I have had some mild success with a project but am looking for some advice.
I am measuring the distance a steel pile is driven into the ground. This is done so by a 2T hammer and is dropped from a specified height so is very violent in nature. There is an elastic portion of deformation and a permanent portion. Both are important but the elastic compression and rebound happens very rapidly. The permanent I can always measure accurately after the impact is over.
So far I was able to use a liner potentiometer to measure this movement to a level of accuracy that was acceptable..I verified the movement of the linear potentiometer with a dial indicator. However due to the violent nature sometimes the slide on the potentiometer slipped out of the clamping bracket I made due to the vibrations.
Is their a better way to do this? My ideas so far are:
-Plunger style linear potentiometer. The one I used was a slide style so the lever action and vibration wrecked it after 20 drops of the hammer
-LVDT..this seems over kill and will likely be very costly
-Using a PS2 mouse and placing it close enough to the pile. Will this read fast enough. The movement I am measuring occurs in less than 2 seconds
-Linear optical encoder/ disruptor type of thing could be rotary encoder but it's hard to convert the linear to rotation so rapidly. I thought a linear gradient strip moving through a sensor and reading the disrupts could work. I dont know how to sense if its downwards movement or rebound from the elastic compression. Could I couple this with an accelerometer to work out direction
Ideas I have rulled out:
-double integral from accelerometer due to inaccuracy
Yes it is in the milli seconds. I diddnt measure it as the timenwasnt important but I think I was getting a serial reading every 3 milli seconds and I only have a out 20 or 30 data points showing actual movement at most.
What sort of optics do you have in mind? I can always try to clean things up and shield it somewhat.
I think the amount of force needed to ensure a wheel driving a gear diddnt slip would be hard to apply as the apparatus has to be independent from everything in order to measure the displacement. I could always give it a go though.
regancooper:
Ideas I have rulled out:
-double integral from accelerometer due to inaccuracy
Why? I'd have thought this is a pretty good solution.
The inaccuracy of doubly integrating an accelerometer is drift. Drift that increases with time.
Here you know when the item is stationary, and can clamp position and velocity values during
this time, and calibrate the acceleration due to gravity.
During an impact the readings will be clearly distiguishable due to the large magnitude, so
turn on position estimation only during the impact and rebound. Should be able
to calibrate this from the other sensors.
Issues might be the high accelerations involved, and the need for rapid sampling.