fiduciary

Author Topic: Custom Neopixel Strip Concept and Testing - UPDATED  (Read 3378 times)

0 Members and 1 Guest are viewing this topic.

Offline SaintLuke

  • No Force
  • *
  • Posts: 3
  • Um...Hello?
Custom Neopixel Strip Concept and Testing - UPDATED
« on: January 14, 2020, 08:05:40 PM »
Hello Folks! This post is sort of an introduction, and also includes what's in the title...

My name is Mark and I have been building sabers since 2009. I took a break from the hobby in 2015 for some other projects and am finally back. I cannot tell you how excited I was to see both Neopixels and Smoothswing...wow!

Anyways, I am working on a new saber and was not super happy with the available Neopixel strips. If you go with the "Skinny Mini" 3535 strips you sacrifice some brightness for evenness. On the other hand the 5050 strips are brighter, but harder to diffuse. I saw KR-Saber's new hard-PCB strip and was inspired to try something similar.

I have made some progress on the design and wanted to share it here to see if anyone has suggestions or tips. I'll keep this thread updated with progress.

Here are some renderings of a proof-of-concept model (parallel data lines not shown):

[nofollow]          [nofollow]          [nofollow]
Specs:
  • Width: 6.9 mm
  • Thickness 3.3 mm
  • Density ~144 LED/m
  • Total Cost: $30

So you get smaller width/height than skinny strips and the same pixel density, all with 5050 Neopixels instead of 3535.

In order to drive the LED's with a single set of voltage rails, you have to flip the LEDS on one side of the strip 180-degrees...basically an upside-down mirror image. This, however, causes the data line of one side to run from top to bottom, which is a problem. I had the idea to pull the intermediate data signal from between each LED from the serial side to drive each LED on the reversed-side. Below is a picture of the test-setup I made. Also, here is a link to a short video of it working: Driving Parallel Neopixels - for Lightsaber - YouTube [nofollow]

[nofollow]

For ease of assembly I decided to use axial ceramic capacitors. No, there is not a cap per Neopixel, however having 0.1uF between each set of 2 pixels should be plenty to prevent transients. Unfortunately these capacitors are really the limiting factor of the pixel density (they are also expensive) and I am toying with the idea of using an 1806 SMD package...but durability might be a concern.

Here is a materials list and cost breakdown:
  • Pixels: SK6812 -> $26.99 for 500 pcs-> ~$13.5/blade
  • Rails: 20 AWG 99.9% copper wire -> $6.99 for 22 ft -> ~$2.20/blade
  • Capacitors: 0.1uF Axial Ceramic -> $6.50 for 100 pcs -> ~$9.5/blade
  • Data Wire: 30 AWG Mil Spec Solid Core PTFE -> $7.98 for 20 ft -> ~$4/blade
  • Heat Shrink: 8 mm OD Clear -> $4.12 for 20 ft -> ~$0.65/blade
  • Time: I estimate 15-20 hrs -> whatever your time is worth to you -> ~$0/blade
TOTAL: ~$29.85....let's just say $30

So, you can build this whole blade for less than it costs to buy one 144/m strip. Of course, this may be a garbage idea/design that will break in the first few swings, which would make the low cost not so important. (Not to mention the tremendous amount of time needed to make one... :rolleyes:)

Please let me know your thoughts/suggestions/criticisms. I am not married to this design and have open ears for any pointers!


UPDATE 1/15/2020

Okay, so I found some 1206 SMD capacitors (much cheaper) and changed the wire size to 18 awg (0.0403"). The new capacitors will fit between the two neopixels (like a sandwich), which allows the neopixels to theoretically have no gap between them. Of course, I left 0.5mm to give the strip some play and make room for data wires. Here are the specs with the new arrangement:
  • Width: 5.77 mm
  • Thickness 4.11 mm
  • Density ~166 LED/m -> 322 for 35" blade
  • Total Cost: $25

A few screenshots/renderings:

[nofollow]

[nofollow] [nofollow]

Thanks,
Mark
« Last Edit: January 15, 2020, 12:30:30 PM by SaintLuke »

Offline ShtokyD

  • Mining Colony Members
  • Force User
  • *
  • Posts: 220
  • whatever
    • My facebook page
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #1 on: January 15, 2020, 04:26:05 AM »
I would say just buy a ready to use strip from KR-sabers when it's released :) Because as you said in your design there are a few obvious issues:

1) very fragile, pixels joints points will brake immidiately when it's slightly bent
2) a lot of time to assemble the strip (and not easy to do), so it won't cost $30 really...

