Halloween Experience - North Baddesley's haunted walkthrough!



	> 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

A name to conjure with...

I decided that the walkthrough would be called the ‘Tunnel of Hell’, owing to the series of ghoulish scenes I had developed showing various figures in tortuous situations, and demon-like monsters that would suddenly move towards visitors. I visualised a big imposing entrance, like an archway, with the name as a large sign directly above where people walked in.

A lot of the inspiration for this came from reading parts of Dante’s Inferno, one of the three parts of his 14th century work The Divine Comedy. Dante conceives Hell as an enormous pit made up of nine circles, containing trapped souls guilty of sins such as gluttony, fraud, treachery…each one darker than the last. He journeys down through each circle and witnesses horrific torments inflicted on these souls.

All this was very powerful and scary imagery, and it gave me a lot of ideas. Some of them made it into the final tunnel design.

Image credit: Daniel Brokstad
danielbrokstad.com

See my first layout plans and descriptions of effects
with this set of early concept ideas (pdf, 2.0mb)

Triggering the scares...

I wanted the tunnel to have the feel of a ghost train at a funfair, in the sense that as you went through it, things happened to YOU! Visitors wandering through the dimly lit passageways would encounter a number of shock effects, triggered as they got near to each window. I wanted it to make people jump!

But how to trigger these effects? Since 2006, I’d used PIR sensors to detect motion in the garden, and these would then trigger whatever effect was on the end of them. PIRs could pick up motion in a wide area, which was particularly useful when detecting people first entering the garden. But in the tunnel, there were clearly defined points where I wanted the effects to happen. The answer was to use infra-red barriers.

< Set into the walls at the required points in the tunnel were four of these devices – small cylinders about 20mm in diameter. These emitted an invisible infra-red beam.

> On the wall directly opposite, a small reflective plate was fixed. Under normal conditions, the infra-red light would be reflected back at the sensor. As soon as someone walked through the beam, the light would stop being reflected and the sensor would trigger. This worked very reliably.

This year, the system in charge of controlling everything was set up in a shed situated over a wall behind the tunnel. In here was the airbrush compressor (from 2015) for connection to the pneumatics, shelves of amplifiers for all the sounds (the whole setup used ten speakers!) and the main control box.

Inside this box were four Arduino Uno microcontroller boards, each controlling a scene. There were also two Adafruit Audio FX boards, containing sounds for two of the scenes. Each of the Arduino boards controlled relays on a 16-way relay board. This was where everything to be switched (12V air valves, mains relays for lights, trigger pins for sounds etc.) was assigned a relay.

When I was developing this system, I gave each of the six scenes a letter, so they were simply known as A, B, C, D, E and F, with A being the first scene and F the final scene. The C and D scenes would be two figures opposite each other that were part of the same animation sequence, so they were known as 'C/D' and shared the same controller. Scene E did not require a controller as it was to be animated using motors that would run continuously, instead of being a momentary, triggered effect.

The order of the scenes going through the tunnel was to be this:

SCENE A (IR triggered)
- swarm of wasps stinging man
(motorised figure with projection and sound)

SCENE B (IR triggered)
- wailing prisoner in chains
(pneumatic figure with sound and light)

SCENE C/D (IR triggered)
– trolls wielding rocks
(two pneumatic figures with sound and light)

SCENE E
- floating souls and reaper
(non-triggering scene with sound)

SCENE F (IR triggered)
– devil with fire
(pneumatic figure with sound)

There was also to be another separate effect in the porch, just outside the tunnel exit (known as M, the mirror scene). As it was to feature another pneumatic figure, it also needed an Arduino controller. To hold this, a separate control box with a 4-way relay board was installed inside the house, which connected to the main box outside.

There were also three computers, each running DMX timelines on VenueMagic SC+. Two of these were used to control lighting and sound for scenes (C/D and M). The C/D computer was triggered from the outside control box by the relevant Arduino board. The third computer ran a looped sequence to control the dimmers which powered all the flickering PAR 16s inside the tunnel, as well as the two torches either side of the entrance, and other LED PAR 56s used for scene lighting. It also ran the main soundtrack music of the tunnel on a seamless loop.

Adding the pneumatics

The moving figures in scenes B, C/D, F and M were fitted with 100mm-stroke pneumatic cylinders, replacing the air muscles I used in 2015. These cylinders connected via a network of 6mm air pipe to a box located in the closed-off space in the centre of the tunnel. This box contained a number of 12V 4-way air valves that were responsible for allowing air from the compressor into and out of the cylinders. These were switched on and off by the relevant Arduino Uno boards and relays in the control box in the shed.

