fiduciary

Author Topic: Neopixel blade assembly - an experience exchange board  (Read 16801 times)

0 Members and 1 Guest are viewing this topic.

Offline Obi_1

  • Mining Colony Members
  • Experienced Force User
  • *
  • Posts: 465
  • Creator of DIYino - first open source FX-board
Neopixel blade assembly - an experience exchange board
« on: February 05, 2017, 12:53:00 PM »

I start this thread as a general board to exchange ideas/experiences/recommendations and tons of pictures on the subject of:

How to assemble a neopixel blade?

Neopixel stripes are around for some time now. Since their debuting (see LED string saber evolution - this might change the way we think about LED string ) in early 2016, a lot of lightsaber builders took up this new technology due to it's unparalelled possibilities compared to both high-power LED and segmented LED string blades.

Quite a long time passed since. After DIYino Prime v1, the first lightsaber board to fully support neopixel blades, other saber board manufacturers also jumped on the train, so it can be safely said that now it is the 3rd pillar of lighsaber illumination (the first two being in chronological order: 1. segmented LED strings and 2. high-power LED type blades aka Luxeon/Cree)

This thread is about how to assemble such a blade. It is supposed to give others not yet familiar with this build type an idea:
- where to source good quality neopixel stripes
- what is their theory of operation
- common pitfalls
- brigthness and supplying schemes
- etc.

as well as to offer experiences builders a platform to exchange new ideas.

Offline Obi_1

  • Mining Colony Members
  • Experienced Force User
  • *
  • Posts: 465
  • Creator of DIYino - first open source FX-board
Re: Neopixel blade assembly - an experience exchange board
« Reply #1 on: February 05, 2017, 01:26:40 PM »
Now, after this introduction, let's have the first post.

A disclaimer: I'm a hobby sabersmith and a saber board designer. But I'm far away from being a skilled blade builder. I made my share of segmented LED strings and multiple neopixel based builds, and I'm happy with the results, but I'm fully aware - and quite happy about - that other are way better at it as I'm, so I can learn from them.

So here is my simple neopixel blade assembly guide to launch the topic:

Rules:
- buy good quality neopixel stripes. I warn from cheap sites, especially some sources from AliExpress have dubious quality. My latest one I sourced from this site: http://www.ebay.de/itm/191956702454?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT . It is a German site, but what I wanted to share is that since my first neopixel posting price of the sripes drop fast, now for 40$ you can have 2m of the densest stripe, one year ago it easily hit the 100$ mark.

- try not to solder the stripes. I have some skills with soldering, but the stripes are not made for it. You can cut, not problemo, soldering it is a xxx! Nevertheless with some skill you will manage it. It a bigger problem that it will break the integrity of the strip, which might impact dual worthiness or generaly speaking robustness of the blade.

- buy the 144LED/m type, less dense stripes are great, but not for sabers. You need the dense stripe to avoid pixeling.


Now to some pictures I took from my latest blade assembly:

Most stripes come with connectors on both sides to be able to chain multiple segments together. This comes in handly, because instead of soldering directly on the stripes you can join the wires of the connectors. The - so far - most common design is to attach 2 stripes back to back, with the first pixel of both stripes at the bottom of the blade. Wiring is easy: connect GND to GND, 5V (supply) to 5V and DIN to DIN, as seen on the next pic (the stripes I ordered so far all had wire assignment as follows: white - GND, red - 5V (supply), green - DIN (serial input)


The basic unit of a stripe seems to be a 0.5m long pixel chain, which is attached to another with the same size, so all stripes are assembled using 0,5m segments. At the joint of the segments there are bigger landings for soldering (see next pic, in the middle), making the distance between two adjecent pixels greater than at other positions along the stripe. My first though was logically, that this distance will appear as an irregularity in the blade, therefore I started to desolder, cut, solder again. I s***d up and had to buy a new one, so I do not recommend doing so. In my newest build I did not touch it and I can conform: no irregularity is visible, this distance is still way too small to be recognized.



