What does standing out look like?
Below are some examples where I stood out above my peers.
Multilumen from single lumens
At MPC one of my projects involved a 3 single lumen extrusions that lead into a single 3 lumen extrution. The previous die design had the transision between these as long as possible. Reasoning like this is sometimes what our intuition says (the longer the transistion the easier to create it), but when you look at what's going on during the formation having a shorter bond length actually makes the mandrels going through the plastic better supported, and less likely to cause cross talk. So I made the transition as short as possible, reducing cross talk.
The same above project had a callout on the customer drawing that a mandrel of a certain Ø should be able to pass "freely" through the round bottom lumen to the corresponding single lumen. Since the single lumen hole was centered, and the multilumen corresponding hole was at the bottom of the extrusion there is some transition between the two. So when the process started creating a sharp transition the "go" mandrel no longer passed freely. During the process, the "forming mandrel" was originally nitinol, which is usually the first mandrel we reach for at MPC. However, in this case we didn't want something that would easily bend causing a sharp transition, so on my recomendation we replaced the niti forming mandrel with an SST forming mandrel and the "go" mandrel started passing freely again.
Transition to suture
One of my MPC projects involved curing silicon over a suture. The customer drawing originally had this silicon plug with a taper that was just a straight taper. In this case this caused any slight misalighnment to make the taper end at a different location on the suture on one side compared to the other, and thin portions that didn't adhere well. There were many rejects to get them samples. A rather seasoned engineer kept chugging away making rejects, but I knew we had to make this more manufacturable. So I looked at it and wondered, where is the atramatic tip like we often make when tipping cathetors? I suggested a radius at the end of the taper like you see on the end of tips. This completely removed the alignment issue and the customer was happy with the end result as well.
Most of the projects at MPC used compression molding. Which involves heating and melting plastic to form it in a die, then cooling it before removing it. Removing before completely cooling can cause major issues. So when a project that involved baking silicon to form HUB's came along, previous engineers had the plates cooling in an ice bath until it was cool to the touch. I realized that we were not melting the silicon, but rather curing it. So I tried opening the plates while still hot and it didn't effect the final product at all. This saved much time and mess since we only had to cool the plates so that they were safe to open, and the parts cool much faster for handling and trimming when removed from the plates.
These same plates were also clamped together using regular hand clamps. I replaced this system with a pneumatic socket driver and screw system which drastically improved yield and visual quality.
LED Fiber System
In order to inspect internal features, you normally need some pretty pricy equipment. However, instead I aquired an LED Fiber system to illuminate the part from the inside out. This made it possible to inspect marker band locations and orientations, voids, and other defects. This system also became quite useful at inspecting the inside of our micro-die's when used in combination with a boom microscope.
Making it happen
When looking for a synchronous lifting hydraulic system there are a few out there, but they all cost too much for a situation that doesn't require such high accuracy. So at a fraction of the price I put together a Synchronous lifting hydraulic controls for ConXtech which synchronously lifted and dropped our beam in and out of our Weld Cell. No product like this exists on the market.
My Mechatronics project Robo Mouse was the only project to successfully implement the "tractor turn" which consists of braking the inner wheel and powering the outside wheel. What made this so difficult? When doing "tank turns" (powering the outside wheel and reversing the inner wheel) you simply need to say go until you hit the line on the inner sensor, then forward. With "tractor turns" we had to set it up to find 4 edges of the line before we made a successful turn.
In my beginning college CAD course I was assigned to properly draw and dimension a 45° pipe with a flange that had holes in it. I thought to myself, how do I draw this without drawing both ends, but showing that both ends are the same. So I put a 22.5° center line through the pipe and just drew one projected view of the end. I had to hold the paper by both ends of the center line for the teacher to see the symmetry after he marked me off.
At ConXtech we were placing stacks of beams next to each other for storage (with the ConX Collar on the ends). I asked one day; "why do you place similarly lengthed beams next to each other, but differently lengthed beams inside each other, why not do the same for simular beams by staggering them?" This may seem obvious looking at the below diagram, but no one else thought of it when looking at the stacks of beams from the side and I improved storage capacity approximately 30% in these situations.
In one of my college classes we were using a program called EES, which allows you to simply input a bunch of equations, and it solves them for you. This particular assignment was to solve for heat output and the mass and mole flows of the combustion of Propane, which boils down to C3H8 + 5O2 = 3CO2 + 4H2O. The teacher said to just use n_dot_C3H8 (the mole flow of propane) = 1 and n_dot_O2 (the mole flow of Oxygen) = 5, ect ect. Then scale these variables to get the "real" n_dot (since it's supposed to be in moles/sec not moles/equation). I said to myself there has to be a better way, and there was. I simply input n_dot_O2/n_dot_C3H8 = 5 and n_dot_CO2/n_dot_C3H8 = 3 ect ect instead. This avoided confusing variables with varying units simplifying the problem.
In one of my thermodynamics course's we were asked to trace the path on a PV chart of a fluid thermodynamic system. The system consisted of a tank with 2 input and 2 output pipes. An input was hot gas entering a tank which combined with the second input which was cool liquid. These combined and then left as saturated liquid and saturated gas separately out the two output pipes. On the PV chart I drew the molecules from both inputs as splitting to both outputs. My teacher marked me off for that suggesting that the input gas went out only the output gas line, and the input liquid went out only the output liquid line. Yet I remembered an early principal in thermodynamics that when a tank is at steady state quality, it's not that no molecules exchange between the gas and the liquid, but that equal amounts of molecules are exchanged between the two. I argued that the outside of the bubbles of the gas likely condensated. He stayed up all night pondering this and took up a good portion of the next class day discussing it, and I got my 3 points back.
In my first thermodynamics course we were asked to program our calculator to linearly interpolate for us (if you know a fluids values at 55° and 60° (from a chart perhaps), but your fluid is at 56.3° you can assume a linear path between the two and get closer to the real values). Everyone else programmed their calculator using single numbered variables. This meant that for every value you wanted to interpolate you had to input 5 other values then re-run for your other values. I thought, there has to be a better way. I opened up my calculators manual and found it supported vectors. So I wrote a program to linearly interpolate two vectors given a single value from the third vector we're looking for.
An IT Story
One issue in IT is computer maintenance, especially when you don't allow your user administrative privileges. As non-administrators they can't even initiate a defrag. And if you need to logon as administrator to do these maintenance tasks you can script it, but you still need to interrupt their work or log on at odd hours and log into a computer often (when you consider how many computers there are). There are some automated maintenance solutions out there that require leaving the computer on, or possibly interrupting the performance of a late working user. So, how do you allow users to run maintenance tasks that require administrative privileges without giving them administrative privileges, and on their scheduale? Well, any logoff script will run as the user, but shutdown scripts run as the computer, which has administrative privileges. So you setup a shutdown script that is visible to the user, that asks if they want to run maintenance. More details can be found at my spiceworks post. This makes it easy for a user to run maintenance on their scheduale, without IT needing to schedule or work around the users schedule, it turns off the computer when done saving energy and wear and tear, and is done when the user doesn't need their computer anyway (for example: shutdown, say yes to maintinance, then leave for the day; or reboot, say yes to maintinance, then go to lunch) saving valuable man hours.