methods to calculate harmonics from voltage and current

pad · Joined: 29 Nov 2007 Posts: 15

hi all,
I am working on a project where I need to show total harmonics distortion of voltage and current signal.
Currently I am taking 32 adc samples of voltage and current in a cycle(20 ms) of 50Hz system and calculate rms values of V & I with almost 2% accuracy. Now I need to calculate V-THD & I-THD from this signals.
I have two choices. 1] - using hardware ie by making LPF of opamp (50 Hz)and calculate rms value of this fundamental wave. At same time another adc will read and calculate unfiltered signal and this is with harmonics.
So (unfiltered - filtered) will give me harmonics content. Is this correct?
and another way 2] Digital filtering - this can also be useful but I have very small knowledge about it. If anybody will help me on this topic,will surely go for it as it requires more software and math and less hardware change. And also can calculate individual harmonic using this.
So Friends, need your help and opinion on this topic.
regards
PAD

dyeatman · Joined: 06 Sep 2003 Posts: 1934 Location: Norman, OK

How will this relate to CCS and the PIC? What PIC are you planning to use?
_________________
Google and Forum Search are some of your best tools!!!!

pad · Joined: 29 Nov 2007 Posts: 15

dyeatman · Joined: 06 Sep 2003 Posts: 1934 Location: Norman, OK

This is a CCS C forum and CCS C and PIC are required by the forum
otherwise the moderator will likely lock the thread as being off topic.

Version 4.020 is a very old and buggy version and will likely give you lots of
problems if you try to use to use it. In addition, it will take quite a bit of
horsepower to perform the kind of analysis you are describing so I would
expect you would need a PIC24 or PIC 30 minimum.
_________________
Google and Forum Search are some of your best tools!!!!

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

It sounds like you need to do a FFT of your data. That will give you the amplitude and phase of each harmonic. If you can live with an update rate of 1 reading/second or so I think any reasonable PIC should handle it. You do want to update your compiler though.
_________________
The search for better is endless. Instead simply find very good and get the job done.

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

I agree, that a 16-Bit PIC would be fine. But nothing has been said about real-time capability, so even complex signal processing, e.g. by FFT would be basically possible with a PIC18. Classically, THD is measured with band stop filters rather than by subtracting power of RMS and fundamental. But an FFT power spectrum (magnitude only) is an interesting option. Taking only 32 samples per period may be a bit low to measure the total harmonic content.

pad · Joined: 29 Nov 2007 Posts: 15

thanks for reply dear FvM & Sherpadog.
I am thinking that once I can find THD of V or I,though slow updated at start( once in a sec.)but will be my first step towards the goal.After that I ca definitely speed up it by higher end controller.
I am trying FFT so that I can remove 50Hz signal samples from input and will compare funda.V & I values with PQ meter.If accuracy is <=5% ,it will absolutely OK for me.
I am searching forums and google for same and have some useful data. mostly code is in assembly for fast calculations.I am new to ASM for PIC but studying it.
So friends,looking forward for your valuable suggestions and help on FFT.

PAD

Ttelmah · Joined: 11 Mar 2010 Posts: 19549

Have a look at:
<http://www.ee.washington.edu/class/472/peckol/code/FFT/brent_fft/>

Not much except changing types to suits CCS, needed.

Best Wishes