Halloween Experience
   
         

  > INTRODUCTION
a fascination with imagination
  > THEME PARK MAGIC
inspirational rides and attractions
  > A PHANTOM PLOT UNFOLDS
a 2003 test from Disneyland ideas
  > THE HAUNTING BEGINS
five years of Halloween, 2004-2008
> BACK FROM THE GRAVE
2010-2011 with new technology
  > A 2013 RESURRECTION
a mix of new and age-old effects
  > DAWN OF THE UNDEAD
2014 show, part one
  > THE ZOMBIES EMERGE
2014 show, part two
  > A CHANGING CONCEPT
2015 show, part one
  > BUILDING PNEUMATIC FIGURES
2015 show, part two
  > ROLL UP, ROLL UP!
2015 show, part three
  > INTO THE TUNNEL...
2016 show, part one
  > MAKING MONSTERS MOVE
2016 show, part two
  > TO HELL AND BACK
2016 show, part three
  > SHARPENING THE SENSES
2017 show, part one
  > A MAGNETIC ATTRACTION
2017 show, part two
  > THE BIG EXECUTION
2017 show, part three
  Improving movement with pneumatics    
  In the last two displays, I had re-introduced movement in some of the figures, however these all used electric motors with limited torque. This meant that only parts of these figures, such as heads and hands, could be moved easily. Using compressed air instead would produce much bigger forces, allowing whole top halves, or heavier sections of figures, to rock or move about.

To create this movement, David introduced me to building McKibben air muscles. These were invented in 1957 by Joseph McKibben, originally to be used to help move artificial limbs. One of David's projects was an impressive set of animatronic legs, moved using McKibben air muscles, to provide life-like human movement.
 
The construction of these muscles involved wrapping a piece of bicycle inner tube with electrical braiding, then plugging both ends with pieces of dowel. One end had a hole drilled through it where the air pipe was glued in. The braiding was then fastened securely to the ends with jubilee clips. Inflation of the tubing would cause the braiding to widen and pull the ends of the muscle inwards.
 
    When compressed air enters the muscle, it contracts lengthways. Then, when emptied of air, it returns to its original length. By attaching these to hinges and levers, movement can be produced.

< Here is one of the first muscles I made. It had a length of about 25cm, and a contraction of about 4cm when inflated. To reduce air leakage, the two dowel bungs needed to be made slightly conical, with the wider end towards the middle of the muscle. This also helped to prevent them from being pulled out when the muscle was in operation!
 
  But how would these muscles move the figures? One creature that I came up with was the ‘wolfman’; an ill-fated hairy fellow who had undergone some unfortunate transformation and ended up with lupine features! I wanted him to rock back and forth and rattle his cage as visitors approached.

> This is one of my early drawings of the figure, showing him with bent knees as if on hind legs, and a hinged torso. I'd originally thought about mounting the air muscle at the back, pulling the body upwards from behind...
   
    ...but David suggested that the muscle could be wrapped around a cylinder, where it would unwind slightly on contraction. So I made sure each figure had a horizontal bar to which the hinged part attached; the muscle could then be fitted to this part, and then wrapped around the bar and secured.

Mounting it very close to the hinge was practical as I found that it had immense pulling strength in operation. This allowed for a greater distance of movement on the figure, from the same small contraction of the muscle.
 
           
                                             
  Making the muscles move  
  The pneumatic system needed a compressor to supply air to the muscles. I found an airbrush compressor online with a 3 litre tank. The advantage of the air tank was that the compressor could store air in reserve, allowing for muscles to be repeatedly filled and emptied in longer bursts, or for multiple muscles to be driven at the same time. I was planning three moving figures, so this sounded like a good idea!  
  The compressor was capable of supplying air up to a pressure of 6 bar.
David informed me that the air muscles would work best at around 1.5 to 3 bar, so this would be more than capable. It also had a mode whereby it would charge the tank up to 4 bar pressure, then switch off. When the pressure dropped to below 3 bar, it would switch on again and recharge.

This would prove to be very useful - it would have been damaging to the compressor to leave it running permanently (it would overheat), but if it was only to come on for short periods to top up the tank, this would ensure a more or less constant supply of air at the required pressure for the figures all evening.

