# Microcontroller ADC and output impedance

I have seen the similar questions here: Required output impedance for ADC input? and Can someone help me understand output impedance to an ADC? but I couldn't apply the answers to my circuit - so I may be reiterating the question but I did search first.

Atmel ATmega328P datasheet, section 24.6.1 says that "The ADC is optimized for analog signals with an output impedance of approximately 10k ohms or less". I'd like to be able to approximate the output impedance of a circuit with a 10k pot and an RC network on the wiper, and change values if needed to fall in or near the recommended range. Some other postings have demonstrated the use of Thevenin's theorem with a pot or voltage divider (for 10k pots th=10k/4, or 2.5k), maybe it's entirely applicable here also.

The input signal is not high frequency, the pot is hand adjusted randomly from 0-5V (it's a paddle controller for a pong type game). To smooth the voltage changes, a series RC network will be connected to the pot wiper. I've tried two RC combinations with the same time constant (R1=10k, C1=100uF and R1=1k, c1=1000uF) and both are acceptable and respond similarly in real life.

Attempting to find Thevenin equivalent impedance here, I've removed the cap (the load) and shorted the pot's voltage supply. R1 and R2 will be the pot divided into two 5k resistors in parallel, with resistor R3 in series. If R3 is 10k, then the result is 12.5k. If the R3 is 1k, the result is 3.5k which appears to be the better selection - or have I missed something, like the effect of the capacitor?

Thanks for looking.

• Please include a schematic of your circuit. Do you expect each of us to redraw it from your text? Sep 14, 2016 at 17:09
• Sorry, I thought it was in there - an edit will be done shortly.
– lrom
Sep 14, 2016 at 17:18

You have done it correctly.

I think that the optimum would be to use the 10K pot with a 7.5K 5% series resistor on the wiper and a 150uF capacitor (or likely 100uF is okay). That's because the cap will be physically smaller than 1000uF and the time constant will vary much less as the pot is turned.