Signal smoothing, half width? sample rate? sample size?

[web18]

The hardware i am using is a 16bit NI-cDAQ-9174 with a NI 9205 analog input module. I am taking Temperature and RH measurements over a long period of time and have my DAQ assist in a while loop so that data graphs upate every second. I am trying to filter out noise using the Smoothing filter express vi, using the moving average setting. Since i want to collect data for up to 2 hours i do not want to have to high of a sampling rate that way i dont have a billion data points. but i also dont want to sample to slowly so the data isnt skewed by noise. i would prefer to be outputting 4-5 data points per second. I figured i could sample at 1000 samples/s and take the average of every 200 samples to get 5 data points per second. Since the i only care about the change in values over a long period of time i dont care about having high resolution of data per second but i do want accurate data.

So my questions are:

Does the smoothing filter express vi do what i described by starting with 1000 samples and averaging them down to 5? Im pretty sure it doesnt and it just weights each data point but i want to confirm.

How could i do what i described by starting with 1000 and ending with 5 data points? I tried the running average subvi from the forums but ran into problems with data types. It used a random number gen not dynamic data.

Would it be ok to just sample at 10 samples per second and use the smoothing express vi? would this not effect the data to much? What half width would you recommend? i have seen some people talking about choosing the wrong half width can effect the data.

Thank you for any insight into this matter.

[KateB]

Moving average gives the same number of points. You must use normal average to average 200 points every 0.2s at 1kS/s sample rate.

[Henrik_Volkers]

Have a look at the

Decimate (continuous) VI

Owning Palette: Signal Operation VIs

Requires: Full Development System

Continuously decimates the input sequence X by the decimating factor and the averaging Boolean control. Wire data to the X input to determine the polymorphic instance to use or manually select the instance.