New Suit Project "Shama La Leon" (construction journal) (final attempt to post)

[Shoku]

Alright somehow the last two times I posted, the post vanished and was blank.  I guess you can’t edit a post.  So take 3!

I’m finally starting on a new suit I’ve been mulling over the idea of for quite a while.  “Shama La Leon” comes from the YouTube channel ZeFrank; it’s his mispronunciation of “chameleon.”

The idea is a plain white fursuit that has thousands of LEDs beneath the fur allowing me to change the color and pattern more or less at-will.  Here are the major bullet points I’m shooting for:

• Dual Color Camera Video Vision
  • This will preserve depth perception, increase viewing angle, and prevent the need for large holes in the fursuit head.
• Arduino-based Video Overlay
  • Surprisingly easy to do with NTSC video.
  • Displayed info includes:
    • System battery
    • Programmable pattern menu display
  • Possible control of other peripherals, such as LED hat and LED POV cane
• Raspberry Pi-based AMOLED Eyes
  • More or less a cut-and-paste of the tutorial on Adafruit, with a few tweaks for the ‘brain’ Arduino to be able to change settings.
• “Hall Monitor” Glove Control
  • A magnet in the thumb can be used to activate 4 hall effect sensors in various positions along the fingers for system control.
• Supersonic Music Trigger
  • This allows me to hide sounds above human hearing in songs to sync light animations to music.
• Teensy 3.2 light display control
  • Teensy 3.2 is able to simultaneously update 8 strings of NeoPixels up to 500 LEDs long.
• WS2812B lighting
  • Also called NeoPixels, individually addressable programmable volitile memory LEDs.
  • Once a color is programatically set on an individual LED, it will maintain that color until given a different command or power is lost.
• LiFePO4 Power Pack
  • LiFePO4, while less energy dense than Li-Poly, is far safer and ideal for wearable applications.
  • 2x Turnigy Nano-tech 3.0 Ah 20C batteries will be used for primary suit power.
    • 20C (capacity) refers to the safe discharge rate of the battery.  In this case, 20C means the battery is rated for 60 amps of continuous current.
  • 4x Turnigy 20A SBEC units will provide regulated voltage for the system.

Additionally, I am attempting to design the fursuit to be modular, so it can be worn as a full suit or as a partial in different configurations depending on what clothing I intend to wear.  To this end the lights will be broken down into 8 zones so they can be individually connected and controlled as necessary.  Each zone will contain no more than 500 NeoPixel LEDs, which also corresponds to the energy usage of 2 zones per SBEC (500 NeoPixels * 20mA typical current = 10A).

Because of the space required for the control circuitry and batteries, I am designing the bulk of the electronics to go inside a hollow tail.  The structure of the tail will be provided mainly by the loops of NeoPixel strips that provide its lighting.  Head-specific functions will be localized to the fursuit head (video overlay, raspi eyes) and the Hall Monitor glove is designed to work with other costumes as well and will be both stand-alone and wireless-capable.  For this project it will likely use a wired I2C connection.

I have mostly finished with the rough math on the project.  My next step is to begin testing individual systems (I’ve done a little bit with the video eyes already).  Heat is a big concern as the system is capable of a peak 80 amps continuous so at some point I will need to get my paws on an infrared thermometer (or a FLIR camera) to make sure the thing isn’t going to flash fry or slow roast me.

The first piece I plan to have functional is the tail, since it will contain the power pack and most of the electronics.  This will be followed by the head (mostly because I’m tired of being blind when wearing one).  The hand paws will be third, with the rest of the body, legs, and foot paws as the final stage.  As my work space is just being set up, I have only an early fur diffusion test to show for now.

Note that this video is marked as unlisted only because it is not up to my usual production standards.  You may freely share the link.

Further updates as events warrant 🙂

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Omni Possum]

Hm, spiffy. How are you inserting the overlay into the video signal, and what’s the delay like? I used a little cmos cam to cover my forward blindspot on an older suit but it always had that slight delay.

With all those lights maybe a reflective material would help deflect the bulk of the heat. Something like a one-sided reflective fabric would deflect heat in one direction.. Unsure how breathable/transmissive it would be from the non-reflective side though. But hey, PCM packs work pretty well for keeping cool. 🙂

Anyway, neat seeing someone else building technical suits here. n.n

-Omni

• This reply was modified 6 years, 7 months ago by Omni Possum.

[Shoku]

Because of the way NTSC works, you can white out pixels by simply spiking the analog signal in the correct timing. The circuit looks for the timing pulse and just superimposes a high signal at the right time.  I haven’t played with the overlay yet but the actual camera to video googles has no perceptible delay.

