These definitely produce an interesting effect and the adjustable spokes will be good for putting a little bit of personality into the display. Something a bit missing when you carbon copy everything several times.
These will definitely look great out on the lawn come Christmas time! Obviously there's some waterproofing to do and more pigtails to solder *sigh* ... but we're good to go on these ones. So queue the bulk print job for another weeks worth of stuff.
Now I've got to say, all this 3D printing is burning through Filament fast but how much you say? Each of the larger props (Christmas Trees and Santa) were less than a 1KG roll each. With each ABS roll costing around $25 AUD, it's not the most economical product but it is allowing me to print things to the dimensions I want. For a lot of the Candy Cane stuff, Christmas Trees etc... there's probably around a role for the lot - but they are thinner prints and less plastic. But it does take a lot of time to do. The Christmas Tree bases alone - 4 of these in addition to the test tree, was say around 1/4 a role and took over a day. Each spoke needs to be printed 5 times, for 4 trees. They take over a day each and probably consumed 1/2 a role. The stars were similar - a day to do those and took around 1/8 a role. So expect around a week of printing, at least on my printer that runs around 30mm per second. I'm sure that more modern printers are able to run much faster these days and with the right infill models in place you can probably reduce a lot of rework.
Where to from here?
Well, more progress has been made. I've built a sample matrix and even found a use for some cheaper Chinese bulbs that I thought I'd trial. These have been done in the last two weeks, but this post is already 18 images deep so I'll leave it for a couple of weeks and showcase those. I've also got myself heavy into the sequencing game so there's probably at least 3 more posts before the actual show.
I must admit - this has been one of the most enjoyable creativity projects I've worked on in years. Being able to build something with your own hands and see it go is great. Watching your son full of enjoyment when the lights come on takes that to a whole new level. Anyway I hope you've enjoyed seeing the pictures and if you're inspired to start working on things, once again I can recommend taking a bit of a look on AusChristmasLighting for some further inspiration. But be warned - it's not something that you're going to be doing if you've only got a budget for $500 or less.
But ... I am keen on working out what kind of show you could build for $500(ish) if that's something people are interested in next year. There are some good shortcuts and devices you can buy for some light sequencing paired up with some traditional AC-style lights that I recon could be a winner. Also - it looks like Bunnings are getting in on the action with a peculiar Christmas Tree that has individually addressable lights. You can be sure that people are hacking these already to work with the WS2811 protocol. Costco even has these 'Twinkly' branded trees at the time of writing for $199. In any case, it's not a cheap hobby and if you can do some of the tinkering yourself and you already have Ardunio's, Raspberry Pi's and are set up with enough to be dangerous, you can do most of it yourself and save on those dollars.
Anyway - until next time. Ciao!
Building a Synchronised Christmas Lights to Music Display - Part 1
With only a couple more lockdowns to Christmas, it's time to get those gifts ready, plans organised and all that other stuff people do in the leadup. I used to despise people who'd start this far out from Christmas, but this year is a little different. My son last year (who is coincidently 1 year old today!) was mesmorised by some Christmas lights we put on the outside of our house last year. Nothing fancy - just some icicles and fairy lights from good ol' Bunnings. They did the blinky blink thing and he really enjoyed it.
No surprises that Covid-19 lockdowns have done a number on people looking for inspiration and last year it really shone through at least where I live with the sheer number of displays up compared to previous years. Some are quite impressive - entire houses covered in more fairy lights that would give Griswald a run for it's money. I'm also reminded of a display I saw a few years ago in the western part of Victoria that seemed to have some crazy Dubstep tunes pumping out to Christmas Lights. While it's not my choice of music, I can appreciate the effort that went into synchronising Arches to the bouncing beats.
I've always wanted to build a full-on Christmas display. I figure that with some programming talent, recent work on 3D printing and a few Raspberry Pi's and Arduinos, I should be able to build something similar albeit without all the Dubstep. Perhaps something a little more traditional and banter-ish.
Amazon sneakily recommended me some WS2811 lights and to make matters worse for me, they sealed the deal with a $5 coupon to go with it 😊. I figured at the very least I can start experimenting with a Raspberry Pi and turning these "NeoPixels" on and off individually.
