Personal Projects



Duration - 1-2 months

For no real reason other than to see if I could, I wanted to build a robot spider. This project took a small amount of my time for a few months. 3 main iterations were designed and built. All three were quadropedal designs. The first used 8 servos, 2 for each leg. CAD:

Spider
Fully built:

Spider
The brain is an arduino nano with a shield which allows for easy wiring of the servos. Power is provided by a small 3s LiPo battery (from my RC planes). Unfortunately this first design suffered from lots of backlash in the legs. This caused it to be very flimsy and I was unable to make it walk. Could get it to do this though:

Spider
This led to version 2, which used Klann linkages and 2 continuous rotation servos:

Spider
The main issue with this design was motors running out of sync, causing it to flop to one side after a while. It also suffered backlash problems, due to the large amount of linkages and pivots.

The third and final (for now at least) version went back to the original design, using 8 servos to control the joints. The design was built from the ground up to be more rigid.

Spider
It can do better squats than me, although that isn't really saying much:

Spider
Most importantly though it can actually walk:

Spider
I have since added a HC-06 bluetooth module, which allows the robot to be controlled through an app on my phone. More importantly, I have created created an excellent new innovation for the biscuit industry:

Spider

- 23/09/18

Duration - 2-3 days

I originally made a basic potential flow simulator back in winter 2017, when I was supposed to be revising aerodynamics. While it was technically procrastination, it did help me understand what was going on. For the uninitiated, potential flow theory is a relatively simple way to visualise and model fluid flows. It assumes the flow is inviscid, steady, and incompressible. The simulation is written in python. It is structured such a flowspace object can be created, which can have various flow elements added to it. Streamlines, velocity potential lines, velocity and stagnation points can then be found and plotted. The main program and a few demonstrations can be found here.

Lifting flow over a cylinder:

Potential Flow
Rankine half body:

Potential Flow
Rankine oval (velocity potential lines):

Potential Flow
Vortex and a sink:

Potential Flow

- 15/08/18

Duration - 2-3 weeks

Although the plane is indeed unmanned, the title isn't a typo, I just haven't named it yet. After my less than successful experience flying the professionally designed simple scout, I decided the best course of action would be to try and design my own plane. After some quick hand sketches, I drew up the aircraft in Solidworks:

Unnamed Plane
The main issue with the simple scout was the fixed wing position. This was a problem as the GC ended up too far back, making the plane very hard to fly. The new plane solves this by having a wing which can move along the fuselage. The wing is attached to the fuselage with some rubber bands, for easy adjustments. Some changes were made from the solidworks model, for example the ailerons, and the circular wingtips.

Unnamed Plane
All the electronics were salvaged from the simple scout. From designing the aircraft myself, as well as consulting a friend, I have made a few mistakes which I won't make again:

The plane uses a few 3D printed components: Unnamed Plane


Unnamed Plane

Most importantly, it actually worked!
After turning the sensitivity down to about 70%, it's much more controllable than my previous attempt, however is still a bit twitchy on the roll axis. I only crashed it once, but that's because of my poor piloting skills rather than poor design. Proof that it works at all:

In hindsight I should have waited until a less windy day, but it still survived.
And yes, the Wednesday frog on the front is necessary for ballast purposes..

Unnamed Plane I should note that after taking the plane out a few times and getting the hang of it, it flew quite well. Up until the point it nosedived too hard, damaging the motor, and burning out the ESC.

- 04/08/18

Duration - approximately a weekend

A seemingly simple hedgehog toy with flexible links. This was a very short project, where most of the time was spent tweaking the tolerances in the links.
The STL files can be found here

Hedgehog Toy Hedgehog Toy

- 23/07/18

Duration - 2-3 weeks

This was my first attempt at building an RC plane from scratch. It was designed by FliteTest, and the plans can be found here.

Simple Scout
The plane did techincally fly..

Simple Scout
Unfortunately, the plane suffered from an aft CG, making it very difficult to control, especially for a beginner. I also discovered afterwards that new pilots should always limit control surface movements to around 50%-80% of their highest values. After a few semi successful flights, it flew for its last time.

Simple Scout
Fortunately, all the electronics survived the crash, and could be repurposed for my next attempt. The fuselage was easy to tape up, however it will never fly again. It now hangs on my wall as a reminder of things that could have been.
- 13/07/18

Duration - Ongoing since March 2018

In late February 2018, I decided I wanted a 3D printer, so I bought a cheap one from China, the Tronxy P802m. Including a 1kg spool of PLA filament, the total cost was about £150. Not bad if you ask me. After lots of assembly, and some required modifications, it was put together:

3D printer
One of the main issues to start, was the hole for the Y axis belt holder was in the wrong place. This meant some of the parts had to be at very awkward angles. Luckily, it printed just well enough for me to design and print a replacement part:

3D printer
Another problem to start with was completely my own fault. The belts were far too loose so the prints were terrible. After proper tightening, prints started coming out decent:

3D printer Cali Cat
After much more tweaks of settings, and proper bed levelling, the printer now performs well with an acceptable level of reliability.

3D printer
Flexi Rex Wednesday Frog
- 10/06/18

Duration - 2 Weeks

After winter exams in 2017/18, I wanted to make something, and decided on a simple arduino powered clock. To start, I acquired a DS1307 real time clock (RTC) module. This is what actually keeps the time. Following that, I made a prototype with the resources I had (arduino kit provided by the university):

Clock
Once I knew the concept could work, I figured out how to use the RTC module in the arduino code. I also purchased a large box of assorted LEDs and a few arduino nanos.
I then drew up a box in Solidworks, and laser cut the parts in the University workshop. After assembly, the internals looked like this (messy I know):

Clock
The final product looks like this:
Clock Clock

To tell the time, the top row reads the hours, middle row minutes, and bottom row seconds, in binary. The 2 buttons on the back are for adjusting the time. The clock is powered by a mini USB cable, plugged into the wall via a 5V adapter.

If I was to do this project again, I would take more consideration with cable management. Designing and ordering a custom PCB would have made it significantly tidier, and less likely for loose connections to occur.

- 10/02/18