RMS calculation of non-coherent sampled data

max3031

Hi all,

I need to calculate RMS of the voltage signal that is in the form of the square-waves of certain frequency (10 to 15Hz). The problem is because my system is sampling this signal at the constant rate 128Hz so I am basically doing non-coherent or asynchronous sampling. Because of this RMS calculation has errors. So what would be the correct way to resolve this?

By the way i can't synchronize my system with incoming signal. Not possible in hardware. So all corrections need to be performed by software.

Regards

over 5 years ago

0 Tom Hendrick over 5 years ago

TI__Guru**** 172656 points

Hi Max,

Can you let us know what you are using to sample this signal?

0 max3031 over 5 years ago in reply to Tom Hendrick

Prodigy 110 points

sar adc is used. It is properly limited in frequency.

No need to worry on hardware.

Everything is fine there.

All where the problem is because I am sampling at 128Hz (non-coherent) and square wave signal fundamental frequency is 10 to 15 Hz.

So to get proper RMS without errors, what needs to be done with this sampled signal?

Here is a application note that is related to my problem. It is discussing sine signals (one frequency), in my case it is square wave so multiple frequencies. slaa570.pdf

0 Tom Hendrick over 5 years ago in reply to max3031

TI__Guru**** 172656 points

Hi Max,

I get the impression that your application is a little different than the one described in SLAA570. That document is describing how to deal with the frequency drift of an AC mains signal (one sine wave) by adjusting a PLL and controlling the sampling rate of a delta-sigma converter (a hardware solution). Without knowing the exact frequency of the signal you are measuring, you will always have some error in your calculation. Assuming you have a signal that periodically varies from 10Hz to 15Hz, can you just assume you have a 12.5Hz signal? Your RMS would then be amplitude*SQRT(Duty Cycle).

0 max3031 over 5 years ago in reply to Tom Hendrick

Prodigy 110 points

Hi Tom,

So my square wave signal is going to change its frequency as well in the range of 10-15Hz. I am doing moving RMS average on it but since I am sampling it incoherently there is a problem in terms of RMS precision and convergence to stable value. Here is a related work to my problem. G3.pdf

0 Tom Hendrick over 5 years ago in reply to max3031

TI__Guru**** 172656 points

Tell me what processor you are using and perhaps I can find a software guy to help you.

0 max3031 over 5 years ago in reply to Tom Hendrick

Prodigy 110 points

Did you check the second paper that I sent? It looks like you did not.

0 Tom Hendrick over 5 years ago in reply to max3031

TI__Guru**** 172656 points

If you are talking about the G3.PDF file and the "PC DAQ" - I saw that. I also saw that the paper is referring to non-coherent sinusoidal wave-forms.

0 Tom Hendrick over 5 years ago in reply to Tom Hendrick

TI__Guru**** 172656 points

Hi max3031,

Do you have any additional detail for us? This really sounds like a software conundrum and without any detail on specific Texas Instruments parts associated with your project, I'm not sure what we can do to help you through this difficult problem.

Audio

Audio forum

RMS calculation of non-coherent sampled data