In the meantime, I put my Google-Fu to good use and found some displays which looked interesting nearby but nothing out where I live in the eastern part of Victoria. I found the one in the western part of the state that I saw as well as another that had a large Mega Tree. There's just something not quite right to me though - don't get me wrong, it's a perfectly good display but not something that I could really see fitting my front yard.
That lead me to another forum (yes - active forums still exist apparently!) AusChristmasLighting. I must have spent 3, 4 maybe even 5 hours working through the Display Videos section - there were some pretty awesome synchronised displays there. With some YouTube-Fu, a guy by the name of Tom BetGeorge kept coming up and I eventually found the display and song that would inspire me now to build some props and try something simple. The music sequencing is near perfect in time to music, the banter the Trees have is fantastic - this is the sort of thing I'd been wanting to do for a while. It's amusing, kid friendly and has the right amount of joy for me that would make it worth doing.
The forum I mentioned earlier - at first I didn't know it was an active forum until I started looking through the comments section of those videos. Definitely a treasure-trove full of advice (conflicting sometimes, but at least you get to hear lots of different opinions as this can get expensive very quickly if you make a mistake). The one consistent theme was a lot of praise for some free software written by an Aussie called xLights. On face value, it's sequencing is inspired by interfaces similar to FL Studio which I use a lot for sequencing music, and thought I'd jump right in and mock up a layout - see what this thing could do. It was fairly intuitive I've got to say. It doesn't look well organised when you first load it up on a 4K screen, but I was able to, having not read any manuals or watched any YouTube videos, roughly work out what to do. So much so that I had created my first rough layout.
Those Christmas Trees and Santa were part of what I would find out later to be a supplier of 'Coroplast' Christmas Props. I did have to wonder what kind of rock I'd been under to not realise this stuff was readily available but unfortunately shipping would prove to be a bit much and the sizing would not be compatible under the roof line - given the decorative elements already there.
Those lights I ordered eventually arrived so I started sketching up a template that should fit the bulbs. The best way I find to work with 3D printing, especially when it comes to measuring up things is to take a measurement, but then put a hole slightly bigger in as well. Print a few test templates. I wrote a few lines of code using a Python library for WS2811 LEDs.
Great! The next step would be to climb up the side of the house and on the roof to get some measurements and think about custom designing some of these props. They would need to have a particular wiring order (configurable in xLights) and an appropriate size. Having drawn half a tree in Sketchup, I eyeballed the placement of the correct dimension for those bullet pixels. Once done, it was copied and flipped horizontally for symmetry and I began drawing Jigsaw lines in (I had experimented with a couple of joining techniques, this one works well for ABS at least. As you can tell, this isn't going to leave much room for error, but nothing a Mallet can't solve.
The print time would take around 5 days to complete. This method isn't perfect by any means. As it turns out, either the print diameter or the bulbs themselves are not identical meaning some fit well, some are too loose and some are too tight. Also, the eye balling of bullet placement should have taken into consideration strings of 50 to avoid lots of cutting, splicing and soldering. If I were to do this again, I'd probably more densely pack pixels and perhaps shape up more of a mouth to get it to 200 pixels and look to more evenly space the pixels including changing the hole design so the bulbs "clip" in rather than hard pressed. So while not perfect, the dimensions will certainly fit within the desired space.
With all the bulbs pushed in using a tool from this guy (which I should have found before I started, it would have saved my hands a significant number of cuts not too dissimilar to 90s era PCs), it was time to fire up xLights, and an OS for the Raspberry Pi called Falcon Player. Having connected the 12V string of 156 bulbs to the Raspberry Pi and injecting some 12V 5A into the string, and a face singing a song, well... here is the result (albeit with a bulb or two that requires better remapping).
There has been a lot of progress in the last few weeks on this including the Mini Tree, Santa and Candy Cane. I'm still yet to determine how i'll complete a Matrix, but I've got some ideas using some PVC Conduit and some 3D Printed templates to drill in. I'm also waiting on some WS2812B LED Strips to come in from AliExpress to do the house outline with. I've also sequenced a couple of songs already to get the hang of xLights (hint: there are a tonne of YouTube videos explaining concepts to help you accelerate your way to creating effects). Those will certainly be a bit of fun to work with when the time comes, but none the less this year's Christmas display is shaping up to be something awesome.