New KR-sabers strips are single-piece rigid pcbs (but still flexible in a needed range) designed from scratch specifically for use in sabers. I helped Khal to test and improve the strip design for absolutely max efficiency, so it will be the best and brightest neopixel strip on the market for sabers. Also they have pixel density even more than 144 leds/meter and are 8 mm wide with 5050 pixels. And it won't cost more than 2 regular flexible strips.
« Last Edit: January 15, 2020, 04:33:24 AM by ShtokyD »

Offline SirRawThunderMan

  • Master Force User
  • *****
  • Posts: 708
  • Very, Very English Poster
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #2 on: January 15, 2020, 04:35:07 AM »
First thought: I'm no neopixel expert, I've only made the one pixel blade saber and I don't really intend to make more (component blades aren't my thing). Off the top of my head though, my concern would be waste heat dissipation. I was under the impression that the metallic strip/PCB that the individual addressable LEDs are mounted on are there in part to serve as a heat sink to keep the 144p/m (x2) from cooking too fast from the 15+ amps being pumped though them in operation. taking that away entirely seems to me like a recipe for releasing the Magic Blue Smoke.

Second thought: okay, so, LED string blades have mostly died because it was a right pain to bend the legs of 100 5mm LEDs and solder them together, and Neopixels came along and surpassed them in most aspects. It seems to me that what you're proposing here is taking the PITA complexity of making an LED string blade, and complicating it further by adding soldering to SMD sized pads. I dunno, it just sounds overly complicated to me.

Third thought: maybe replace the single core wire on the data line with stranded core wire. if the single core wire breaks, your whole blade will die up to the break.

Despite my negative outlook, I do look forward to seeing what you can do. This hobby thrives on people trying out their craziest ideas. Who knows, maybe you've got the next big thing here  :smiley:
Better late than never.

Offline ShtokyD

  • Mining Colony Members
  • Force User
  • *
  • Posts: 220
  • whatever
    • My facebook page
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #3 on: January 15, 2020, 06:58:25 AM »
First thought: I'm no neopixel expert, I've only made the one pixel blade saber and I don't really intend to make more (component blades aren't my thing). Off the top of my head though, my concern would be waste heat dissipation. I was under the impression that the metallic strip/PCB that the individual addressable LEDs are mounted on are there in part to serve as a heat sink to keep the 144p/m (x2) from cooking too fast from the 15+ amps being pumped though them in operation. taking that away entirely seems to me like a recipe for releasing the Magic Blue Smoke.

Second thought: okay, so, LED string blades have mostly died because it was a right pain to bend the legs of 100 5mm LEDs and solder them together, and Neopixels came along and surpassed them in most aspects. It seems to me that what you're proposing here is taking the PITA complexity of making an LED string blade, and complicating it further by adding soldering to SMD sized pads. I dunno, it just sounds overly complicated to me.

Third thought: maybe replace the single core wire on the data line with stranded core wire. if the single core wire breaks, your whole blade will die up to the break.

Despite my negative outlook, I do look forward to seeing what you can do. This hobby thrives on people trying out their craziest ideas. Who knows, maybe you've got the next big thing here  :smiley:

Neopixel leds do not run at 15+ amps, it's a misconception. They run at 16-18 mA max per color, which is 50 mA max per each pixel total (it's 0.05 amps). 10-15 amps you are talking about is for the Positive and Negative power "rails" that go along the whole strip to power all 144 pixels. So if you multiply 50 mA (max each pixel can draw when it shines a White color - all 3 RGB colors mixed: Red + Green + Blue = White) by 144 pixels in the strip you will get 7.2 amps total draw per strip, and since we have 2 strips we get 7.2 * 2 = 14.4 amps per blade. Thats how much your blade connector and your wiring inside the saber must handle. But each strip individually only 7 amps.

Second thing - strips are not "metal core", they are regular flexible pcbs which just have wide Positive and Negative "power rails" on sides for same reason - handle 7 amps. And they get hot because heat is building up from each individual pixel (144 total) and also from high current flowing through these Positive and Negative "power rails".
« Last Edit: January 15, 2020, 07:01:28 AM by ShtokyD »

Offline SaintLuke

  • No Force
  • *
  • Posts: 3
  • Um...Hello?
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #4 on: January 15, 2020, 10:21:14 AM »
I would say just buy a ready to use strip from KR-sabers when it's released :) Because as you said in your design there are a few obvious issues:

1) very fragile, pixels joints points will brake immidiately when it's slightly bent
2) a lot of time to assemble the strip (and not easy to do), so it won't cost $30 really...

New KR-sabers strips are single-piece rigid pcbs (but still flexible in a needed range) designed from scratch specifically for use in sabers. I helped Khal to test and improve the strip design for absolutely max efficiency, so it will be the best and brightest neopixel strip on the market for sabers. Also they have pixel density even more than 144 leds/meter and are 8 mm wide with 5050 pixels. And it won't cost more than 2 regular flexible strips.

Thanks for the reply ShtokyD. I agree with you completely. Of course, I already bought the parts so I will probably make at least one blade just to see how it works. Definitely not for dueling.

I did think about gluing the top/bottom pairs of NPXLs together, which would add strength, since it would have 4 solder joints per pair.