One other great benefit was that due to its relatively small size and power, it was very quiet when running. It would have been impractical to have had a massive compressor rattling away at 90 or 100dB, despite the substantial increase in air supply! Installed in a shed, behind a wall at the back of the garden, the airbrush model would be completely inaudible to visitors.
     
  Controlling the movements  
  With a supply of air now sorted, I turned to the challenge of controlling the muscles in the sequence required to animate the figures. Each figure would only have one muscle, moving a single section. The wolfman, as described above, would be hinged so that his whole upper half rocked back and forth. The good news was that I wasn't after sophisticated movements, just a bit of erratic shaking and lunging!

David explained that as the muscle could only pull in one direction, I should attach some strong elastic cord to the hinged part, so that once the muscle had pulled it one way and then been emptied of air, the elastic would restore the moving part back to where it started. In the case of the wolfman, repeatedly filling and emptying the attached muscle would cause his torso to move backwards and forwards.
 
    The switching of the air supply into each muscle would be done using 12V solenoid air valves, arranged as in the photo to the left. Quite simply, when 12V was applied across them, they opened and air could pass through. When they were switched off, they shut again.

Two valves were used per air muscle. The leftmost valve in the photo ('IN') was responsible for allowing air into the muscle. Air came into this via the black pipe visible in the bottom left of the photo. The other valve ('EXHAUST') released the air from the muscle. Restrictors (the silver rotary part on the leftmost valve) were useful to control how fast the air could flow into the valve and up to the muscle, which allowed me to adjust the speed of movement.

The photo also shows a regulator, which was used to limit the air pressure going to the muscle of the figure. The wolfman needed 3 bar to move; the other two figures used around 2 bar.  The pipe extending out of the regulator and off the right hand side of the photo was connected to the muscle. All the muscles, fittings and valves were connected using 6mm nylon pipe.
 
                                             
  Due to the way in which the valves were connected with the regulator and outlet to the air muscle (in a T-shape arrangement, like in the photo), it became clear that the 'EXHAUST' valve could only be opened when the 'IN' valve was closed - otherwise I would be draining air from the compressor and the muscle at the same time! So the procedure to move the muscle would be this:

1. Switch on 'IN' valve to open it. Air flows into muscle and the muscle contracts. (Red arrows in the photo above).

2. Switch off 'IN' valve to close it once the muscle has fully contracted. Air supply from the compressor to the muscle is now closed.

3. Switch on 'EXHAUST' valve to open it. Air drains from the muscle out via this valve. (Blue arrows). Muscle returns to its normal length.

4. Switch off 'EXHAUST' valve once the muscle has been fully drained of air.

 
 
  This process would need to be electronically controlled, to manage the switching of the valves needed to move multiple figures, as well as receive initial triggers from motion sensors. It would be necessary for each figure to have a pre-programmed sequence of commands that controlled its associated air valves to specific timings, in order to produce the movement I wanted.

I decided to use a series of Arduino Uno microcontrollers, connected to a 16-way 12V relay board. Each controller would be connected to its own group of relays, switching them on and off as programmed when activated by an external trigger.

I mounted this all in a large moulded electrical box (see photo right - excuse the cable mess, this was taken during building!) This system would be the 'brains' of the display. Also inside the box were some 12V relays that were capable of switching 230V, which was useful for controlling mains lights and a smoke machine used for a separate fire-breather effect (more about him later!).

   
    I wrote some code for each Arduino board, which would run each time the effects were triggered. David later came back with the suggestion that I could have used the board's 1k of EEPROM to record the moves of each figure 'as live'.

By connecting the controller to push switches which also ran to each valve, I could press these to animate the figure how I wanted whilst recording the switching into the board. However, I'd already finished writing the code so I stuck with that method, although puppeting the moves with each figure in front of me would definitely have been a lot more fun! In either case, there was also the possibility to create more than one animation sequence for a figure, and have the controller pick one at random upon triggering.

Each board was programmed to wait for around 10 seconds after its animation sequence had completed. My air muscle construction wasn't perfect, and they did leak a bit, so I was needing more air to move the relatively heavy figure frames - it was sensible to give the compressor a bit of bonus time to recharge before the effects could be re-triggered.
 
