Development work need not always be ridiculously expensive. Sometimes ordinary hardware store tools and basic shop skills will do a pretty good job.
Recently, I was approached by a customer to design a load sensor for an industrial application. This particular application had great potential for high-volume manufacturing-possibly hundreds of thousands of units, or even into the millions.
The customer needed an inexpensive load sensor design. I investigated several strain-gauge type sensors that were very expensive, oftentimes with prices in the several hundred dollars per unit range. These units were intended for testing lab applications, not mass production.
Granted, these were small quantities, nevertheless several hundred dollars each was going to be excessive. Foil strain gauges are great, but their technology is almost 80 years old. I continued to dig around on the Internet to locate a force sensor that would fill the bill at a moderate cost, and found one.
The sensor assembly consisted of several layers of Mylar that had a conductive coating. As force increased, the conductance increased with a corresponding decrease in resistance. Since this was a proof of concept design, I assembled a small bench-test model using washers, nuts and bolts. The sensor itself was a solid disk of approximately 1 inch diameter. In order to use it, I would need to cut a small hole through the center in order to slip the bolt through it. Conversations with the manufacturer’s technicians confirm that this would not affect the sensor much, and should function without any problems. The general idea was to connect a voltmeter to the sensor’s output leads while gradually tightening or loosening the nut and reading the change in resistance on the voltmeter. It worked perfectly.
Now I needed a place to attach the sensor to the structural assembly provided by the customer. The structure the sensor was intended to go on had many bolted connections. I acquired the CAD model for the structural assembly. Fortunately, the assembly contained many bolted joints, which provided a location for the load sensor. I needed a specific place that would provide a compressive load proportional to the stress placed on the structure. I converted the CAD model to a finite element model, proceeded to modify the contact elements so that the loads would pass through the bolted connections and into the various bolts and washers in the assembly. This was followed by applying a variable force on the assembly and noting whether the stress on a particular bolted joint would respond proportionally to the external load. This was very straightforward, and several locations were identified.
Along the way, I constructed a test stand for the load sensors and the structural assembly provided by the customer. This was nothing more than lumber, nails and some bolted connections. To apply the loads-which at this point all they needed to be was qualitative-I used a two-ton bottle jack I had in my shop, which if I recall correctly cost me around 11 bucks.
Having several potential locations confirmed, the next step was to convert the output signal to a voltage or current that was proportional to the applied load. This assembly was available through the load sensor manufacturer, or my customer could develop a circuit assembly themselves, which would be the most likely outcome because of their unique requirements, and then to redesign the load sensor for manufacturability.
Unfortunately, the project did not progress beyond the point of proof of concept testing. This project was interesting and a heck of a lot of fun as it involved a small amount of electrical knowledge, mechanical and structural design and engineering, and put together a finite element analysis model with ANSYS. In addition, I got to use a number of tools in my garage shop!
Norman T. Neher, P.E.
Analytical Engineering Services, Inc.
Elko New Market, MN
www.aesmn.org