measuring low frequency by cyclone

[Gino]

I want implementing in FPGA some algorythm which will measure input frequencies near 50Hz. I need robust solution, not only counter and zero crosing method.

I read topic
system design using cyclone II
http://www.alteraforum.com/forum/sho...ight=transform
I like this metod but I am worry to make low pass filter at this low frequency. I appreciate some idea from any DSP Guru, because I am new in FPGA and DPS design and I haven't enough experience with digital filtration. I'd heard that it is possible to create. I won't waste time by this trying if it is blind way.

I think also about dq0 or park transform but I haven't matlab to try modeling this solution.
http://www.mathworks.com/help/toolbo...formation.html

I appreciate any advice. thank's

[FvM]

All possible digital filter solutions are based on a digitized analog signal. Do you already have a digital signal? Otherwise an analog low- or bandpass filter and comparator would be the most simple solution.

I usually won't implement a digital filter for a simple frequency measurement. A multi period measurement with a one-shot filter to supress ringing zero crossings would be sufficient even for a higly distorted power grid.

If you want to exactly reproduce the mains fundamental in presence of strong distortions and other interferences, a (signal processing) PLL can be helpful. A simple second order IIR band pass will be sufficient in many cases, however.

[Gino]

All possible digital filter solutions are based on a digitized analog signal. Do you already have a digital signal? Otherwise an analog low- or bandpass filter and comparator would be the most simple solution.

I want make whole evaluation the frequency by digitalized analog signal in Cyclone. On DPS would be only AD converter. In this way will be considered later, if the DSP isn't suitable for this solution.

I usually won't implement a digital filter for a simple frequency measurement. A multi period measurement with a one-shot filter to supress ringing zero crossings would be sufficient even for a higly distorted power grid.

the one-shot filter; What do you mean exactly? Make simply digital filtr and then estimate frequency value by zero crossings?

If you want to exactly reproduce the mains fundamental in presence of strong distortions and other interferences, a (signal processing) PLL can be helpful. A simple second order IIR band pass will be sufficient in many cases, however.

This variant I already try simulate with system in Quartus (IP) and ModelSim. I use by filter the CIC core, this core have band pass dependent on decimation factor, and this I don't like. My sinusoidal wave(input signal) is sampled by 25,4kSps and I want process signal fast, idealy every sample. Decimation by my opinion slows computing. I don't no experience with IIR filter's, over this, I will try implement IIR filter on my ADPLL. Thank's for this idea.