SparkFun Forums 

Where electronics enthusiasts find answers.

Everything ARM and LPC
By treez
#159487
Hello,

We are reading a voltage rail with the on-chip ADC of the SAM3N00B microcontroller via a potential divider.
We are trying to assess what is the maximum allowable Ohmic value of the upper divider resistor.
The ADC reading will draw some current through this upper divider resistor, so we must not make this resistor too big, otherwise it will affect the potential divider voltage reading.

Do you know what is the current drawn by the ADC when an ADC reading is taken?


SAM3N00B MICROCONTROLLER DATASHEET:
http://www.atmel.com/Images/doc11011.pdf
By treez
#159539
yes i saw with out 15k upper divider res it would allow a reading (TRACKTIM) every 2.5us by their calculation......thats just not a sensible figure..........so that TRACKTIM calc in the datasheet must have been not relevant?
By mlu
#159561
When reading the voltage rail scaled down by a voltage divider the upper resistance will effectively be the source impedance. We can assume the
voltage rail can supply enough current.

As to a time of 2.5uS, why is this not sensible ? Its a sampling frequency of 400kHz, much lower than the ADC max, but you are feeding it through a
fairly big resistor (source impedance), at least when it comes to high speed adc 15k its a big value.
By jremington
#159562
It is somewhat counter intuitive, but the impedance of a voltage divider consisting of two resistors R1 and R2 is the parallel value, or 1/(1/R1 + 1/R2). Use the Thevenin equivalent circuit to see why.
By mlu
#159563
sure and the smaller resistance will give the effective impedance :) , I guessed that the top one is smaller if sensing the voltage rail of your own processor
By mlu
#159566
A most enlightening response for sure. I do know how to calculate impedances for both series and parallell
coupled inputs, I also can calculate which part of the parallell input impedances with give the largets contribution
to the total. And if you claim that is not the smaller impedance, or resistance in this case, then perhaps you could
try to prove that. In our case with a voltage divider it will be smaller than both R1 and R2 but never smaller than
half of the smaller. And in this application, we are not discussing the exact 10% resistance but the right order of
magnitue. Its not like designing highly tuned resonators.