I got a meter of RGB dotstars recently and I'm working on stringing chains of them together. The solder pad on the 144 LED/M strips are tiny so I'm leaving a sacrifice pixel at each end of the strip. I.e. if I want 10 pixels, I cut 12 off the main strip. Then I solder the "IN" between 1 and 2, and I solder the "OUT" between 11 and 12.
This is leading to two issues.
- The 11th pixel is both the last one on strip 1 and the first one on strip 2.
- If I wire 2 'tens' together and configure my code for 20 pixels the last pixel is showing constant blue (i.e. the end sacrifice on the second strip). This is because it's not getting initialized and is showing some "random" value. If I configure my strip for 21 pixels it works fine.
What's the best way to disable my sacrifice pixels so they won't work ever?
I don’t think what you’re envisioning is possible: there’s no way to split a chain into two branches partway down the chain and have each branch be individually addressable. If you don’t want an LED to light, desolder it.
I'm not trying to run both branches. That's what I'm trying to stop.
Looking at the strip more carefully I can see the solder points. So if I use some desoldering braid I should be able to remove the LED off the end of the first strip? I'll try that. Thanks.
I'm not trying to run both branches. That's what I'm trying to stop.
Except that actually
is what you’re trying to do: have one branch do something different than the other, namely for one branch to remain off while the other lights up. Because of the bucket-brigade way they’re controlled (there’s no “smarts” inside of them with regards to addressing), there’s no way to send different signals to both branches, even if one of those two signals would be “Color: #000000”.
I’d use hot air to desolder. Or just cut not on the dotted line, but as close to the sacrificial LED as possible using a knife. And of course you could always solder wires directly onto the LED if the pads are too small.
Maybe we are talking semantics here. I understand that what I have wired up is in fact a place where as far as the circuit is concerned there are two position 11 (i.e. two branches). That wasn't intentional though it's just a byproduct of me leaving the sacrifice LED on the strand. I never want that 11'th LED on the first strand to light up which is why I'm going to disable/remove it. I probably should have known this would happen before I wired it up but I wasn't thinking about it.
I was wondering what's the best way to disable it. I didn't know if crushing it with pliers or drilling a hole through it or something of that nature would cause a short. I'll do what you suggested and desolder the LED or pry it up somehow.
I have a hot air gun I use for shrink tubing. Is that what you are suggesting? Wouldn't that cause the joints on the adjacent LED to loosen as well? They are pretty close. Do you not actually melt the solder all the way, just make it malleable enough that you can pry the LED off the strip?
In principle, if you remove the LED you must jumper the data and clock in-out pins. CKI==CKO SDI==SDO
I'm not sure what size dotstars you have but the issue is the 144/m density; you'll need to practice your micro soldering! You might be able to jump across the pads from the missing LED - see pinout. Use thin strands of wire rather than fight with solder blobs. Lock and insulate in place with dots of enamel paint or nail varnish.
In principle, if you remove the LED you must jumper the data and clock in-out pins. CKI==CKO SDI==SDO
I'm not sure what size dotstars you have but the issue is the 144/m density; you'll need to practice your micro soldering! You might be able to jump across the pads from the missing LED - see pinout. Use thin strands of wire rather than fight with solder blobs. Lock and insulate in place with dots of enamel paint or nail varnish.
It sounds like the LED removed is at the end of the string, and not driving anything so nothing needs to be jumpered.
But its not 100% clear to me without a diagram.
What's the best way to disable my sacrifice pixels so they won't work ever?
I understand your worry: the substrate is so delicate, and my fingers so sausage-y, I find it hard to reliably remove/disable a single LED without damaging the flexible strip (causing a "slight short" of the "bus bars" in the flexible strip). And that's with the cheaper 72 LEDs per meter ones, too. I mean, I can definitely get the LED off, but the problem is doing so without damaging the flexible strip and the two neighboring LEDs.
Have you considered using a paper guilliotine to cut the extra LED in half, then carefully cutting the soldered legs, leaving small "stumps" on the strip?
Unlike scissors, the guilliotine would cut (nearly) perpendicular to the strip, and wouldn't push the "bus bars" in the flexible strip around. With good sharp cutters, it is easier to cut the legs –– if visible and accessible –– than to desolder them.
You might also be able to just remove the phosphor coating, and then remove the actual LED chip(s).. but I'd worry about the leftover bond wires maybe causing some unnecessary load and/or shorts; it would leave the LED driver chip intact.
Someone like BigClive (bigclivedotcom at Youtube, bigclive.com) might know a reliable way; he plays a lot with these LED strips.
OP: if you desolder, you must melt the solder all the way. Solder doesn’t really have much of a soft phase (eutectic solder doesn’t have one at all), and you’ll just rip the pads off.
Nominal: there’s no phosphor layer in RGB LEDs.
OP: if you desolder, you must melt the solder all the way. Solder doesn’t really have much of a soft phase (eutectic solder doesn’t have one at all), and you’ll just rip the pads off.
Nominal: there’s no phosphor layer in RGB LEDs.
D'oh!
Right. I meant that transparent goop/stuff that usually sits on top of the individual LED chips, some kind of silicone or epoxy.
Ok thanks for the help guys. The paper cutter idea got me thinking. I ended up using a small pair of wire cutters to snip the legs. Being careful I was able to remove the LED without any obvious damage to the strip. Then I trimmed the remaining solder/leg as close to the strip as I could and I'll heatshrink the ends to prevent accidental shorts.
Did a quick test after snipping the two LED's and both strips still work fine.
I assume you have a 3 dimension array that represents each element of the string. If you want the 11th led to be off just set the eleventh array element to all zeros. And that applies to any one you want off.
My preferred way to remove them is with heated tweezers that can melt all 4 or 6 pads at once. The tiny MHP30 hotplate also works, and I've even re-used LEDs removed with it : attempts with other desoldering methods have damaged the LED.
The packaging isn't very good on these parts. The white body plastic has quite a low melting temperature so movement of the pins while hot tends to break the internal bond wires, and the seal between lens and body seems to easily break and let moist air in. Neither of these matter if you're just removing one led and not trying to replace it.
I've found the dense (96led/m and up) the hardest to solder to : the pads on the lower density ones are larger. So I tend to remove one led where I want to end the strip and make the cut in the centre of the led site. That leaves me a full round pad instead of the half-pad I get when I cut on the line.
It's sometimes easier to solder to the back of the strip, though you have to remove the adhesive tape if present. Also note that only the data lines need to be connected right at the end : you can solder the data line at the end and the power lines one led down. This makes the connection less congested.
If you are removing any surface mount part with hot air, it is smart to mask adjacent parts with Kapton tape whilst heating.
Kapton is a clear 'polyimide' tape that tolerates high tempertures to 250c - and beyond in short duration exposure when melting lead free solder. Remember to always flux the part being removed
A reel of Kapton tape is a need to have on any EEs workbench.