For a new project I have selected the ADG1407 dual 8:1 mux. Three of these are used as part of a programmable 6-pole Butterworth filter (8 cut off frequencies) and there are multiple channels per instrument, with each channel having it's own filter.The filter is a cascade design, 3 stages of 2-pole Sallen-Key.
During prototyping I have noticed that the cut-of frequency fc of all filters is low. For example, for the 1kHz filter it is ~910Hz. This is true for all channels and the fc variation across channels is low.
I believe that the on capacitance of the switches are affecting the filter. The ADG1407 datasheet gives typical values of 70pF for CdON and CsON. This however is at f=1MHz while my frequency range is 100Hz to 20kHz.
I have some questions:
1) Do CdON and CsON vary with frequency (an approximate answer would be ok)?
2) How do CdON and CsON vary with temperature (again an approximate variation or even a guesstimate would be ok)?
3) Can you provide me with a reasonable model for the ADG1407 mux when one switch is closed? I looked at MT-088 Figure 8. It shows CDS(ON) across the switch. But the datasheet doesn't give a value for this. What I have is 8 inputs connected to S1a to S1h and one output DA. So say S1a is switched to DA. Should I include CD(OFF) for the 7 switches which are not closed as they are internally connected to the drain? If I could get a reasonable model I could probably determine the resistor values needed for each stage.
4) I looked but could not find a SPICE model, I guess it is not available?
Thanks for any help you can provide.