If you choose the back-to-back design, attaching the stripes is made confortable by the double sided, clear adhesive tape present on many of the commercially available stripes. I choose to attach them so that there is a pixel on a side between 2 pixels of the opposite side, i.e. interleaved. I figured that it migth give an even smoother illumination along the blade. The adhesive is insulating, therefore you do not have to worry about the small landings on the back of the stripe touching, but even if your stripe does not come with adhesive, or it happens to be non-insulating or bad quality, an interleaved configuration makes sure the landings do not overlap (you only have to be cautions at the soldered joins as mentioned above).



As you can see, assembling this basic design is not at all difficult, just make sure you align the stripes and press them well together before moving onto the next few inches.

A note on voltage drop and adding wires: I saw few people mention that their stripe seems to be more dark at the end. This can well be due to voltage drop. A logical solution - which led me to extensive soldering and frustration - is to add additional wires to lower the IR drop along the stripe. It can be done, migth help, but there is a drawback: if you but these wires between the sripes, there will be a gap. And you do not want that gap, because it will be visible as shadow in the blade. For achieving a good and uniform brightness along the blade, pressing the stripes together best as possible is crucial.

I will do some more pics on how I diffused the blade. I used basically the same technique I applied to segmented LED string blades:
- TCSS transwhite blade
- with the accompanying diffuser rod (!!!)
- and a packing foam wrapped around the stripes.

To the number of LEDs/pixels: a standard 32" blade needs around 120 pixels in a row (i.e. 240 if you use the back-to-back technique), that is why I geared LSOS to support that number or less.

Offline jbkuma

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 804
  • Pixels, everywhere.
    • Mad Science Workshoppe
Re: Neopixel blade assembly - an experience exchange board
« Reply #2 on: February 05, 2017, 02:52:06 PM »
My best blade yet was two strips back to back, 4 layers of white packing foam, 1 layer of toilet paper, 4 ft of clear gift wrap, clear thin wall blade.  The hardest part of construction is definitely cutting the TP.  My blade was perfectly even with no noticeable shadows, except for a ragged seam in the paper.  I considered trying PTFE heatshrink as a diffuser, but it's pretty expensive to try.  I have also been considering ordering a 7/8 transwhite tube to use a diffuser within the clear tube.  Without the TP the blade looked pretty good, except from directly viewing perpendicular to the plane of the strips.  The individual LEDs were visible from dead on, but not when viewing at an angle.  Adding a 6th layer of the foam didn't help with diffusing very much.

From my experience in prototyping, I'm pretty sure the "browning" effect is due to wiring issues or poor battery.  Once I moved from my lower gauge jumper wiring and breadboard to 28awg soldered connections the blade lit up perfectly without any dimming.  If insufficient wiring or connectors are used, this can easily result in dimming.  Secondly, not all batteries can deliver the current required to light 240 LEDs.  A quality cell is essential.  It goes without saying, that if powered from a bench or other plugged in source during testing, the source most be able to deliver the proper current.

Unfortunately I don't have any worthwhile photos to share, but I'll be sure to document the construction of my next version.  I may swap out the TP for waxed paper, which should be easier to cut cleanly and stiff enough to make into a tube.

Offline profezzorn

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 886
  • May the source be with you.
    • Hubbe's Corner
Re: Neopixel blade assembly - an experience exchange board
« Reply #3 on: February 05, 2017, 04:55:30 PM »
The best way to tell what's causing browning is probably a multimeter.
Just power the whole thing on and measure the voltage at the base and end.
If the voltage at the base is low (less than 3.7 on a full battery) get a new battery.
If the voltage at the end is significantly lower than at the base (I have not measured this myself, so I don't know what to expect.) resistance in the blade itself is the problem, and adding wires might help. (Or getting a better strip.)

Offline theholyduck

  • No Force
  • *
  • Posts: 40
  • Um...Hello?
Re: Neopixel blade assembly - an experience exchange board
« Reply #4 on: February 05, 2017, 04:58:01 PM »
How well do you find those i assume 10mm wide strips work in the blades? I know profezon used 7mm wide strips. On 1 hand it seems narrower strips would allow more diffusion foam and thus better diffusion. on the other hand. wider strips have more thermal mass and wider traces for more current carrying capability.

