Bench Talk

I had an environmental monitoring project at Salt Creek in Aspermont, Texas, a few years back. If you have never been to Aspermont, you can get the total experience with a video of a slow tumbleweed. Salt Creek in Aspermont has off-the-charts salinity. There’s a huge salt dome that contaminates the spring source for the creek, and this super-salty water eventually meets up with the Brazos River. The group that owned the property wanted to know the flow, volume, and salinity of the creek. We were to set up monitoring at the base of a small valley, near a concrete bridge.

My consulting partner, a geologist from Oregon, worked out the volume calculations with a flow table and we used a conductivity transducer to help us determine salinity. It was so salty that we were bumping the upper limit of the scale, but it was August and had not rained in a while, so that made sense. A pressure transducer determined the flow and volume via a creek bed flow table and some calculations. We ran the signals from both sensors through PVC pipe to a panel on the bank. It was hot. I was used to working outdoors in that weather, but Mr. Rock Hammer was near to passing out.

We set up a controller with a small solar panel and a battery charging unit. Part of our set up used the Geostationary Operational Environmental Satellite (GOES) system to upload the data. This natural creek was contaminating the Brazos, in effect, and GOES can be used at no charge for approved environmental projects. Eventually, the raw data is uploaded to a server. As a backup, we had the data also going to a memory stick inserted into the port on the controller. Everything checked out when we left, but after a few weeks the uploads stopped.

I am versed in controllers and I understood the sensors and other hardware, but I had never done telemetry in such a remote location. It was frustrating to try and troubleshoot over the phone, perched on the top of a ridge covered in rattlesnakes, while shouting suggestions from the controller company to Mr. Rock Hammer below. Lessons learned: we should have gotten the whole thing totally, 100% working on a bench first, but Mr. Hammer had done this before. If I had more time and better working conditions, I would have banged away at it until the problem was solved. I suspected that a fanless panel for the equipment was to blame. I have seen some crazy random output when a controller gets too hot. But in this case the controller was still sending reasonable data to the stick, so it wasn’t the controller. But I knew little about the satellite link we were initiating with GOES, which made it frustrating to troubleshoot.

Open source hardware reduces the frustration factor. Not all of us are experts at everything we are asked to do. Engineers learn for a living. We solve problems. Had pieces of the Salt Creek system been modular and open source, putting it together and then trying to understand the gritty details of satellite communication might not have been an issue. In solving difficult problems, engineers end up knowing the problem very well. Knowledge is a tool in our tool box, but one person can only know so much. Open source can hand us a bucket of tools, for free, as long as we are willing to learn to use them. As time goes on, hopefully we will get more open source tools to do more things. Maybe I’ll tackle Salt Creek again someday. It would be a steep learning curve for me, but maybe I would finally know what happened. Several years later, the memory stick is still swapped out every 6 months, so the data is being gathered, but I still think about the communications aspect of that project.