btrettel wrote:D_Hall, when using strain sensors, is something like a half or full bridge circuit adequate to remove the bending stress that would be problematic? To get only longitudinal stress multiple strain gauges can be used, one on the top and the other on the bottom of a beam. The strains will (or should) cancel each other out.
As for post processing, will determining the fundamental frequency and filtering out multiples of it be adequate? I can do that without too much difficulty. I could see this being problematic for stiffer barrels though.
In strain gauge scales, the common practice is to place the gauge on fulcrum points with the load supported on pivot beams so the only direction of force on the gauge is directly into and out of the gauge. In the case of a large object, the bias voltages are all wired in parallel and all the outputs are summed in a resistive adding network. I learned this servicing truck scales at a cement plant.
The scale has 6 sensors. Each sensor was a short rod with supports on both ends and one in the middle. The scale was tied down on one end so trucks stopping and starting could not shift the scale bed, but the truck weight was full rested on the 6 sensors. There is one in each corner and one on the middle of each side. The scale was used to measure sand, gravel, water, and cement to mix a batch in cement trucks. The loading of the recipe was automated and was set for the desired finished yardage.
One sensor went bad. I got to figure out which one. For that I used a small heavy load, a steam roller and weighed it in the 4 corners and on each side as well as the middle. The faulty sensor was quickly located.
I know it's too much info for this thread, but I figured someone may want to learn something about big scales.
The ones I serviced looked like these. They have two support points on one side and one on the other in the middle. They measured strain of bending the bar.