October 14, 2015
Repeatability vs. Desired Outcome in Precision Shaft Alignment, Part 1
By Stan Riddle
When mechanics align rotating machinery, they expect two outcomes. One, they expect their alignment values to be repeatable. Two, they expect shimming and moving machinery to produce a desired outcome-namely, to be in tolerance.
In this paper, we’ll discuss repeatability, which is, performing the same measurement twice, and getting the same results.
If alignment results are not repeatable, after measuring two or more times, most people blame the tool! In fact, a laser alignment tool is simply a calculator with laser beams for measuring, and it simply does not have the capacity to lie! On my calculator, if I put in 2+2, it will ALWAYS equal 4! If it doesn’t, then I didn’t do something right. If your alignment results aren’t repeating, it could be something you’re missing.
So, what should you do if you measure twice, one right after the other, and your values are not repeatable?
1. THINK! What could be changing (moving)? What could you have done differently?
2. Check for looseness. In the brackets, rods, and transmitter/detectors. In the coupling hubs (if you are mounted on them). In the movable and stationary machine foot bolts.
3. Observe if there is any vibration from external sources, such as from a nearby machine. Perform a repeatability test if your laser system has this function. Shut the offending machine down momentarily, if you can.
4. Make sure you are rotating the shafts in the same direction during the measurement process. If you really want to troubleshoot for accuracy, try to measure from the same angular positions each time. Use inclinometer values.
5. Make sure the transmitter/detector lenses are clean. Dirt and dust can refract the laser beams.
6. Confirm that the bearings in the machines you are aligning are OK by performing a lift check.
7. Make sure you are not using the brackets as “handles” to rotate the shafts. And make sure they aren’t bumping against bolts, conduit, oilers, guards, or anything else that could move them.
Lastly, you need to be realistic about how “repeatable” is “repeatable”. If your coupling values are changing a few ten thousandths of an inch, should you care? If you measure three times, and all three measurements are minimally different, but each one is in tolerance, shouldn’t you just save it, and pack it up?
Next time, we’ll talk about desired outcomes, or what happens when 2 minus 2 doesn’t equal zero!