fsk modulator ic
FvM said:
The 1 nF capacitor, as far as I understand, is a kind of loop filter, some element of this kind is most likely required for regular operation.
Loop filter in PLL is usually a low pass filter. A series capacitor will behave like a high pass filter, isn't it? This looks like a conflict.
FvM said:
I don't understand your hypothetical digital frequency difference concept.
I mean, the frequency difference between two inputs can have only two different values for all the time. For example 10kHz and 20kHz, or 0Hz and 1MHz, or 0Hz and 100kHz; no values between these two can exist. FSK is a digital modulation, isn't it? There are only two different frequencies in the modulated signal. For example, a 900kHz sinusiodal represnts a logic 0, and a 910kHz sinusoidal represents a logic 1. No values between these two frequencies can exist in the system theoratically.
What I asked was, if there are no frequencies except for two frequencies, and they are clearly sperated (e.g., 900kHz and 910kHz exist in the inputs, no frequencies like 908kHz or 901kHz exist), then do we still need a 3rd order filter?
FvM said:
A PLL FSK demodulator depends on tracking the input frequency, involving also transient phase differences. The unfiltered frequency output would show some ripple then, and a first order filter can hardly provide fast response and no modulation resiuals at the same time.
I think the higher order filter is required for the transition regions between two frequency components. If the frequency difference between the two inputs are constant we expect the voltage level at the 7th pin to be DC (thus, we don't need a low-pass filter). But at the moments of transition, 7th pin output can have highly rippled voltage levels, therefore we need a higher order filter for this worst case. Am I right here?