Makerbrett

A story of making things with technology and fire. I'm Brett Bond. I'm founder of LiveSpark, a special effects company producing flame effects, Velocity, a motion sensor analysis company, and Concrete Interactive a software development agency in Santa Monica.  I love to make, or at least learn how to make, everything I can see in this world. This is where I share what I'm doing and learning, and I invite you to join me in these adventures.

[makerbrett]

View this post on Instagram

🔥🎃🥁What happens when a special effects company invites a couple drummer friends over for Halloween? (at Venice, California) https://www.instagram.com/p/BpiEtR6H6fs/?utm_source=ig_tumblr_share&igshid=1md0xeif7niye

[makerbrett]

For my daughter’s bed it’s important to have some solid railings since she’ll be up high. I decided to use finished select pine boards with dowel pins. I made a small dowel jig so drilling holes for the rails would be quick and uniform and I used my digital calipers to measure spacing on the upper and lower boards.

[makerbrett]

This very knifey looking thing should win any LARP battle I encounter but in fact it's a torch intended for a dance group. This prototype will be the first of 7 or more fully self contained, finger trigger flame effect systems. The interface is a set of Force Sensitive Resistors, a disc about the size of a fingertip pad that changes resistance when it is pressed. What separates these FSRs from normal buttons is that in place of just an on/off state, the FSR gives a full range of values, which in turn we use to determine flame height. With my friend and expert mechanical engineer Mike Thielvoldt, we were able to knock out this prototype in just 3 days. Our tools and techniques were primarily CAD, waterjet, bending and just a little MIG welding. I did the design in Rhino, with Mike putting the mechanical drawings together in SolidWorks, both tools applied in my opinion for the tasks they do best. We chose lightweight and easy to obtain 6061T6 aluminum for the vertical slats, and 14ga cold rolled steel for the 5 star plates that hold everything together. Control is via an Arduino Leonardo, with a custom shield that performs the RS-422 communication with the FirePixel. A 6S Lithium Polymer battery supplies ample power at about the 24V the FirePixel requires, and we step that down to 18V suitable for the Arduino power using a linear voltage regulator. Fuel is supplied by a single 14oz propane tank mounted in the torch base, regulated to 30psi and sent up via LPG rubber hose through the torch handle and up to the FirePixel. Flow rate is 72 Liters Per Minute maximum which creates 4 feet of flame in still air. Weight is around 22LB, something I'd like to reduce, but anyone handling this sort of thing ought not to flinch at that poundage. Software takes sensor input and in real-time mode transmits commands to the FirePixel to adjust flame height. The squeezing motion is quite precise and natural. A second sensor starts pattern recording which can then be played back in a loop. You can store two patterns at a time and trigger either one of them with a touch of the middle or ring finger buttons. Next steps are to build more, and find a dance group that will be creative (and not scared) with these so we can make a cool video. Look for that as soon as I can get some good dancers signed up, or contact me if you're interested.

[makerbrett]

For a project I'm working on, I need control from a Raspberry Pi embedded computer to an 8-way relay. This device is capable of switching large amounts of current, high voltages, and in a compact package that can be controlled by the Raspberry Pi's input/output pins.

As part of the project, I also wanted to see how fast I could switch these relays on and off from the ruby programming language. I discovered a nice little ruby library (called a gem, in ruby parlance) that makes it very easy to control the inputs and outputs. The 8-way relay can be purchased for about $13 here. 

And for my fellow coders, here's my little ruby program that is running in the video below.

require 'pi_piper' $pins = [9,10,11] $relays = [] def setup $pins.each { |pin| $relays.push PiPiper::Pin.new(:pin => pin, :direction => :out) } end def flash(relay, delay) relay.on sleep delay relay.off end def flash_forward(delay) $relays.each {|relay| flash(relay, delay) } end def flash_reverse(delay) $relays.reverse_each {|relay| flash(relay, delay) } end setup n=1 while n < 10 (1..20).each{|d| flash_forward 0.1*(1/d.to_f) } (1..20).reverse_each{|d| flash_forward 0.1*(1/d.to_f) } n=n+1 end puts "Program complete."

[makerbrett]

Power is always a question that comes up when connecting large numbers of FirePixels, or anything that draws significant current in fact. 

What we've attempted to achieve with the LiveSpark PowerPlate, is a simple solution that accepts multiple chains of effects. By internally handling data and emergency stop, the Controller remains in charge of the effect display, and a single flip of a switch can kill hundreds of effects in an instant.

This video is a quick tour of how the PowerPlate works in combination with 3 chains of FirePixels and their Controller.

[makerbrett]

Theme parks certainly call for spectacular special effects, and so when asked to partner with Orlando Special Effects at the International Association of Amusement Parks and Attractions (IAAPA) show in Florida this year, I said yes.

But I wanted to create something that would make people understand the versatility of our fire technology, so showing a line of FirePixels across a piece of truss was not going to be good enough.

I eventually settled on building a hanging pyramid of fire. It turned out to be a great balance between showing a real application while still leaving room to imagine what else could be achieved with these effects.

Here's a short video on how we made it. To begin with, I wanted to build all plumbing and cabling inside the tubes that make up the pyramid shape. The trick is all of this had to flat pack for shipment to Florida. 

It took many extra connectors, plumbing fittings and cables, but in the end the look was cleaner and more professional, without wires and hose sticking out all over the place.

Using Hollaender speed rail cast aluminum elbows allowed me to create the pyramid out of re-usable elements. I also decided to put a flat black powder coat on all the pieces, including the elbows and tees so the emphasis was on the FirePixel, not the parts and pipes. 

