LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

how do I increase fft resolution?

I'm currently running an fft calc on a real time accelerometer DAQ signal. By tweaking the DAQ settings or setting my own time average using shift registers I can reduce the dt which does increase the resolution but the number of data points seems to be limited to 50 regardless of what I do, i.e. reducing the dt decreases range of frequencies visible in the freq-spectrum.
 
Thanks,
 
Rob
0 Kudos
Message 1 of 4
(5,245 Views)

Here are simple guidelines

1) the number of lines (answers in the FFT) are equal to number of dat samples passed to the FFT divided by two.  Most vibration experts use 2.56 as the divisor due to slow roll-off of antialiasing filters

2) the maximum frequency in the FFT is the sample rate divided by 2.  Most vibration experts use 2.56 as the divisor to calculate the maximum frequency in the fft, or FMAX. 

so, if you sample at 10kSA/sec and get a 10kSA block of data - you have a 5k maximum frequency in the FFT and there are 5k answers or lines with delta f = 1.  To get a delta f =0.5, sample for two seconds to increase the block size to 20kSA.

 

Preston Johnson
Solutions Manager, Industrial IoT: Condition Monitoring and Predictive Analytics
cbt
512 431 2371
preston.johnson@cbtechinc
0 Kudos
Message 2 of 4
(5,238 Views)
Preston,

If I understand your answer correctly you are saying that by doubling the data sample size
and not changing the sampling rate you can get out of the FFT freuencies twice the range.

I guess I'm missing something because it seems like you'd need to double your sampling rate
in order to get higher freuencies, not just double the data sampling block size.
In the example you gave, a sampling rate of 10kSa/Sec leads to a max freuency of 5k - makes sense.
But the same 10kSa/Sec should not increase the freuency range.
If extend your logic, sounds like I could increase the sampling block to 100M and get freuencies of up to 100M,
while the sampling rate is slower than that.

What am I missing?

Thanks
Zvi

0 Kudos
Message 3 of 4
(5,193 Views)
When you increase the block size, you increase the resolution on the frequency axis.  In my example, the max frequency in the FFT remains at 5kHz, there are now twice as many answers in the FFT and each frequency bin is now 0.5 hz apart as compared to 1hz apart. 
Preston Johnson
Solutions Manager, Industrial IoT: Condition Monitoring and Predictive Analytics
cbt
512 431 2371
preston.johnson@cbtechinc
0 Kudos
Message 4 of 4
(5,171 Views)