Wednesday, September 10, 2014

The Making of Gertie - Part 3: Engineering the Jump

The next challenge was to create the right springs for Gertie's hopping mechanism. They would fit in between the leg and the servo. The springs would compress as Gertie landed from his first jump, storing some of the energy as potential energy, that could be released on his next jump. This would improve the height of Gertie's jumps, while taking some of the stress off the servo motors as well.

In the end, I was unable to find the right springs for my needs: ones that had the right strength/resistance and diameter. I decided to make my own springs, so that I could make them to my exact specifications, using the music wire below:

To create the springs, I created a wooden jig, after some trial and error--and eventually got some great results! To research spring-making, I referred to this immensely helpful website: How to Make Springs. The spring-making jig I created was different from the designs on that site, but I was able to leverage many of the concepts they described. I'll go into more detail on my particular jig, and how it works, in another post!

Here you can see the springs attached to Gertie's legs, and assembled as a bare skeleton! This was one of the first working iterations, and it was very exciting to finally see all the pieces together.

Here is a video of Gertie jumping for the first time, and the delta kinematics at work:

Monday, September 1, 2014

The Making of Gertie - Part 2: Early Designs and 3D printing

For some background on Gertie, check out Part 1: Starting Out!

After figuring out what I wanted in the robot's design and function, I decided to start with my first 3D printing experiment. Using Maya, a 3D graphics program, I created a digital mockup of Gertie's motor mount, which would hold the 3 servos (one for each of the delta robot's arms). The print turned out great, and was an eye-opening experience! My digital modeling skills from working in computer animation translated so well to this process. Realizing it made me feel like I'd just evolved into a more powerful Pokemon! You can see below, how closely the print matched the digital model:

Each pair of arms wrapped around one servo. There was a satisfying "click" as each motor fit precisely into its frame! The accuracy of printing, even at a moderate resolution, was very exciting.

I continued to refine the design in Maya, creating a complete digital model of Gertie. Even the parts that weren't 3D printed were modeled, so I could judge the full design and whether the parts fit well.

This model below is one of the earliest versions: you can see the first part I tried printing (it's upside down from the images above). Each of the three legs hooks into the servo at the top, which could rotate quickly and cause the leg to contract and expand to cause Gertie to jump.

This motor mount design wasn't structurally sound in the end because the large motors were very heavy and strong and kept popping out of socket. The motors needed to be secured by being bolted on instead of the previous snap-in method. 

While designing the new motor mount, researching for new motors led me to the consideration of a better weight to power ratio. There were many different servos to choose from, and the most powerful ones were obviously attractive for higher jumps. They were also bigger and heavier, a trade-off that would shorten the jump too. I ended up doing a good deal of research, and created a spreadsheet of the different servos, ranked by power to weight ratio (  Ultimately I chose a smaller motor with better weight to power ratio, the TGY-306G-HV

I created and printed another version, with a new frame design and the much smaller servos:

Getting closer now! In the next post, I'll talk about creating springs for Gertie's legs, and getting to see him finally move.

The Making of Gertie - Part 1: Locomotion concepts

In existing toys, there were a couple of examples I looked at. One was the flipping dog toy, which used a rack and pinion to compress a spring. The pinion (the gear) only has teeth in a specific part of the wheel so that when it rotates to the part that has no teeth, the built up spring tension is released and the rack impacts the ground and makes the dog jump!

One of the limitations of this design is that you can only jump at one height, so I looked into a scissoring mechanism that would give me this level of control:

This design led me to the delta robot, which consists of three main arms connected at the tip. The three arms allow the tip to move all directions in space. It's not limited to just an up and down motion like the scissoring mechanism above. This mechanism is typically used in 3D printers and "pick-n-place" machines:

My first experiments with 3D printing come in the next post! I start to test different designs, and need to get a bit creative with making my own parts.