Oscar – Getting ready for user testing

In order to commence user testing, we needed to make Oscar more durable and sturdy. This would also aid in more reliable and controllable results as our drawing aesthetic is reliant on precision and control. Some of the initial prototyping with Open Beam, which we posted earlier, showed that the “2 car” approach to the X axis was not sturdy enough to: (1) create a levelled and even movement along the Y axis, but also (2) it was too weak to support the movement of the Z axis cart and pen.

Below is the “2 car” approach. We treated each of the “cars” separately and relied on the 2 servo motors to evenly pull them along. The rails going across the x axis were not sturdy and there was flex along the axis.
2 car approach

The prototyped pen cart was created to have the pen in the centre, but what we found was that the central position would actually lift the cart off the rails if the pen was not properly positioned. This caused the cart to be dragged by the belt in a way that was not controlled or consistent. We also experienced that having the pen point straight down made it subject to pressure and traction on all sides. For example, if the cart moved to the left, the pen would be caused to press slightly to the right and the connection became looser and looser.

Below is the initial improvised pen cart. We very quickly learnt that the cart needed to have more weight so that it was not pushed off the rails by the pen-paper contact. We also found that the rails needed to be rigid and sturdy as the movement back and forth of the cart resulted in them moving and the structure loosening.
pen cart

The final major learning was to do with the belt along the X axis. While the Y axis belts were functioning well and seemed to be the right level of tightness, the Y axis was too loose and the meant that the pulleys could not control the movement accurately. We had additional two things which directly informed what we needed to do in future iterations of the belt/X axis design. We had a set width of the paper at 47cm, and we had a maximum belt length of 30cm. This meant that we had to be creative and project the pulleys over the paper rather than hide them in the structure as we had initially wanted to. However, projecting the pulleys over the paper without being able to have a strong support on the ground immediately below (due to the paper being wider than the belt, meant that we needed to create some customised parts or drill new holes in the Open Beam materials we had.
belt is loose

With the above limitations, we performe an initial test of Oscar drawing an image. This video was the drawing result of this initial prototype.
Video evidence of Prototype 1 work

The art work generated during the prototype 1 test session are below. Note that the eye was not drawn by the machine, but by us.
Oscar artwork 1

We took the learnings from the initial prototype and sought to make some improvements so that we could commence user testing.

Improvement 1: X Axis rigidity.
To make the x axis sturdier, we changes the “2 car” approach to a “1 car” approach. What we did was created the entire axis 1 car with 1 wheel on the extreme left, and the other on the extreme right. This created better rigidity and more stability for the pen cart.
Rigid 1 cart

Improvement 2: Customised and projecting pulley structure on right side of X axis.
While we built this structure, we very quickly realised that there we had a difficulty in fixing the pulley. It was unstable because the bolts kept unwinding with the movement of the belt. We will need to re-think how we build the right side of the X Axis. However for the 2nd iteration, we improvised the solution by placing the bulldog clip on top of the nut.
bulldog clip

Improvement 3: Springs and movement in Z Axis. Smaller cart design. Cart weight.
We recognised the need for the Z Axis to be less rigid and more responsive to the varying levels of the drawing surface. As such we added some springs which would absorb minor variations in the drawing surface height. We also added a servo which would control the placement of the pen on the page. This is functionality our previous iteration didn’t have.
pen cart
To add more weight into the cart we experimented with the addition of larger nuts and bolts as well as extensive use of washers.

Improvement 4: Position of the pen.
We changed the position and angle of the pen so that it resembles more closely to the way a human holds and lowers the pen. The new angle of the pen seemed to lend itself more to the drawing of continuous lines and resulted in less vibrations.
pen position

While testing the 2nd iteration, we found that the X Axis still had a bit of flex. We played with the position of the wheels as a way to mitigate the impact of the flex.
wheel on left
wheel position change

This is the video evidence of the drawing process by the 2nd iteration of the Oscar prototype.
Video evidence of Prototype 2 drawing

The input image is below.
Screen shot of user test image

The interpretation of this image by the 2nd iteration of the Oscar prototype is below:
Oscar artwork 2

The X axis also seemed to be moving more smoothly.
The pulley on the right side of the X axis needs to be sturdier. We will create additional support structure for it.
Drawing process is very slow.
The axis were opposite to what we expected.
We need to use of softer pen which bleeds out more is needed in order to speed up the drawing process.


One response to “Oscar – Getting ready for user testing

  1. Pingback: Oscar – Final Documentation | IDEA Studio 2 | Creative Robotics·

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s