FabLab Academy 2012

Manchester Lab

David Forgham-Bailey

Week 20: Final Project Presentation update

This section is designed to update and complete the documentation for my final project and reflect on the course...

When we were directed to plan our final project, in the first week, I decided to produce an old style stage ventriloquist dummy. My original intention was to make a manually controlled head unit and seperate body. Throughout the course, as new units were taught, I modified and updated the design. It became clear that with the new possibilities introduced by the electronics and programming sections that my final design must include electronic control.

A major restriction in design was size. I experimented with various methods of trying to squeeze motors, pushrods etc into such a small space, and still retain functionality. I used Rhino to create the block head out of slices - then removed sections for the eye unit and mouth unit. The mouth would be created using the slice method, but the eye unit was far more complex.

The design of the eye unit was perfectly suited for 3D printing. I experimented by printing small prototype eyes to determine the minimum gap size - 2mm was too small - 5mm gap produced a good fit and smooth rotation. It took two days to dissolve the support material - so desiging a less bulky eye was required.

To prove the control system, and assist in the design process, I manufactured a series of test rigs from easily obtained items (25mm wooden beads for eyes - meccano for the frame - springs, rods and screws from a local craft shop).

The first system used servo motors for eye and mouth movements. The control was programmed using an Arduino Mega. A nintendo Wii nunchuck acted as the hand control. It was intended to replace all this with a modified hello servo board, and the hand control would use the most appropriate input method (hello button/hello load...) But for now this will remain on the list for future improvements.

Although the movement produced by servos was good, the noise they produced was not.

The next rig used stepper motors - these were considerably quieter, the control they offered was good but their size, and weight offset these advantages.

Each test unit made could also be operated manually.

The task now was to create a final design which would use as many FabLab operations as possible.

In order to complete this project in time I've had to make some adjustments to the specifications.
A combination of machine availability, design time required and a flood in the FabLab means that a more basic ventriloquist dummy will be made. The full original specification will be produced as time and machine availabilty permits, in the future.

The design elements of this project posed quite a challenge. I had decided early on in the course to use Rhino for 3D design. I had not previously used this software, and for the duration of the course, a free beta version of the software was available for the mac. However this was a stripped down version with most of the plugins unavailable.

Deciding on the best method to use for the manufacture of the head was tricky. Traditionally the head is carved from a block of wood or cast in sections from a mould. I considered milling a mould using the modella or the shopbot, but this was too complex and time consuming - maybe for the future...

I also investigated using Autodesk's '123D Make'. This software had just been released and could slice a 3D model into sections. I conducted trials with cardboard and plyboard, using the Epilogue Laser cutter, but the structures were too flimsy.

I eventually decided to construct a similar model using Rhino. This would be a basic block head shape which could be cut using the shopbot. This gave a result which was more like the carved block method.

The body and arms were fairly straight forward. A frame for the hips and shoulders was cut using the Shopbot and the structure was finished using dowels and heavy duty muslin. This material was also used for the limbs. The material was cut to size using the Laser cutter and stitched using the Sewing Machine.

Electronic control was required early on to prove the eye and mouth mechanisms. Here the Arduino made life easier. Routines were written to control both servo motors and stepper motors. Control was provided by a Nintendo Wii nunchuck. This required some investigation and experimentation to implement, but proved to be a slick way to interface. If time had permitted this contol would have been manufactured using modified hello.servo and hello.input boards. 

To overcome the problem of giving the dummy it'a own individual character, I decided to lasercut and etch a selection of faces which would fit onto the dummy head. Cut-outs for the eyes and mouth would reveal the working eyes and enable the mouth to move. As an interim measure I used cardboard printouts.

The Fab Academy has been an amazing experience. I have found the range of material covered and pace of the course quite a challenge. The list of subjects for me to follow up and study in more depth has continually grown. I have particularly found the electronics and programming modules to be valuable. I now have a long list of potential projects to develop.

Video of completed dummy in operation here.