We are using LV Order Analysis to assist balancing fans. There is one tachometer pulse per revolution and the speeds can range 500-5000 rpm. The order analysis is an in-house LV DLL called from LabWindows. The DLL returns the speed and the first order magnitude/phase of six accelerometers (NI-9234 & NI-4472 hardware). We have sample length and sample rate as parameters passed to the DLL. However, the length of time required by the DLL to acquire the data does not seem much affected by the choice of sample size/rate. Is this expected behavior? Do we only need sample size/rate that fills one period of revolution?
We are able to balance fans to acceptable levels in most cases, but want to understand how the sample rate/size might effect the quality of results and the ability to balance fans at lower speeds.