Surely I should have used 4-way valves?

Good question! There’s a bit of a story here…The cylinders were of the double-acting type, which means that air can be forced in from either end depending on which direction the cylinder needs to move (ie. out-stroke or in-stroke) – the exact job of a 4-way valve.

Originally, I did connect the cylinders to this type of valve, as in the diagram to the left. Their configuration allowed air into the cylinders such that they remained fixed at one extreme (eg. piston rod fully in-stroked). Then, when the valve was turned on, it ‘flip-flopped’ over so that the air now flowed in through the other end of the cylinder, and it moved to its opposite state (piston rod fully out-stroked). When the valve was de-powered, the direction was reversed again and the rod returned inwards.

See me doing an early test of a figure frame with a 4-way valve and pneumatic cylinder
(.mp4, 4.0mb)

This all seemed to work during testing...but when it came to fitting the whole system of five valves together during the installation week, I just couldn’t get them air-tight enough for the compressor to maintain the required air pressure! I definitely had the correct type of fittings in the valves, they were of the right thread and they were super tight, but collectively they still leaked enough that the compressor could never reach its 4-bar target, and as a result, would never switch off! It would have overheated if it ran continuously, so some urgent re-thinking was required.

The solution (from bits I had available) was to use two standard 12V 1-way air valves per cylinder, and operate them as single-acting, similar to how the air muscles in the 2015 display worked. I called the valves for each cylinder IN and EXHAUST. They were connected with a T-piece, as shown in the picture to the right. (click it to enlarge)

The idea was that each cylinder would have a resting state (piston rod fully in-stroked), and the moving part of each figure was attached to a fixed point with strong elastic cord. When the first valve (IN) was turned on, with the EXHAUST valve closed, air would enter the cylinder and the piston rod would out-stroke. Then the IN valve was closed again, and when it was time for the figure to move back, the EXHAUST valve was opened. The air would escape from the cylinder, and the elastic cord would return the piston rod and figure back to its starting position.

This may not have been the system I expected to create (!) but it did have the advantage of needing less air pipe overall! Now only one piece of pipe needed to run between the valves and the cylinder, rather than two. I was also able to finally achieve a very air-tight system, capable of supporting the necessary air pressure. (The figures all operated using around 3-bar pressure).

There were five pneumatic effects in the display (one each in scenes B, C, D, F and M). Eight valves were installed in a box (for scenes B, C, D and F) and the two valves for the M scene in the porch were positioned directly behind the wall for this scene. These two valves were controlled by the Arduino board and relays inside the house.

The eight valves in the central box were surrounded with acoustic foam in an attempt to dampen their noise a bit – the switching and releasing of air was quite loud! But once they were in place behind the tunnel walls, you could hardly hear them and they did a great job throughout the whole week leading up to the big night with never a problem.

As I had a total of five pneumatic figures this year, I programmed each scene controller to wait for a period of about 10 seconds after each figure had done its thing, before being able to be triggered again. There was the very likely possibility of multiple figures being triggered at the same time, and so a higher air demand would be put on the compressor. With this delay in place, the compressor would have a bit more of a chance to restore the air pressure level before more triggering took place. As another precaution, I designed all the figures in such a way that even if they didn’t move at all due to low air pressure, their resting positions were such that even with just their accompanying lights and sound active, they still looked effective. As it turned out, I needn't have worried - the compressor did a fine job, firing on all five cylinders...!

Building the figures...and recycling old ones!

The plastic pipe frames of the 2015 figures were modified to create the new ones. The wolfman became the devil (scene F, right), as he was already floor-standing, and just needed his legs straightening! The fortune teller frame was given legs and became the man being stung by wasps (scene A), and the grotesque frame was used for the mirror scene (M).

Frames for the static figures of 2015 were modified to have hinged pieces that attached to the cylinders. These were used for the C/D troll figures (left), and a new frame was made for the scene B figure.

The figures were built up using polystyrene heads with latex masks attached (including one from an Irish company called Rubber Johnnies... hmm). The arms and legs were ‘bulked up’ using foam lagging, then covered in different materials, such as hessian, gauze and muslin.

Here are some pictures of the tunnel interior before the scenes were added in.

Apart from my little urgent pneumatics reworking, the rest of the tunnel had gone together a treat! By Sunday 30th October, the only things left to install was some extra lighting in the porch, and the signs to go outside the entrance of the tunnel. The whole interior of the tunnel was fully working and tested, and so it meant that for the first time ever, we could all have a little preview of the show ourselves that evening, a day early!

The next day came, the finishing touches were completed, and in the evening, it was time for the tunnel to open!...