Thursday, October 25, 2012

Sunday, October 14, 2012

Pneumatic Jack in the Box, Part 3

To support the mask I built a frame out of perforated metal tape. I basically fit each piece into the mask from the inside, then attached the pieces together with small nuts and bolts. The frame was then screwed into a stand made of three pieces of 1x2 wood left over from the torsion box coffin lid project.

Thursday, September 20, 2012

Pneumatic Jack in the Box, Part 2

Make sure you view all the videos posted in Part 1 before continuing.

The ELK-120 sound board can pull up to 2A@12VDC and the mouth servo may require more current @5VDC than the Arduino's voltage regulator can support, so I went with a dual-output power supply that provides 12VDC and 5VDC outputs.

However, the Arduino and the DF Robot Relay boards all need 9VDC.

Tuesday, July 17, 2012

Pneumatic Jack in the Box, Part 1

This year's new prop is a large Jack in the Box - a two foot cube, powered by the Arduino microcontroller. It will have an Insteon InlineLinc relay module inside so that the prop can be controlled with a hand-held remote or an Insteon motion sensor. When the Arduino sees that the InlineLinc has been activated, it will execute the following sequence of actions:
  1. Turn on a spotlight aimed at the prop so people pay attention to it
  2. Spin crank handle (powered by motor) while playing a sound effect
  3. Open the top lid
  4. Pop up a large scary clown head
  5. Play a sound effect while moving the clown's lips with a servo
  6. Lower the head
  7. Close the lid
  8. Turn off the light

Wednesday, June 13, 2012

Costume-friendly remote for prop control

I've moved away from using motion sensors to trigger props for several reasons: Some props are too scary for really young trick-or-treaters, motion sensors don't know if the person is looking at the prop or which direction they are walking, and manual operation results in better timing.

However, the buttons on most remote controls are too small for a costumed operator. My werewolf claws can't tell if they are on the button.

Costume Remote Fail

Additionally, the battery life and effective range of the remotes I've tried is less than ideal.

So I set out to design a remote control that is more costume-friendly.

My requirements:
  • Large buttons
  • Control multiple devices
  • Long battery life
First I needed to select a control system. I decided to go with Smarthome's INSTEON technology. A remote can control multiple devices, and "dual band" modules (such as the InlineLinc 2475SDB) receive their control signals wirelessly. The RemoteLinc2 can control 4-8 devices and is the size of a small stack of business cards.

Next, the physical interface. I decided on a black PVC tube, with a large selector knob on one end and a large activation button on the other end. The selector knob will set the operating mode (OFF/1/2/3/4/PROGRAM) and the large button does whatever the current operating mode is set to.

However, I need a way to determine what the current operating mode is, in the dark, without rotating the knob back to OFF and counting clicks forward. I decided to include a 7-segment LED behind a red lens to show the current device number (or a period when in PROGRAM mode)

In order for the selector switch to control power, device selection, program mode and 7-segment LED digits, it needs to be a "four pole" rotary switch. Although the RemoteLinc2 supports 8 devices, I decided 4 was enough. 4 plus OFF and PROGRAM means the selector switch needs 6 positions, so it needs to be a "four pole six throw" (4P6T) rotary switch.

Newark sells one made by Electroswitch that is quite well made. Part number is C4D0406S-A. It was a special order. This is the "shorting" type rather than the "non-shorting" type. Shorting, or "make before break" is requried since one of the four poles carries power to the InlineLinc and it needs to stay ON between mode changes 2-6 (1-4 plus PROGRAM).

I came up with a simple way for the selector switch to control the 7-segment LED so it shows 1-4 based on the switch position - using a network of diodes. The diodes prevent current from flowing in the opposite direction and turning on all the connected segments. However, since current flows from negative to positive the 7-segment LED needs to be of the Common Anode (anode=positive) variety.

The RemoteLinc2 requires a regulated 3.9VDC power supply to operate correctly. I wanted to be able to run the remote off of either 9-volt (used on my UV LEDs or blinking eyes) or AA (five times the life) batteries and the voltage regulator actually made this possible. I can connect a 4xAA pack or a single 9V battery to the battery snap connector that goes to the voltage regulator and I will always get 3.9VDC. The voltage regulator circuit and the diode network went on a couple of round PCBs purchased from the neighborhood Radio Shack.

