Sorry for the delay in this but I now have a response from the team
RMS and TRUE RMS both require an integer number of cycles to perform the measurement. Each time the waveform rises past, falls and rises back to the median value is a considered cycle.
With that in mind consider channel A of the psdata file:
The PicoScope rulers are set at 0.4-116.7ms. The nearest integer number of cycles to this is 9.99-110.0ms. Using excel this gives us 196.587 (cells B1231:B5320) which is pretty close to PicoScope 196.6.
You have chosen cells B782:B5622 (-0.984-117.4ms) which gives the wrong result.
Even if we were to use PicoScope ruler values of 0.4-116.7ms (cells B839:B5596) we get 195.0 which is still a lot closer than the customers result.
Hope this clears things up