Offline profezzorn

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 886
  • May the source be with you.
    • Hubbe's Corner
Re: Neopixel blade assembly - an experience exchange board
« Reply #5 on: February 05, 2017, 07:32:51 PM »
I don't think more copper is going to make much difference for heat, since it's all going to be wrapped in some insulator anyways.
It could make a difference for current, but it hasn't been a problem with mine. There is a voltage drop between base and tip, but any unevenness in light/color is pretty minimal. (And only when the battery is close to the cutoff.)

Offline Obi_1

  • Mining Colony Members
  • Experienced Force User
  • *
  • Posts: 465
  • Creator of DIYino - first open source FX-board
Re: Neopixel blade assembly - an experience exchange board
« Reply #6 on: February 06, 2017, 01:49:37 AM »
On that note it's worth to mention that especially protected batteries might not be able to deliver the current needed for 240 pixels. In my current build I had to limit the PWM to 150 (max is 255), because otherwise the protection circuitry shuts down the battery. Not the case if you have an unprotected, high draw cell (but everyone should try at own risk!).

As to the voltage drop, I did measure and did not see any drop worth to mention between base and end, i.e. the drop should be minimal over the quite broad copper traces. However your wiring/mosfets/everytjing else in the way will make a difference, even with a fully charged 18650 I measured only ~3.1V at the base when switching on the blade.

It just tells that an optimal supply of the neopixels is still to be found. I will definitely redo my build with an unprotected cell (at my risk), in which case I need to see how  much brightness is coming from the pixels and how much from glowing wires ?  :evil:

Offline Azsde

  • No Force
  • *
  • Posts: 26
  • Um...Hello?
Re: Neopixel blade assembly - an experience exchange board
« Reply #7 on: May 19, 2017, 03:10:48 AM »

I will do some more pics on how I diffused the blade. I used basically the same technique I applied to segmented LED string blades:
- TCSS transwhite blade
- with the accompanying diffuser rod (!!!)
- and a packing foam wrapped around the stripes.

I have a few questions:

  • What thickness did you get for your blade ?
  • What is the diffuser rod you speak about ?
  • What is the thickness of your packing foam and how many layers did you wrap ?

I thank you in advance for your answers.

Regards,

Azsde.
« Last Edit: May 19, 2017, 03:18:08 AM by Azsde »

Offline jbkuma

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 804
  • Pixels, everywhere.
    • Mad Science Workshoppe
Re: Neopixel blade assembly - an experience exchange board
« Reply #8 on: May 19, 2017, 06:48:35 AM »
I've done quite a bit of experimenting, and I've built and rebuilt about 5 or 6 times now.

Trans-white blades don't seem to significantly improve diffusion in the way that is necessary for eliminating the visibility of the individual pixels.  Shadows in general don't really seem to be the major problem everyone theorizes they should be, but to my eye it actually looks somewhat worse with the trans-white than with clear.  In the end, the effect of the trans-white seems similar to what you'd get from a regular star LED: a thicker looking blade.  I'll have more on this once I've given the blade the silicon treatment.  I've never been a fan of the white blade, so for a standard blade (not covered in silicon) I think I'll stick to clear blades.

Experiments with 3D printed tips have yielded some desirable results.  I'm printing hollow bullet shape tips in clear (natural) PETG and with a few layers of thickness, the layer lines do a great job of catching and diffusing the light.  They aren't as nice looking as a the machined tips from TCSS, but they come much closer to matching the blade when illuminated.  The only catch hear is that since it is clear the layer changes are pretty obvious as the plastic is denser there.

There are some more tricks up my sleeve, and I'll share more about then when I've had a chance to test them.

Offline Master Cram-Fu

  • No Force
  • *
  • Posts: 17
Re: Neopixel blade assembly - an experience exchange board
« Reply #9 on: June 18, 2017, 11:31:21 PM »
So, I'm relatively new to this, but I just clear heatshrunk two strips back to back, stuck them in a tcss thick wall diffuser, then put that inside a tcss thin wall diffuser, and that went in an enhanced blade.. Here's some pictures with the enhanced red.

