Actual 16f886 analog impedance?

[Mosaic]

Hi all,

The 10K impedance required for adc sampling 'within the specified time frame' does not actually indicate what the tru load of the analog adc line is.

I'd like to know what it is to determine whether its effect requires an op amp buffer for measuring conductivity etc. Also i can allocate up to around 1millisec of 'sampling time' to improve accuracy of the sample.


thx

 

[ericgibbs]

Hi all,

The 10K impedance required for adc sampling 'within the specified time frame' does not actually indicate what the tru load of the analog adc line is.

I'd like to know what it is to determine whether its effect requires an op amp buffer for measuring conductivity etc. Also i can allocate up to around 1millisec of 'sampling time' to improve accuracy of the sample.


thx

hi,
This is a LTSpice plot of the PIC's ADC circuit.

The internal 'impedance' is approx 8K into a 120pF capacitor.

A 1mSec sampling time is not necessary.

E.

 

[Mosaic]

I am using a capacitive 'PCB' home made sensor as a moisture detector for soil. Co planar type. The capacitance is about 50Pf to 400pf dry/moist range measured @ 100Khz. Reads about 4 times that @ 1Khz.

Anyway, I note that the leakage current on a schmitt PIC pin is about 1uA max, presumably @ 5V Vdd. Does this mean the lowest resistive loading is about 5 M ohms on the input pin?

Trying to figure out the rate of discharge of the sensor into the input pin and determine a load resistor in parallel to make for a reasonably measurable time to discharge to a logic low.
PIc is running @ 8Mhz.

 

[Mr RB]

May I ask why you are using the ADC to measure a capacitive sensor?

It's more typical and many times more accurate to use the capacitive sensor as the timing element of an oscillator and use the PIC to measure period, (or the average of many periods for high accuracy). That also gives quite a linear conversion from capacitance->period.

 

[Mosaic]

Sorry for the unclear post.
I am presently using a tristated schmitt input to measure the cap discharge in uSec.
The cap is charged by the pin's high dig output and discharge is measured by the same tristated pin.

 

[Mr RB]

Still a little unclear (to me). I guess you are discharging the cap slowly, and using the ADC to tell when the cap voltage is <X volts? I think your cap of 50-400pF is too small for this method to work well, which might be why you have trouble with impedances.

If your PIC has a comparator, that will give a fairly precise voltage setpoint for charge/discharge and can make a capacitor dependant oscillator with very few parts count, just 4 resistors.

I used a similar system as a high resolution capacitance meter, the basic circuit would be ideal to work with your 50-400pG capacitive sensor. Here's the cap meter project showing how to use the comparator to measure a cap;
https://www.romanblack.com/onesec/CapMeter.htm