Thanks to Orlando Special Effects for partnering with LiveSpark to produce a killer booth.

[makerbrett]

No other holiday says fire like Halloween, and we love to get creative this time of year. We were having some fun with pumpkins over at TechShop, Arizona. TechShop is a great place for makers, and even us nerds love to get down. Yes we use dremels instead of those cheap-o orange pumpkin knives of yesteryear :)

[makerbrett]

If your fire shooting item has a problem—it should text you.

When working with the City of Las Vegas Fire Marshal to obtain a permit, we pursued what is known as a Group V flame effect classification as described in the National Fire Protection Association (NFPA) 160 code. This classification states, "A large individual or group flame effect that is permanently installed inside or outside any structure designed for unattended operation without a main show supervisory control system."

Essentially, our Las Vegas Blvd sculpture is open to the public and should not be climbed. So we built a proximity sensing system using an ultrasonic sensor that can detect a person near the base of the sculpture.

If a person gets too close, the monitoring system signals a site called Xively.com with the data containing the person's proximity. When that data goes below a "too-close" value, xively sends an API message to another site called Zapier.com. Finally, Zapier passes the message on to a text messaging service called Twilio. Please not all three of these companies follow the standard issue 3-vowel, 3-consonant naming convention for such enterprises.

In short, all this happens within a second or two and sends a text message to the flame effect operator on duty—this person doesn't have to be present, but they do have to have their phone on.

[makerbrett]

For years people have said "Oh like a Bellagio Fountains of fire..." so it is high time we put our fire in the pool. Here's a concept rendering straight from the paintbrush (cad modeler really) of LiveSpark designer Erik Walker. We also imagined a home version, cliffside infinity pool not included.

[makerbrett]

If you love something...drown it? I'll admit it, the first few times I performed these tests I was afraid I'd break my beloved creation. And I did. We had water leaks inside the case, sparks were shooting all over the place, corrosion set in on the circuit board and some internal components remarkably quickly. In short, we broke a bunch of things that were expensive and hard to build.

But after patching, more testing, and finally redesigning components and assembly steps, we were able to pass more and more rigorous tests. It became fun—break it, take it apart, fix it, try to break it again until it just doesn't break.

The last thing in the world I'd ever want is to create something dangerous. Fire is always dangerous you say? Well of course, that is true. But when treated properly the risks can be limited to safe levels (similar to gas tanks in cars). What I mean by dangerous is something that behaves unpredictably, so that it may not be possible to mitigate those risks.

When it comes to risk assessment, there is no substitute for real-world testing. There are some benefits of course to third-party testing, and it is on our shortlist to pursue independent validation of our results. However, to my mind it is more important that we perform the tests ourselves and get them to pass, so that I can personally know what the product is capable of, and what makes it fail.

And product failure in itself is ok; failure can be predictable. Drive over it with a forklift and it will fail. But it should always fail gracefully. A good way to achieve graceful failure can be derived from a methodology known as Failure Mode Effect Analysis.

For example, we use a microcontroller, an electronic component that is like a small computer, inside each FirePixel. And it is quite common for a microcontroller output pin to fail, say if it is overheated. When one if its outputs fails it may be stuck in a high or low state (a binary 1 or a binary 0). If that output happens to be controlling our fuel valve, that could mean that the FirePixel just blasts fuel even when it's not supposed to. However, that is not what happens thankfully due to a practice of FMEA.

By making the valve driver circuit depend on a changing state (flipping between high and low repeatedly), the only way the valve may open and release fuel is based on the rate that its input signal changes. So in the case where the microcontroller output sticks to a single state, either high or low, the valve will always close.

[makerbrett]

As we finalize water tightness of the Model 11 FirePixel, the possibilities of fire in water become enticing- even in a dumb old green bin I think.

[makerbrett]

Our newest software tool Altium renders beautiful 3d images of our circuit board. If only it simulated our fire too :)

[makerbrett]

A precision machined tool helps align the igniter position. The ends are milled to within 0.001" which takes a difficult measuring task and turns it into a simple assembly task.

[makerbrett]

Planning for IAAPA, the trade show in Orlando, Florida focusing on theme parks and amusements it is also a meeting place for the special effects industry. This image shows a banner we made for the show last year. In real life the middle FirePixel is missing from the banner and there is a stand holding an actual physical unit in front of it. 

Last year we went with a simulated approach showing the banner and a video because bringing live fire to the convention center is so much more involved. This year we're going for it, and LiveSpark will present a live fire display with our partner Orlando Special FX. More to come as we develop our booth and presentation specifics, but it's definitely going to be a lot of fun.

[makerbrett]

Scenes from FirePixel assembly. Getting the pipeline up and running to scale LiveSpark's manufacturing process.

[makerbrett]

In my quest to learn about manufacturing and become self-sufficient enough to make at least a few dozen FirePixels each month, I have started building some more tooling to help various steps of our assembly process. It’s amazing how much faster and more consistently you can build a sub-assembly just by holding a bunch of them as I’m doing here with igniters, or lining two things up with a jig that removes measurement from the process. Sometimes it’s not the tooling but rather the order of steps that matters. The first burner we made here in Las Vegas took 2 hours, and ended up with slightly bent fins. By adding a tool and change the assembly steps we just made 12 of them in about 10 minutes total.

[makerbrett]

A reel that sizzles. When putting this together I culled as much video from past projects as I could readily get my hands on. It's true there's not as much detail or explanation, and I added a bunch of reaction shots.

The effect is trying to make people think the fire is neat, and want to know more. Other videos go into greater detail or show more definitively what LiveSpark does.