Here's a picture of the blade up close. Ignoring my scratches from trying to jam it into the emitter, you can see that there is no corn cob effect.


Unfortunately, there is an obvious line down the side that is darker. You can see it here compared to the darth vader master replicas. Keep in mind, these are both at 1/8" brightness, so the RGBW setting is (31.875 0 0 0). The darker spots are a bit of a tradeoff for brightness... either way, its quite a bit brighter than the MR on 1/8 brightness (which is good enough in my book!). On full power, it will light up a dark room with ease, but it dies in about 5 minutes haha...

As far as difussion goes, I did attempt to use several combinations of 1/16" thick packing foam and light box diffuser to no avail... I always ended up with a very obvious line where the foam and/or fabric was cut.





BTW, am using this battery: Turnigy Graphene 950mAh 1S 65C LiPo Pack w/ JST-SYP-2P (AU Warehouse)
Also, here is a link to the exact strip that I purchased: Adafruit NeoPixel Digital RGBW LED Strip - White PCB 144 LED/m [1m] ID: 2847 -

I wouldn't really recommend the rgbw due to cost. You can definitely make some different color combos with the extra diode in each LED, but you cannot really crank it without heating up to much. My thermocouple was clocking the back to back strips on full blue at around 90F in a 75ish degree room before putting it in the blade, so I'd imagine it would get very hot with all of the LEDS on (not going to risk it).

Anyhow, hope this is of use to someone. I'd still like to find a better way to diffuse these without taking apart the master replicas and using the tube from it. I have also built a traditional string blade using some 120 degree 5mm red leds from a website called "www.ledssuperbright.com" which appears to be dead right now, otherwise I'd link them. They were inconsistent in brightness but about as bright as the neopixel strips on full brightness while using much less power than the neopixel strips. I may be able to hook that strip up and show you guys if there is any interest.

Offline profezzorn

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 886
  • May the source be with you.
    • Hubbe's Corner
Re: Neopixel blade assembly - an experience exchange board
« Reply #10 on: June 19, 2017, 01:04:11 AM »
The slim neopixel strips are easier to diffuse, but I'm not sure if it's possible to completely get rid of the side-shade. (Although I'm using 7/8"-blades, which cut into the diffusion a bit.) Building a string of PL9823 or APA102 leds can get around that, but it's a lot more work.

Offline Sligs78

  • Padawan Learner
  • **
  • Posts: 56
  • Um...Hello?
Re: Neopixel blade assembly - an experience exchange board
« Reply #11 on: June 29, 2017, 12:30:27 AM »
I know it's suggested to use a capacitor to mitigate the voltage spike when turning on the neopixel strip, but does anybody use one?

Offline Obi_1

  • Mining Colony Members
  • Experienced Force User
  • *
  • Posts: 465
  • Creator of DIYino - first open source FX-board
Re: Neopixel blade assembly - an experience exchange board
« Reply #12 on: June 29, 2017, 02:58:10 AM »
I know it's suggested to use a capacitor to mitigate the voltage spike when turning on the neopixel strip, but does anybody use one?

The stripes already have at each pixel an SMD cap, I guess they take care about the spikes. I haven't seen so far anyone actually using extra caps other than those on the stripes or the driving boards.

Offline Azsde

  • No Force
  • *
  • Posts: 26
  • Um...Hello?
Re: Neopixel blade assembly - an experience exchange board
« Reply #13 on: July 05, 2017, 01:44:23 PM »
I am having a huge issue with my build, I have an arduino nano clone, it's being powered by a 3.7V 5200mAh battery.

Since it is not powered in 5V, it seems that the clock doesn't work as intended and prevent all timed and pwm related feature from working.

What solution do I have?

Offline profezzorn

  • Mining Colony Members
  • Master Force User
  • *
  • Posts: 886
  • May the source be with you.
    • Hubbe's Corner
Re: Neopixel blade assembly - an experience exchange board
« Reply #14 on: July 05, 2017, 04:11:35 PM »
Either use a booster to get the proper voltage, switch to a board that can handle the lower voltage. (Like a TeensySaber V2, which has a booster on it.)