We've mentionned 3 so far, there are a few others. Decide which sampling technique you need.That's why you have to sample the signal at the beginning of every sampling period and hold it still while the conversion takes place. Because the sampling period has a finite length, the signal must remain constant during the conversion. Design a filter to remove undesirable frequencies (to prevent aliasing). ![]() the equivalent dynamic range is then -20 * log (1 / 256) ≈ 48dB for 16-bit, ≈ 96dB) for a singal 5V p-p, using 8 bits, the smallest interval would be 1/256 * 5V ≈ 20mV. How sensitive your sampling will be (what's the smallest variation in voltage amplitude can be sampled) ? That will give you the resolution (in bits) you need.What is the maximum frequency to want to sample ? From there, by the Nyquist law, you would find your sampling rate (e.g.Characterize the signal you want to digitize.I put more than you probably need but that gives you a big picture : ![]()
0 Comments
Leave a Reply. |