Description of the problem
How to improve the performance errors when running power analysis with math calculations (Full FFT, Harmonics, ...) at high acquisition rates (15 MHz for example)?
Description of the solution
The first and most important step to solving this issue is to check that you have the option to Store IB buffers for vector channels turned OFF. You do this by going to the Options > Settings > Performance > Memory > Store IB buffers for vector channels and unchecking the circle on the right.
An additional measure you can take is to divide the calculation frequency, meaning the rate at which data is acquired will remain the same, meanwhile the rate at which math calculations are done will be reduced, improving the calculation speed.
To do this you will have to set up a new math channel. Go to the Channel setup > Math > Formula. Then pick the channel you want to perform calculations on from the dropdown list on the right side. Name the math channel at your leisure. Then go to the Time axis and choose Sync for the timebase. For the SRDiv choose a factor (number) by which the calculation frequency will be divided. In this example the chosen factor is 10, meaning the math like Full FFT or Harmonics will only be calculated on every 10th sample. The frequency at which the computer will have to perform calculations will therefore be 10 times smaller than the acquisition rate.
Then you will have to change the Frequency source and the Waveform channel. You do so by going to the Channel setup > Power analysis. Here change the Frequency source to Arbitrary channel and the Waveform channel to the math channel you had set up to divide the calculation frequency.
Additional information
Power analysis, Power quality analysis