Today’s other small update is the battery charger. I’m using a Q6 Plus as recommended by a fellow techie fur, however I present exhibit A:

https://photos.app.goo.gl/y1R07uWEfsHXP1ag1

The connectors between the 2 batteries I have and the charger aren’t the same. Then I realized I’ve seen this before.

https://photos.app.goo.gl/Vv9o8LhqpakRZRQD3

These are actually solder tabs. They want me to make my own cable. So I plan to use Dean’s connectors (highest amperage connector I’ll be using) and I can create adapters for that as necessary.

The two batteries shown are intended for my light up top hat (large single pack) and my light up cane (thin dual pack) which aren’t actually part of the suit but use similar lighting systems.

• This reply was modified 6 years, 7 months ago by Shoku.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

(mistake post)

• This reply was modified 6 years, 7 months ago by Shoku.
• This reply was modified 6 years, 7 months ago by Shoku.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Omni Possum]

Spiffy. ^_^

[Shoku]

Well the batteries were out for delivery but never made it to me today, meaning I can’t design the tail enclosure just yet.  So instead, I ran another diffusion test.  The original test (YouTube above) assumed a 1 inch spacing, however 30/m LED strips are actually about 1.3 inch spacing.  This also gives me an excuse to set up and drive 150 NeoPixels at once and test one of my reels.

Another test I performed not seen in the video; temperature.  I’ve had the sequence running for about 20 minutes now and the LEDs inside the fur are just barely noticably warm.  The reel its self is what I’d call “safe to wash a baby in” warm.  I don’t remember where I got that from lol.  Interestingly, my power supply is showing a maximum draw of about 1.8 amps, where my math says it should be closer to 3 amps.  The system may be less power hungry than I expected, and this is a good thing.

Also, I re-ran the test at 5.5 volts.  2 reasons for this:  First, wires lose voltage over long runs meaning by the end of the run and I want to make sure we have at least 5 volts available for logic.  Second, the power regulators I’m using are jumper selectable and my options are 5 and 5.5 volts.  I am happy to announce the NeoPixels did not explode.

• This reply was modified 6 years, 7 months ago by Shoku. Reason: Changed "m" to "inch" because tired wolf pulled a NASA

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Omni Possum]

I like the striped pattern. Mm, only thing I can suggest for the  diffusion would be a flexible semi-transparent white plastic(vinyl?) sheet under the fur to act as a diffuser/softbox. Anyway, looks cool. ^_^

[Shoku]

The curse of the vanishing replies strikes again.

Today’s update includes a lot more light system testing.  It was about time to see how recognizable anything I put up on the furscreen was, so I did.

• Here we have a heart (aww) https://photos.app.goo.gl/Jq4rqYa69p8dJzbi1
• Here we have a wolf howling at the moon (aww) https://photos.app.goo.gl/rQRPJ8u6dsdjsKe13
• Here we have the Google Chrome logo https://photos.app.goo.gl/D5wYhQvBshovNbtU2
• Here we (kind of) have Mario https://photos.app.goo.gl/jP7HLZn7y2AlsDbX2

All these images were created by me manually typing in color values and pixel locations into the console.  Here’s a view of the workflow for that. https://photos.app.goo.gl/SvVe9RLUo2dQAFQv1  The system was set up so if I type in a number it either sets that pixel to the last color code I used, or turns it off.  Appending a number with a hash told it I was sending a hex color code.

While that system sort of worked, the whole idea of the project is to use a program called Processing to generate still frames and animations for the code to run off an SD card.  Today I finally got that portion of the project working and here are the results:

Here is a look at Processing in action as it creates that same file:

Because the overhanging pixels were not set with a location, I think they’re defaulting to <0,0> which is the top left corner of the video.  In the final fursuit there won’t be any unused, unapped LEDs so there shouldn’t be any stray colors.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

Another minor update with 2 more videos.  First, here’s a fire effect on the fur:

And second, here’s a timing test.  I added code to play the video at 25 FPS, but there’s a problem.  Either Processing didn’t actually output 25 fps, the source video isn’t really 25 fps, or the crystal on the Arduino Uno is slow because I had to speed my video up to around 115% of what I shot it at to get it to sync to the actual video (Rymdreglage, 8 Bit Trip).

The timing is based on the millis() function which counts the total number of milliseconds since the board was powered on.  Because I measure from the first frame rendered instead of from frame to frame the frame delay SHOULD be a perfect 40 milliseconds (which comes out to 25 fps) which my console output shows but within just a few seconds the videos were out of sync.  More experimentation will be necessary.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

Most MEMS mics you can buy are digital I2S devices.  I managed to order pretty much the only analog one Amazon sells.  *headdesk*  It cost me $8.50 and the shipping to return it is $8.  …. I think I’ll keep it.  Though I now have the CORRECT mic ordered and I’ll be able to start testing hidden audio cues in about a week.

