Detecting Fundamental Frequency From a Signal

Are there any ICs available which I can use to extract the fundamental frequency signal from a random signal. I know the frequency of the signal i have to extract.

If not what other methods can i use to extract the fundamental component from a signal?? Maybe microcontrollers??

Strictly spoken, a random signal has no fundamental frequency. So there may be doubts, what you exactly want to detect.

A dominant frequency component in a limited frequency band can be detected with a PLL, e.g. a NE567 tone decoder IC. Also a bandpass filter with zero crossing detector can work. A FFT is a more general approach for signal analysis, but the frequency resolution will be limited by the discrete frequency representation,

Is the fundamental frequency constant, within a reletively small range +/- 5% ?. If yes, use a PLL to decode the signal for example a NE567 or 74HC4046A. If no, than you have to do much more work and the best way is use a DSP and perform a FFT is possible.

Yes in fact the fundamental frequency is constant at 50Hz. Could you give me some more details about PLL using the above mentioned ICs. Preferably some link where this has been implemented that you may know..

The said IC's are self-contained PLL devices, the datasheets can be found at the internet. You should tell a bit more about the nature of the signal, how much is the harmonic content and non-harmonic signal components. This would decide about the requirements for the PLL phase detector.

4046 has a digital signal input, so it would need pre-filtering and and a comparator to process an analog signal. NE567 (or LM567, which is the same) has an analog input, but is demodulating the input with a squarewave, so it has a certain sensitivity to harmonics. But it would be usually able to lock on a fundamental in a complex signal. If strong interferences exist, an additional bandpass filtering may be helpful.

Hi FvM, Your incorrect 74HCT4046 also has an analog input with a 20mV sensitivity, same as the NE567. The NE567 is a little more easy to work with and the datasheet will help you select components for the frequency and bandwidth. Best is to use the NE567 for you since it's very easy to work with. Check the datasheet with google and a simple application diagram is available. Take the fundamental freq. at 50 Hz and give yoyr filter a bandwidth of some 10%. use only 1% resistors or maybe even a potentiometer to tune the circuit to 50 Hz.

No additional filters are needed with the NE567 if you make your bandwidth filter on the NE567 narrow at 10%, check the datasheet for this. I did use this chip for locking on a signal in the video signal that was only small compared to the total video signal and it worked perfect.

To protect the input its good to place two diodes antiparallel at the input signal !.

It looks like your signal is the 50 Hz mains frequency, if you select the components for your NE567 you can also lower the bandwidth to 5% but not below this since 50 Hz mains frequency can vary with +/-2% add to this the component errors and you have 5%.

regards, Paul.

The signal is actually the current consumed by a non linear load which is being fed by a generator operating at 50Hz. So i need to remove the 50Hz signal from the current pattern of the non linear load, so that i can used the "50Hz free" signal to reproduce all the harmonic currents locally through inverters so that the generator does need to generate the harmonics demanded by the load.

Will PLLs with NE567 be helpful to me? If so, could give some hints on removing the 50Hz component from the current signal using the NE567

Hi, Simple question. Is the current constant too ?. I think its better to use a DSP for harmonic content removal, dsPIC33 or 30 would be a good choice and not so complex since your freq. is very low at 50 Hz.

Hi FvM, Your incorrect 74HCT4046 also has an analog input with a 20mV sensitivity, same as the NE567.

Click to expand...

Yes and no. Yes, it has a self biasing feature and an inverter chain acting as a comparator. No because it's not an analog synchronous demodulator as implemented in NE567. I forgot about this, because I always used 4046 with digital input signals. The difference is important in presence of stronger interfering signals. NE567 can detect a weak signal in noise, 4046 can't.

i need to remove the 50Hz signal from the current pattern of the non linear load

Click to expand...

That's quite different from the usual understanding of "Detecting Fundamental Frequency ". Actually you need to substract the signal with exact magnitude and phase, respectively filter it from the input signal.

As far as I know from APF (active power filter) literature, it's often operating based on fourier transformation. But you should review the available literature yourself.

So i am guessing PLLs wont work for me , right?

Do i have to used FFTs or some filter like Notch filter might work?

As FvM already told you have to use APF type of filters and they use a notch filter structure. FFT is not used for this. They are completely digital and track the fundamental signal and suppress this signal in the digital frequency domain. PLL's are not usefull for this, its possible but it requires a lot of OPAMP's and phase shifters and filters to get what you want. Biggest disadvantange of this setup is the fact that it drifts overtime and needs recallibration every now and then, digital circuits based on xtals and dsp's do not have this problem.

But is the current stable ?. if yes you could maybe use an True RMS measuring solution. measure the total current with a True RMS chip and subtract from this the known fundamental current, leaving the harmonic content for further processing.

I have seen many APF papers referring to FFT and/or PLL filters. There's a lot of options.