PVC Remote Diode Array PVC Remote Diode Network section PVC Remote Voltage Regulator section

The round PCBs are glued to three nylon rods spaced equidistant from eachother and the whole assembly slides into the PVC housing.

It's fairly easy to hold while wearing a costume, and one hand can slip to either the selector knob or the activation button while the other hand holds onto the body.

PVC Remote Fits

Sunday, May 27, 2012

Pneumatic Toe Pincher Coffin, Part 4

I decided to revisit this project and make some improvements. The solid plywood coffin lid tends to "flap" when opened, and the pneumatics are not smooth. It takes a while for the cylinder to overcome the weight of the closed lid, and if I open up the flow control too much the lid opens with too much force.

This year I am replacing the lid with a "torsion box" style lid, which will include a nice routed edge profile. Torsion box construction will result in a very rigid lid, eliminating "flap" - and I am also making some changes to the pneumatics.

The problem is that since the coffin is laying on the ground horizontally, the weight of the lid decreases as it is opened. It takes more air pressure to start the lid opening than it takes to complete the arc to the fully opened position. The cylinder needs an initial burst of air to overcome the weight of the lid, then a slow flow of air to open the lid slowly.

Bimba makes accumulators (air tanks) in a variety of sizes. It only takes a small volume cylinder to add enough burst pressure to overcome the lid. For my project I selected the Bimba D-1500-A-3. I also added another flow control. See the diagram below.

The air compressor connects to the coffin pneumatics via standard 1/4" industrial quick-disconnect fittings. The first component is the air filter, to prevent debris from entering any other components. Next is an optional large accumulator. If large enough, this tank can be "charged" with enough air to run the coffin for an entire evening, but at a minimum it serves to buffer any air line pressure drops due to other props consuming air from the same compressor. The check valve keeps this tank pressurized even after the air line is disconnected from the coffin.

Next is flow control #1. This flow control determines the speed at which the coffin lid will open AFTER the initial weight of the lid has been overcome by the small accumulator and flow control #2. Flow control #2 should be open a bit more than flow control #1, but adjusted so that overall lid movement is smooth after the coffin has been placed into the environment where it will operate. The angle of the coffin where it rests on the ground will affect the weight of the lid through its opening arc.

Flow control #2 connects to the 3/2 valve which connects to the cylinder. The other side of the valve connects to flow control #3 which controls the rate the coffin lid closes when the valve is turned off. This connects to a small exhaust muffler to reduce air hissing noise.

Another major change is the shift towards push-to-connect hose and fittings. Such much easier to work with than coiled rigid hose and 1/8" threaded connectors. Automation Direct sells a broad selection of push-to-connect fittings, tubing, and flow controls.

The new lid is made out of two sheets of 1/8" doorskin with 3/4" pine beams that are approximately 5/16" wide. Grid spacing is 4.5" and the "bottom" lid is 1/8" wider on each edge than the coffin. The "top" is 1/8" narrower. Edges are 1.5" x 3/4" pine, and are routed with a Whiteside 3172 "Wavy Edge" bit.

The hinge side of the lid and the bottom are made of 5.5" wide pine rather than the 1.5" so that there is support for screwing the hinges and cylinder clevis to the lid.
Coffin lid torsion box internals       Coffin lid routed and base coat  Whiteside 3172 Wavy Edge

Return to: Pneumatic Toe Pincher Coffin, Part 1

Friday, May 25, 2012

Momentary pushbutton debounced into toggle flip-flop

Here's how to debounce a momentary switch to turn an AC-powered device on and off, and provide power status indication at each pushbutton switch.

Wednesday, March 7, 2012

Polycom Headset Adapter

Thought I could simply plug in my wonderful GN Netcom 4800 headset into my Polycom Soundpoint Pro home office phone but no, a special adapter is required to adapt the RJ9 connection from the Netcom to the 3/32" subminiature phone jack on the Polycom.

I found an adapter on the internet with Polycom part number 2457-11095-002 and ordered it, only to find that its bulky shape didn't quite work with the location of the jack on the Polycom phone.

Disassembly was required in order to make re-orientation possible.

In the process, I documented how the adapter is wired. There did not appear to be any resistors, capacitors or diodes inside the adapter housing.

Polycom Pinout