Today’s experiment is about heat.  The code from my last post runs through all the animation loops I’ve created nonstop, and so makes for a good stress test of the system (so far).  I let that run for about an hour and then turned my new toy…. I mean new tool on it.  My thermal imaging camera.  The first picture shows one of the lights hanging past the fur.  The second picture is the fur its self.  The third picture is the power wiring junction (left in Celsius because I was testing my 3D printer before hand).

https://photos.app.goo.gl/rGiaFrZwdcE5tH5p2
https://photos.app.goo.gl/4S6ddNIPk6m8XKwA2
https://photos.app.goo.gl/FzoGTC8b3iquwPZ73

As long as we’re at it, here’s a picture of my 3D printer working on a test piece for an unrelated project.  I find it quite interesting that you can clearly see where the fan is cooling the heated bed.

https://photos.app.goo.gl/jRuLpTbpcI2aqe7H3

This did get me thinking a lot about thermal properties and I’ve decided I need to add a cooling fan to the tail.  Talk about your tail wind.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

Today is all about sound, and ideally sound you can’t hear.  I got my MEMS mic successfully hooked up and working with the Teensy 3.5 board (after running out of curse words and needing to invent some I might add).  This first video is showing an FFT, basically a band analyzer showing the strength of the sound it detects at various frequencies.

And the second video is the actual tone detection test (warning, high frequency sounds are annoying, don’t play loud because I decided to test lower frequencies than 20K too).

My next test will be to see if I can reliably use the tone to reliably sync an animation track to a music track, but that will come later on.  However, this is the system that will let me auto-sync animations to music tracks for doing performances (of which I have a few kicking around in my head already).

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

Well a little while since my last post.  Tonight’s is mostly info and pictures.  Sorry, no cool new videos just yet.

I’ve been mainly working on collecting supplies and organizing my work area.  Now that my project for Azure Boxcat (https://photos.app.goo.gl/XJPuwOQtTC3H45j43) is done I can focus entirely on Shama again.  To that end I have obtained the Deans connectors, a bunch of shrink tubing, the heavy and light gauge silicone wire, and the rest of the power plant.

https://photos.app.goo.gl/C00I4ghUuoBCxzYv2 (SBECs)
https://photos.app.goo.gl/6zVvaWMi2COC76Wr2 (replacement battery connectors)

Past that I’ve been working on drafts of the tail skeleton.  Both the rigid section (https://photos.app.goo.gl/NtcV4ifaoSqzuMHH3) and the flexible “snake” portion (https://photos.app.goo.gl/jvXjk9LSFLh5IuwW2).

There will be a base mounting plate of some sort worn on my person.  Ideally I want to contour this to my lower back (I’m a stickler for anatomically correct tail placement, much of a pain in the ass as it is) and I’m looking at photogrammatry software to scan my back to create the fitted plate from.  The plate will serve as a connection for the universal joint (which will house the power cooling fan) connecting to the electronics portion, as well as a terminal block for the rest of the suit lights and data lines.

I opted to go with a universal joint first to hide the fact that a sizeable portion (149mm) of the tail is actually rigid.  Beyond that the rest of the tail will be flexible.  I’m intending to simply use soft upholstry foam as compression springs to give snake portion a return-to-center type of effect and resist motion just a little bit.

Much as I was hoping to have some lights working in time for tomorrow’s picnic, there just wasn’t time.  Issues with the printer and rushing to finish the cat’s cooling gun prevented me from getting in as much proto time as I would have needed, to say nothing of waiting for payday.  Which was today.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Shoku]

Side note, I’m doing a few upgrades to the printer to speed things up.  I’ve added set goto points to the OctoPrint interface to allow me to level the bed (https://photos.app.goo.gl/xv7K12gephRN1cqy2) and I have a digital depth probe ordered from Amazon which will allow me to level the bed much faster than I can do using the feeler gauge method.  I am also considering printing at least some of the tail pieces from carbon fiber infused filament, but that stuff super expensive (generally $80 plus per 1kg spool) so I haven’t gotten any of it yet.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Omni Possum]

You actually might be able to use a thermoplastic like Worbla to make a base plate and save some render time. Unsure what the lower end of it’s molding temp is though..just don’t put it against bare skin while it’s hot. Once you get the shape you could layer or reinforce it with another material, just a thought. The tail is looking neat….actuated?

[Shoku]

Not actuated, at least not in this version.  In the past I found simply creating a spine-like core was enough to get some pretty decent movement out of it.  Also I haven’t liked most of the Bowden cable movement I’ve seen.  The tip always curls before the wag motion starts, doesn’t look good to me.

"The problem is not that there are too many idiots in the world, the problem is the distribution of lightning."
- Mark Twain

[Author]

Posts