4 month Internship - Thalmic Labs
From September 2014 to December 2014, I worked as a mechanical/ manufacturing intern at Thalmic Labs, a startup working on wearable technology. At the time, Thalmic had about 75 employees and was in the process of beginning to manufacture their first product, the Myo armband pictured to the left. The Myo is an armband that reads the EMG levels in your arm muscles to determine the type of hand gestures that you make. It wirelessly connects to devices over bluetooth and can be used as an input remote for things like presentations, music or anything else that has an app made for it.
​
When I started there, production had not yet begun and it began to ramp up to ~400 units a day by the time I left. My job involved developing equipment to improve the processes used in manufacturing. The details of two of the main projects that I worked on can be found in the sections below.
​
For background, pictured above, the myo armband is comprised of 8 individual plastic pods which are each made of a front and back that snap together with plastic clips. The band is a flexible plastic/ rubber that interfaces with the pods as they are snapped closed, holding everything together and is joined between one of the 8 pods. One of the top/bottom bands is embedded with a flexible PCB that connects the pods together and is over moulded in the plastic/ rubber.
Pod press for Myo assembly
​
My first project while at Thalmic was to solve an urgent assembly problem that begun just before I arrived. Waterproofing was initially achieved through a conformal coating over all ridged PCBs. While effective in preventing shorts, if water penetrated the mechanical pod and a drop sat correctly over the capacitive coupling on the flexible PCB, unwanted noise could be introduced. The solution was to use a clear silicon based adhesive to waterproof the pod enclosures themselves.
​
The main problem with the new form of waterproofing was that as the adhesive cured, if the pod was not pressed together hard enough, they had the tendency to pop apart. Applying a significant amount of force to close the pods seemed to seat them well enough that this did not happen, but it was unrealistic to do this by hand. Initial product runs were producing only a 60% yield with almost all problems related to insufficient or uneven closure force.
​
To solve this, I designed a pneumatic press that the Myo slides into and within 4 seconds, is adequately pressed without risk of damaging to the pods. The press is small enough to sit on the work benches of those assembling and is safe to use without additional protection due to its guards, lockouts and electro-magnetic latch while in operation. Each pod receives an independently calibrated amount of force via spring loaded pads. If the components are misaligned on setup, the sprung pads will attempt to align them and will prevent damage to the components, it can then be reworked if necessary. To queue up the Myo and its components to be pressed, I designed a fitted plastic "carrier" that properly held the loosely connected components and made the assembly process easier when inserting connectors, components and silicon glue. Lastly, as the armband is still flat at this point and only 7 of the 8 pods assembled, the Myo is rolled and the 8th pod loosely assembled. A hand operated press, working on similar sprung pad principles is then used to press the last remaining pod, this same press can also be used for rework.
In the end, the addition of these two pieces of equipment and the processes around them prevented the pods from opening almost entirely bringing the over all yield of the product up to ~98%.
​
Jig for flexible PCB insertion prior to overmoulding
​
My second major project was one that I took over from the mechanical team lead. He was working on a jig to improve the processes used in the injection moulding plant when inserting the flexible PCB that links all the pods together. As mentioned above, all of the pods are linked by a flexible PCB that is embedded in the flexible plastic/ rubber band. The first injection moulding shot of this part leaves a channel for the PCB which is stuck to the surface with double sided tape prior to the second shot that over-moulds the PCB. This is a challenging process as the channel follows the "zig-zag" contour of the part and along the way, each of the 8 pods must be lined up within ~0.5mm of their openings with a 90 degree crease that has to be formed in the part which is over-moulded too.
​
When I took over, the process by which the jig operated had been more or less established and it was my responsibility to finalize the design details, get the jig manufactured/assembled and make it work.
​
Finalizing the design primary involved taking the geometries of the parts and assemblies, and turning them into a jig that no longer had interferences, could be easily operated by a human, and was manufacturable. This ended up being a fairly complicated process given the complexity and number of moving parts/ interfaces across the design, many of the original parts had to be broken up into assemblies and adjusted to make it all work.
​
Once the design was finished, I made drawings of all parts for manufacturing and sent the parts out for fabrication. Unfortunately, by the time the parts came back, there was only about a week and a half left in my internship giving me little time to work on perfecting the process. Never the less, I was able to assemble the jig and conduct preliminary testing on the various processes involved.
Related Project Links
This page outlines only my most recent 3 year's involvement with the First Robotics Competition (FRC). In my last two years, I mentored a world renowned team (1114) and made it to the world finals both years in a pool of 4000+ teams worldwide.
I made and programmed this CNC mill with a friend. It is almost entirely made out of lego and uses the lego robotics system (mindstorms NXT). It reads industry standard G-code and cuts out the desired shapes in soft floral foam using a spinning drill bit. Pictured are some of its milled creations.
I designed my own high end 3D printer instead of buying one for my own projects. My design featured many innovative features that I developed through significant research and testing allowing for unmatched print stability and consistency.
This project was part of the 2016 First Robotics Competition and was a perfect example of a "moonshot" project. The design was unlike anything any of the other 4000+ teams had seen and was superior than anything else that had been developed.