Being a Real Engineer

Double feature today.

Now that I’ve been in a real job for a while, I can make a few statements about a job and how well Harvey Mudd prepares you for it. OK, it’s not a real job, it’s an internship, but I’m getting paid to do it.

First, a quick overview of my job. I’m an intern at CornerTurn. They’re a small company that broke off from a larger firm so that they could do engineering without all the bureaucratic methodology. They specialize in special application boards, although they’ll do a variety of tasks. One of their claims to fame is a special internet router designed for boats. People on their yachts want high speed broad band internet, but that requires a giant satellite dish. So, they designed a special router that can use multiple antennas and combine their signals into a single internet connection. This means that they can use a couple of small antenna, and treat it as a single Ethernet connection to your computer. Pretty neat, or at least I think so.

My particular job is to do whatever they want. This so far means that I’m working on three separate projects in parallel. First, I am building a printed circuit board for monitoring liquid levels in a storage tank. Second, building scripts for hashing data from a remote sensor that’s been acting up. And lastly, trying to understand and apply energy harvesting for another remote sensor.

The cool thing is I’ve actually been helpful. Not in the thanks for doing the grunt work helpful, but some of my ideas have actually found important results. Looking at the second project, I’ve been building MATLAB scripts to parse data that we get from a sensor. It’s supposed to find itself using GPS, and then report. The problem is that for some reason we’ve been having periods where there’s no GPS data in the database. So, I’ve been parsing NMEA protocol and console dumps trying to help track down the problem.

One of the things I did was to use what I’ve termed gap analysis, or plotting the duration of the time gap in the signals. I’ve recently learned that gap analysis is a term for something else, so probably delta analysis would have been better. This creates an interesting graph where these peaks show up, indicating a giant gap in the data stream. I’ve also created ways to compare these gaps in the NMEA packets the database makes with the attempts of the sensor to try and transmit a GPS fix, and to track throughput of packets. One key element we found was the sensor would try to transmit, but either it’s getting lost or the database drops them. It’s still unsolved, but hopefully we’re moving forward.

As far as Mudd getting us ready, I think the answer is they’re doing a good job. I have had to put my knowledge of energy and computer science into practice already. I’ve done some physics and mechanical applications. As a bonus, the theory has given me a really strong base. Sure I don’t know everything, but they can start explaining something and I’ve figured out what’s going on and what they want by the end of their explanation. Plus it’s not all spoon feeding, I’ve been working mostly on my own to figure out the energy harvesting, and that’s been quite the experience. It turns out that there are lots of options, not just solar cells. There happens to be vibration and heat as well, but those only seem to work well if there are strong driving forces. Not so great in the middle of nowhere, which just so happens to be where we want this sensor to be.

All in all, it’s been pretty good. I even get my own desk that’s getting filled with lots and lots of datasheets. That’s right, better get used to scouring datasheets people, they’ll be with you for many years to come.