Thanks.
here:
View attachment 242415

Your small capacitance values to measure with the large resistance will also have to contend with stray capacitance as well as the input impedance of the A/D input. If you stick with this idea, consider adding a very low bias current op amp follower before the A/D input. Your errors should be not too difficult to approximate. Typically, the A/D converter will spec a minimum source impedance for whatever is setting the voltage on the input. Perhaps start there. Again, much of this goes away with a voltage follower with very low bias current buffering the capacitance to be measured. Then, calibrate away the stray capacitance.

 

Your small capacitance values to measure with the large resistance will also have to contend with stray capacitance as well as the input impedance of the A/D input. If you stick with this idea, consider adding a very low bias current op amp follower before the A/D input. Your errors should be not too difficult to approximate. Typically, the A/D converter will spec a minimum source impedance for whatever is setting the voltage on the input. Perhaps start there. Again, much of this goes away with a voltage follower with very low bias current buffering the capacitance to be measured. Then, calibrate away the stray capacitance.

Thanks.
I am looking for better solution, won't insist on a plan now.

 

You still need to select a range of capacitance values that you want to measure.

 

GOhm scale resistors sometimes are presented in catalogs like Octopart, Farnell, Digikey etc. However I am sniffing out the danger if You ever try to use it for high-voltage divider. If 2W full sized DIP then 750V is max max max over the single resistor body, or 250V over 0.5W and 150 V over 0.125 W and 50...150V for different size of SMD. So, for example for my 36 kV DC divider I used stack of pcb, where each pcb took 4 lines of resistors with some 50 pcs in line and such plates was 8. So 8x4x50x50=80 kV instead of demanded 72 kV. Simultaneously this also is an answer where to get GOhms when only MOhms are available

 

GOhm scale resistors sometimes are presented in catalogs like Octopart, Farnell, Digikey etc. However I am sniffing out the danger if You ever try to use it for high-voltage divider. If 2W full sized DIP then 750V is max max max over the single resistor body, or 250V over 0.5W and 150 V over 0.125 W and 50...150V for different size of SMD. So, for example for my 36 kV DC divider I used stack of pcb, where each pcb took 4 lines of resistors with some 50 pcs in line and such plates was 8. So 8x4x50x50=80 kV instead of demanded 72 kV. Simultaneously this also is an answer where to get GOhms when only MOhms are available

Thanks.
I'll take in mind.

 

Yep. The HV divider the the CVS-8 is a bunch of 625K (I think) resistors in series.