Offline jbkuma

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 980
  • Pixels, everywhere.
    • Mad Science Workshoppe
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #5 on: January 15, 2020, 01:07:42 PM »
I look forward to seeing how these perform. Heat is definitely a major issue in real world performance.

Offline Onli-Won Kanomi

  • Master Force User
  • *****
  • Posts: 4689
  • Jedi Knight Of Faith
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #6 on: January 15, 2020, 10:54:11 PM »
What if one put a thin metal strip up the blade to act as a heatsink and mounted the pixel strips to both sides of that? Would that help with the heat issues?
To DREAM the IMPOSSIBLE DREAM. To FIGHT the unbeatable foe. To BEAR with unbearable sorrow. To RUN where the brave dare not go. To RIGHT the unrightable wrong. To LOVE, PURE AND CHASTE, FROM AFAR [-sigh-]. To TRY, when your arms are too weary; to REACH the unreachable Star!... This is my Quest; To follow that Star, no matter how hopeless, no matter how far...

Offline ShtokyD

  • Mining Colony Members
  • Force User
  • *
  • Posts: 220
  • whatever
    • My facebook page
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #7 on: January 16, 2020, 03:06:35 AM »
What if one put a thin metal strip up the blade to act as a heatsink and mounted the pixel strips to both sides of that? Would that help with the heat issues?
No, because it will still heat up but no way to dissipate the heat from the heat sink metal strip (must be only aluminum) since you have 2 neopixel strips from both sides and all enclosed in heat shrink wrap, cellophane, foam diffuser and plastic tube... There is no heat issue really, strips heat up and it's normal.

Offline Onli-Won Kanomi

  • Master Force User
  • *****
  • Posts: 4689
  • Jedi Knight Of Faith
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #8 on: January 17, 2020, 04:22:15 AM »
Hmmmm...yeah that makes sense that it wouldnt work to dissipate heat unless there were a way to conduct heat from the aluminum strip back into the hilt's mass the way in-hilt LED sabers do by connecting their heatsinks to the bladeholder which obviously doesn't work with present neopixel blade designs and so would require new kind of connectors on both the blade and hilt side of the system probably with separated electrical and heatsinking conduction paths eh?

But if heat isn't a problem with present pixel strips then yeah there's no real point to adding superfluous complexity for an as-yet unneeded solution. 
To DREAM the IMPOSSIBLE DREAM. To FIGHT the unbeatable foe. To BEAR with unbearable sorrow. To RUN where the brave dare not go. To RIGHT the unrightable wrong. To LOVE, PURE AND CHASTE, FROM AFAR [-sigh-]. To TRY, when your arms are too weary; to REACH the unreachable Star!... This is my Quest; To follow that Star, no matter how hopeless, no matter how far...

Offline jbkuma

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 980
  • Pixels, everywhere.
    • Mad Science Workshoppe
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #9 on: January 17, 2020, 09:17:22 AM »
Heat is very much an issue.

Offline ChrisFields

  • No Force
  • *
  • Posts: 32
  • Um...Hello?
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #10 on: March 20, 2020, 08:34:05 AM »
What's the heat issue? I'm still getting back into lightsabers myself, and like saintluke, i was happy to see neopixel and smooth swing. my normal job is thermal analyst in aerospace as a mechanical engineer, so would love to help with any thermal issues if i can.

Offline jbkuma

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 980
  • Pixels, everywhere.
    • Mad Science Workshoppe
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #11 on: April 01, 2020, 01:01:48 PM »
As LEDs heat up their output decrease and the wavelength of the output changes.  This is a result of physics not a manufacturing limitation.

Offline ChrisFields

  • No Force
  • *
  • Posts: 32
  • Um...Hello?
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #12 on: April 07, 2020, 02:16:48 PM »
How are they currently mounted? Or what is the current method of heat transfer off the LEDs?

Offline ChrisFields

  • No Force
  • *
  • Posts: 32
  • Um...Hello?
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #13 on: April 09, 2020, 11:37:43 AM »
Looking at what you have above... i know copper is heavy, but if you could mount the LEDs or back them with some copper strips running all the way down on the bottom of the LEDS, maybe .030" thk-ish... you would provide an extra path for heat flow, and at least increase the effective watt density of each LED which should help a little. You'll probably have to cover them in kapton tape so they don't ground or short anything. unless you can use them as a ground themselves. not sure how much it will help til you try it. plus it will help with hots spots that may happen, if that's an issue.

I do have the ability to run thermal models here, so if you need any help, let me know.
« Last Edit: April 09, 2020, 11:40:53 AM by ChrisFields »

Offline JonKClark

  • No Force
  • *
  • Posts: 40
  • Just a guy looking to make saber stuff!
Re: Custom Neopixel Strip Concept and Testing - UPDATED
« Reply #14 on: April 26, 2020, 06:36:04 AM »
I would say just buy a ready to use strip from KR-sabers when it's released :)
ShtockyD,
Any idea when they are going to be released? Just curious...
ThanksJon
Remember: Han shot first!
Jedi Atu'karek

 

retrousse