My thoughts
The technicolors will never be what I consider water resistant. They are just pushed into a hole and you can drive a truck into the led leads and board.
The bullets filled with silicone work great for water resistance. But the silicone is so hard gaps form between wire jacket and the silicone. My solution is to dip pixels into polyurethane which flows and fills even the tiniest gap.
Ideally the jackets of the wire would be covered with soft rtv. Then it would be encapsulated. That way no stress or temperature change would make the tiny wicking gap that sucks water up to the board
nutz4lights wrote:I think the Technicolor strings with the bulb are $30 for 50 and I was willing to pay that... I think your chemical numbers are definitely a worse case scenario. In a higher volume environment, those chemical costs would become more diluted (no pun intended).Louie
How does one goes around to have these mass produced? (Well, by mass being a few hundred). Record a video for Ray and have him quote on the process?
There's probably something else I'd do here that will maybe add another $2 to $5 per string - which is to create a clip that goes around a strap (a James-style strap with holes in it), and then create a little standoff on the other side where you clip the raw pixel to before potting. That will serve 4 purposes:
a) Hold the pixel at the same position in all C9's during potting.
b) Keep the wires at the back apart to ensure epoxy penetrates between them.
c) Makes the potting process simpler (just string a strap between 2 hooks in your garage, and go for it).
d) After potting, keep the C9 oriented outwards at a 90 degree angle to the strip.
It will also displace some of the space that need to be filled with Epoxy. However, cost of Epoxy and cost of ABS is about the same (assuming a 3d printed part). Can do it for cheaper with injection molding, but that requires a 5000 strand order to justify the cost of creating a mold (to get a $10k mold to amortize to < 4c per pixel).
Of course, it makes the lights unserviceable on the tree (you'll have to take down the strap, cut off the offending lights & break their clips, then re-pot new lights and have them dry). However, I like the dilutive effect of the white strapping, but that requires wires to be at the back of the strap, which in effect requires potting to be done while on strap.
To instead have serviceable lights, you'll need to use a Walter Monkhouse style strap, with the wiring in front. Probably would still need a little clip to orient them in the right direction. I don't think the normal style of using the wires to hold them out will work on the heavier pixel. On the plus side, a clip may make the Monkhouse straps a little bit simpler to construct. Probably only need insulation tape on one side.
deonb wrote:I've treated bullets with everything imaginable. In general, treatment just makes things worse. I found one highly impractical way to do it, but it's not worth it.
I've asked this somewhere in this thread i think and I don't remember seeing anybody post in, but are your failing bullet style pixels epoxy or injection style? I have around 5,000 bullet and 10,000 square style, all epoxy, not injection style, and have almost zero failures after two years, spread across probably one year of ordering (i.e. different batches). Now, I've had failures on arrival that get screened our during burn-in, but...
deonb wrote: I think the nanocoating that Louie suggested may work, but it's not a process that you can do at home. It requires pulling a vacuum on the object, injecting a fluid which then turns into gas, then introducing plasma to polymerize the liquid. Even so it would be much better doing that on the boards before potting.
Actually not true! Yes, plasma polymerization is popular for nanocoatings, but there is a straight liquid approach as well. I've seen plenty of good test data already as well (3M has products on the market and you can also search for the drywired 101x product line, there are several in that line). I've been assured that the liquid approach is just as robust as the plasma approach and the sole discriminator for its suggested use is capital equipment cost and ease of implementation (cost and labor). My understanding of the liquid process is that they take the active component and dilute it heavily. The product is dipped into the tank for application and it permeates. Thickness is less than a micron (human hair is ca. 40 microns), but typically more like 0.1 micron or 100nm.
Louie
nutz4lights wrote:I've asked this somewhere in this thread i think and I don't remember seeing anybody post in, but are your failing bullet style pixels epoxy or injection style? I have around 5,000 bullet and 10,000 square style, all epoxy, not injection style, and have almost zero failures after two years, spread across probably one year of ordering (i.e. different batches). Now, I've had failures on arrival that get screened our during burn-in, but...Louie
Injection style. I know this is bad, but the point that I was making is that if something is manufactured already 'partially' waterproof (like the injection pixels), it's very hard to make it truly waterproof afterwards over whatever attempt at potting it already has - which just serves to defy the additional liquid-based or spray-based waterproofing measures you're trying to take on top.
If you start off with epoxy over a raw pixel, then you're in a much better boat.
nutz4lights wrote:Actually not true! Yes, plasma polymerization is popular for nanocoatings, but there is a straight liquid approach as well. I've seen plenty of good test data already as well (3M has products on the market and you can also search for the drywired 101x product line, there are several in that line). I've been assured that the liquid approach is just as robust as the plasma approach and the sole discriminator for its suggested use is capital equipment cost and ease of implementation (cost and labor). My understanding of the liquid process is that they take the active component and dilute it heavily. The product is dipped into the tank for application and it permeates. Thickness is less than a micron (human hair is ca. 40 microns), but typically more like 0.1 micron or 100nm.
Louie
Wow, I just can't get any pricing on those. Do you have a link for a low volume retailer of either product?
Interesting to see the "outside the box" style of thinking coming out of this!
Yes the Sylgard is way too expensive for something like this but your clear UV protected epoxy would certainly be a better choice. You would have to get the actual material properties of whatever you plan to use to be certain that it would be compatible with things as well as a low air entrapment type of epoxy so that you could make reasonably sure that you would not have a lot of bubbles in it which would just make it self defeating to use.
