An addressable net, or mesh, or curtain, whatever – was what I was searching for initially before I made the fairy lights on vinyl display.
IMO the problem with the bullet ones is that it’s overkill for power and brightness. For a reasonably dense matrix or layered matrix it’s going to add up to a lot of power draw. The fairy lights are so appealing for this reason because they are smaller and draw way less power, but there’s nothing pre made. The sparkfun ones are too delicate IMO if you wanted to make your own. The knock off AliExpress ones are fixed address. So there’s a hole in the market for this.
The AliExpress ones can actually be good in one sense precisely because they do have fixed addresses; they can be cut into segments and then wired in parallel which simplifies some installations, but if you want to use more than 200 or so then you’ll need an output expander to give you a separate channel and address space for the additional strings.
If I remember correctly, @wizard’s 8x8x8 volumetric cube used one 8-way expander per row (so 8 in total), but with the AliExpress strings you could do the same cube with 3 channels of a single output expander.
My little 1.6 x 1m 4cm spaced sheet is 800 pixels. A curtain or net, especially with multiple layers is going to run you into the capacity of 200 per channel on the output expander pretty quick!
Sure, but you can stack 8 boards for 64 channels x 200 for 12,800 pixels. Though at that point maybe multiple PB or expanderverse would be the way to go.
Lots of discussion on this in other posts. It really depends on the pattern and what level of FPS is acceptable.
One example:
There are multiple bottlenecks, processing is just one of them. On a v3, I think in most cases, if you go over 1000 pixels in untuned patterns, you’ll see some FPS drop purely from rendering time… Could be more or less, YMMV
How can you tell the Fairy lights on AliExpress have fixed addresses? The ones I have bought there are labeled as WS2812, for what that’s worth, and there are several versions available. Would it only show itself if you tried to attach a second string in series and it failed or as you suggested, cut one up and wire them in parallel, in which case they would work as expected?
… I have this perhaps irrational focus on something other than the bullet strings because I so much like the more linear arrangement found in Fairy Light strings. I’m just not so much a fan of the “jaggedy” aesthetic of bullet strings (reminds me too much of Christmas lights?) , though in a net the bullet could conceivably be arranged co-parallel with the conductor. That seems to be a basic hurdle that I might have to get over, dunno yet. As yet, I’m focusing on Fairy Light strings which does engender its own beautiful nest of issues.
I get the arrangement, thanks, have not built anything yet. But I’m also not so partial to zip ties for the above aesthetic reason. I know, picky, picky, picky. Again, dreaming of the fastenerless fastener. Let me know if you find one!
@Booli, the fixed-address fairy lights are usually advertised as “SK6812”, they’re connected by 3 copper wires, and all of mine have come with the same packaging:
Two definitive ways to tell that they have fixed addresses are:
The strings sometimes come from the shop with non-zero starting addresses.
I have one 100-LED strand that responds to addresses 20-119; I noticed because end-to-end chaser patterns like KITT or Cylon would disappear off one end and change directions before reaching the other end.
though I haven’t tried one yet. It’s not clear from the photos whether the vertical strands have sequential or parallel data paths, but even if they’re parallel that could be fixed by cutting the data wire at the top and jumpering across the bottom of alternating strands.
Thanks for all the great answers, folks! Really worthy, interesting and appreciated. In short, just a few notes -
@pixie you found the prior discussion of nets Nets Awesome!, and from the pics there I now see that bullets can be co-parallel with the wiring. Some tasty looking translucent shell (cubeoctahedron or higher poly shape perhaps?) might spice it up. @jeff 's reference to 10 x 30 nets from Ali Express Ali Express 300 LED Net shows them now at $240 including shipping. Worthy of consideration, but not “cheap” either. At the bottom of that listing page are several interesting arrays and net-like arrangements including Net-like Array and LED Array on Translucent Film.. This latter much like @gmcmicken filmed array Fairy Light Matrix on Clear Sheet, but his is better!
@gmcmicken, your matrix on a sheet is just really great. I love the whole approach… diffusion, effectively strain-relieving the strings, hot melt to clear sheet yielding wide angle viewing… that vinyl sheet though, did you see any puckering or melt behavior there? Could you reveal what type of sheet you used?
Finally, my fairly elementary observations about bottlenecks and such is that watching framerate behavior real-time is a great teacher and reality check for any given pattern/system combo. Multiple PBs and expansion boards can do a lot! Its not so difficult to wire this stuff up, as long as atention is paid to volts drop, and grounding grounding grounding That is IMO. Also there’s the alternative of running things from @zranger1’s ExpanderVerse, which opens up, well, yet another Verse.
ANd @Pixie thanks for running down the different Fairy Light strings. I have been running two types BrizLabs 200 LEd Fixed Address Strings and those 3-wire strings, where the data line hops back and forth. The former have much heavier wire then the latter. The 200 light strings seem to run fine without power injection, though there is some brightness drop.
Well first I looked on amazon for vinyl sheets and the reviews were all terrible so I had my wife look at the fabric store in town while she was there for something else. She said they had two thicknesses and bought me the thinner one. I almost guarantee this one (6mil?) from the fabric store is 10x better than the ones on amazon. It’s advertised to cover furniture, tables, etc. It did not melt, I didn’t have to be careful, and it didn’t pucker much at all. Though the glue is permanently bonded to the vinyl there’s absolutely no way to scrape it off now.
Thanks for the crystal clear clarification…I just went and check and I have versions of all of them, including 3 wire ones which I now suspect have fixed addresses.
Caveat emptor. It looks like each string might be run in parallel. In the video they are mostly doing the same thing on each string. There’s a little color variation but I suspect that might be incidental.
There’s also some minor flickering/glitches in the video, and complaints about that in the reviews. Reviews also mention setting the controller to 30 LEDs to get things to work.
On the other hand, if these do speak ws2812 you could hack them into separate strings by the connector and run them to a splitter to get things fully addressable.
For anyone interested, I had ordered these, and they are in parallel. It definitely would be easy to make them into a zig zag as suggested by connecting them at the bottom and cutting the top connection, but I haven’t tried yet.
Okay so I ordered these which are almost exactly like the ones linked above. The power/controller, the wire, the led size/shape/construction all look identical except for the shape of the junction boxes (with hooks) for each drop.
Anyway - the million dollar question - how are they wired and what protocol do they use. I’m happy to report they are all wires in series with a “return wire” running from the end of each drop back up to the top to continue onto the next drop. Despite 4 wires from the USB, one is not connected. So it is 3-wire, WS2812-like protocol with GRB order.
Can you simply chop off the controller and run it direct from a pixelblaze - yes you can.