  I took a slight technological step backwards for 2015, in terms of sound effects! Whereas the previous displays were moving steadily away from CD-based soundtracks, this year I had a big stack of CD players at my disposal, so I decided to put the sounds for each figure on discs, and have each Arduino board trigger playback on the associated CD player. This was perhaps another idea inspired from Rumpus Mansion at Blackgang Chine, where its sounds also originate from CD players (1993 technology, remember!) and are triggered in a similar way by a controller for each scene.

The CD players worked well, and one of the players had a 'random' playback function, which I used for the fortune teller figure. On the disc were around ten different phrases; I programmed the Arduino board to start the random function, wait for three different phrases to be played, then stop the disc. This worked very nicely, and gave the figure more variation. The sound effects were the basis for the timings of the figure movements. I wanted it all to match up and be in sync, ie. the wolfman would lunge forward and growl at the same time!
   
                                             
  Open-CD player surgery...  
    < Indulge me on this little electronics musing...take a look inside these two CD players I used. The top machine (JVC XL-E31) is about 25 years old, the bottom one (Technics SL-PG390) about 15.
What a difference in the circuitry required to perform the same task!

> Here's the stack of CD players and amplifiers used for the display. Welcome to 1990...! Each player had pieces of wire cheekily inserted and soldered across their relevant playback buttons for external control by the Arduino Uno boards.

I was born in 1991, very much in the glorious era of the CD... So when better than 24 years later to celebrate that...! I like to think of this as my equivalent of a misty-eyed vinyl fan digging out their old record player!
   
                                             
  Constructing the figures  
  While I was building all this, I was aware that I was rather keeping David in the dark as to what my figures were looking like! I wanted him to see them fully working, rather than as just bits of pipe and hinges! And that's what I've done with this 2015 account...I've kept you in suspense about how they looked! So let's have a sneak peek at how they were made...
 
 
    < I'd come up with the final plan for three moving figures; the wolfman, the grotesque (another caged monster) and the fortune teller. The drawing to the left shows some of my early ideas, which includes a 'magician'. He was included in the display, but as a static figure (shown on the next page).

All three of my figures were built from pieces of 50mm pipe, connected with elbow fittings. This helped to make them lightweight, but the plastic was still surprisingly robust and coped well with movement from the air muscle. The hinged part (the spine of the wolfman, in the photo below, left) was attached with bolts using standard steel hinges. The air muscle was then fixed next to this hinge, and attached to the moving part (eg. the spine or head).

I used foam lagging to 'bulk up' the limbs of the figures, and also to create a 'rib cage' to shape the torsos. The heads were polystyrene, with latex masks attached. Each figure also had latex hands. It was great fun finding clothes for the figures to wear; I got a few odd looks in charity shops buying old jackets and shirts for them...!
 
    < The wolfman frame was the most challenging to build as he was required to be completely free-standing! His pipe legs were tightly fitted to the floor of his cage with more elbow fittings to prevent him from falling over!

> The grotesque and the fortune teller were easier to create as although they were to move in a similar way to the wolfman, they were only 'top halves' with no legs!

The fortune teller's lower half would be hidden behind her table, and the majority of the grotesque's body would be disguised under rags!
   
                                             
  The end of October was approaching, and I was putting the finishing touches to programming the figures. The wolfman hadn't been moving as well as I'd hoped. His hands were fixed to the bars of his cage, and it turned out that as a result he couldn't move forward as much - his arms were preventing it somewhat.

The solution was to cut his foam arms away from his body, so that only his jacket was connecting them to his torso! This gave him a lot more freedom of movement, and suddenly he was rocking and rattling his cage a treat!

Zelda was easier to build and operate - no heavy jacket or arms, just purple fabric, hands, and a lightweight polystyrene head with mask, so she required less air pressure to move.
   
  See the wolfman moving before he was modified - his movements were restricted by his hands and arms
(.mp4, 4.5mb)
  See Zelda working in this early test clip of the figure
(.mp4, 4.6mb)
 
                                             
  I decided to call the display 'The Carnival of Horrors'. As well as the moving figures, I had some other special effects and tricks up my sleeve to spook our visitors! It was finally time for the carnival to open... (and for you to finally see the figures in action!)...