Another consideration in this is how do you plan to "prep" the assemblies after you solder them together? Sonic wash to remove the flux or what other method would one want to consider? The flux will be just as big a problem as moisture unless dealt with and of course you can get other problems when trying to remove the flux and introduce other contaminants onto the surfaces that may interfere with a conformal coating or the epoxy final coating. If you buy the assemblies pre-made then how can you be sure that they are actually clean before you go through your steps to encase them?
Would you have to fill the C9 cover completely as that seems to be overkill and adds a good amount of weight and extra material to the strings? Should be some way to make a mold cover for the board and Led that would let you just cover the assembly with a few mm coating of epoxy and then you could always "weld" the C9 cover on afterwards and use a Nitrogen or CO2 dip to evacuate the air out of the cover before you attach it. That would save on excess epoxy use and weight. I see that the Technicolor pixels with the covers have a ring to secure the cover? Might be able to make a ring or use one that the cover would screw into after you used the ring as a mold cap to hold the few mm's of epoxy onto the board?
deonb wrote:Injection style. I know this is bad, but.....
...If you start off with epoxy over a raw pixel, then you're in a much better boat.Wow, I just can't get any pricing on those. Do you have a link for a low volume retailer of either product?
I don't even know why they sell the injection style anymore... I don't remember much of a price difference is there? The epoxy style have been super reliable for me and I can only assume they're not conformal coated boards... If the more expensive Technicolor bulbs are not conformal coated, that is probably respectable logic.
I will look into the coating materials when I'm back the office next week. I'm pretty sure we were promised 3m material samples... drywired coated samples for us, which we paid for and i haven't been quoted in their material. I can get quotes for their material.
Louie
nutz4lights wrote:I don't even know why they sell the injection style anymore... I don't remember much of a price difference is there?The epoxy style have been super reliable for me and I can only assume they're not conformal coated boards... If the more expensive Technicolor bulbs are not conformal coated, that is probably respectable logic.
They were from an old batch before the new epoxy style existed.
nutz4lights wrote:I will look into the coating materials when I'm back the office next week. I'm pretty sure we were promised 3m material samples... drywired coated samples for us, which we paid for and i haven't been quoted in their material. I can get quotes for their material.
Louie
Ok, thanks. I've ordered some raw pixel strands from that eBay seller to test with. I'm curious to see what will happen if all I do is dip it in conformal coating - without potting. Wonder if that will be enough.
In the mean time I get to go figure out how to build a salt fog chamber to be ready when they arrive... Fun.
If you are going to go to all the trouble to encapsulate and waterproof your own strings I would seriously consider using something like the APA106 pixels. Less components to worry about as the IC chip components are already integrated into the LED housing and less connections to worry about since you now don't need the 8 pin chip on the board either!
Something like this:
http://www.ebay.com/itm/100PCS-APA106-RGB-Pixels-12MM-Waterproof-Addressable-LED-String-Light-Logo-DC-5V-/271673807719
but of course without the epoxy applied.
I can get a price for some strings from a couple places if you would like? Would you like 50 node strings or 100 node ones? Also flat 3 wire cable in black or colors?
Let me know and I will see what they can do. I have already gotten some string done without waterproofing from one place so that I could inspect the boards and connections just to let you know.
Of course these are 5 Volt just so you are aware and not 12 Volt.
1. Would those make for shorter bullet pixels, or would the LED+ circuit board fit more completely into the c9 cover so that the epoxy would seal further down the wire?
2. Why are they seemingly expensive? Is that pricing not accurate because it is ebay and there is free shipping? What would a bare string of 50 cost for reference (without shipping preferably since i know what i pay for shipping from Ray).
I use mostly 5v product. The Technicolor lights were the only 12v lights in my entire display.
It looks like these are compatible with ws2811?
Louie
nutz4lights wrote:1. Would those make for shorter bullet pixels, or would the LED+ circuit board fit more completely into the c9 cover so that the epoxy would seal further down the wire? I don't have a C9 cover to measure but yes these are bullet style. The distance from the back of circuit board to the tip of the LED is generally between 26 mm to 29 mm. The sleeve is 30 mm sleeve and when it is installed on the LED the overall distance is average around 37 mm total length and that puts about 10 mm of sleeve past the rear edge of the circuit board. The circuit boards are 10 mm wide and 15 mm long but they could be brought down some without too much trouble.I have a manufacturer that I am working with on trying to get them to make the circuit board resized as well as a new board that would be like the Technicolor pixels are so that it can be made into the square style with mounting holes built in.2. Why are they seemingly expensive? Is that pricing not accurate because it is ebay and there is free shipping? What would a bare string of 50 cost for reference (without shipping preferably since i know what i pay for shipping from Ray). Well free shipping and those are "what you get". I put in some questions to the seller about buying them custom but am awaiting answers. I have a manufacturer that I am also waiting for answers on whether or not they will sell them without the waterproofing as just a plain 50 node string with black flat 3 wire and no connectors. Will post when I hear from them.
I use mostly 5v product. The Technicolor lights were the only 12v lights in my entire display. I am more of a 5 Volt fan myself as well. Just better power usage and I happen to have a bunch of 5 volt supplies.
It looks like these are compatible with ws2811? The APA106 will run the ws2811 protocol with no problem so far that I have found. I have some strings of them running on an ECG-P2 and they run fine.
Louie
Should have answers by tomorrow and will post then!



