Hi everyone, I have a square signal at the output of a photodetector. The frequency of this signal is MHz. It's a noisy signal. First of all, i want to sample this signal with and ADC (fs=10 MHz, 14 bit). Then filter the signal converted to recover the original signal without noise. Can someone help me about choosing the best platform to that( FPGA, µC, DSP or raspberry ...)please ? I need also you help concerning the choice of the ADC. Thanks in advacnce. Best regards
Guest
#4525433
jeorges F. wrote: > First of all, i want to sample this signal with and ADC (fs=10 MHz, 14 > bit). Then filter the signal converted to recover the original signal > without noise. Hello jeorges, are you sure you require this performance? The converters for this task are not cheap (around 30USD per piece) and so is the input stage for the converter. Next is the performance for "filtering" such a data stream in reatime. Can you describe the problem a little more to check whether other solutions would be easier? rgds
jeorges F. wrote: > to recover the original signal without noise. Is the original signal "noiseless"? Where does it come from? Which part of the signal chain is noisy? What must be done with the signal afterwards?
thank you for you answers Lothar & 6a66. This is the description of the problem: the purpose of the system is to treat a sinusoidal input signal, attenuated and noisy so as to output a signal of the same frequency but having a high ratio of signal to noise. the output of this system will provide a square wave. The input is a sinusoidal signal e(t) of frequency fr, marred by predominantly Gaussian noise and possibly having a DC component. This signal has the form : e (t) = A (t) sin (fr.t) + n (t) (A (t) amplitude of the signal, n (t) is the additive Gaussian noise and DC (t) the DC component). - e(t) is an analog signal supplied by a photodiode and a mounting Transimpedance. The features of e(t)are : -5 V <e (t) <+ 5V A (t)> 20 mV SNR> 5 dB Fr = 1 MHz The goal is to have a square frequency analog output signal, compatible logic level (0 V and 3.3 V low state high State) the phase difference between e(t) and s(t) should remain relatively constant with a phase noise less than 0.05 °. Best regards