nRF52832 SAADC circuit R_INPUT value

[Farhang]

Hi there,

I am getting some error in my SAADC measurements and I believe it's due to the R_INPUT forming a parallel resistance with my voltage divider at the input (the voltage divider is meant to scale down the voltage of interest to under 2.4 Volts so it can be measured by the MCU). The values I have chosen for my divider are 1M and 68K.

I would like to know the typical value of R_INPUT so I can estimate what the equivalent resistance of my voltage divider is. The specification says R_INPUT is >1M but this doesn't help as it covers a huge range and theoretically it covers infinite resistance (open circuit).

infocenter.nordicsemi.com/index.jsp

I suspect there's a good reason for this but if not I would like to know the typical value so that I can scale my measurements accordingly.

[Jørn]

Hello Farhang

If you look at the bottom of the electrical specifications here you can see a more detailed schematic of the ADC input. The 1 MOhm resistance is the input resistor you see in parallel with Cpad. When the pin is configured for the ADC, PSEL closes and Rinput is in parallel with RLadder, which is, according to the SAADC electrical specification, 160 kOhm. The total input resistance seen into the analog pin will then at the very least be 1MOhm||160kOhm=137930 ohm, and at the most be 160kOhm.

Keep the acquisition times in mind when designing your input network. Take a look here

Best regards

Jørn Frøysa

[Farhang]

Jørn, Thanks for the reply, but i need more explanation: First of all, we're bypassing the ladder resistor with (NRF_SAADC_RESISTOR_DISABLED aka SAADC_CH_CONFIG_RESP_Bypass), so that makes things easier. Now my only question is why the exact resistance of R_Input not listed in the spec? Is it not a simple resistance similar to R_Ladder(which is listed exactly 160K without a > sign )? Having it listed as >1M doesn't make things simple to calculate the equivalent: In our circuit, we have a 1M(pull up) and 68K(pull down) resistor voltage divider at the (external to the chip) input, now if i want to calculate the equivalent resistance of the bottom side of our divider, it can be 68K||1M=63.67K or 68K||100M = 67.95K or 68K||(infinite resistance) = 68K! This means that the scale of my divider Rdown/(Rup+Rdown) can vary from 0.064 to 0.060 which is a 7% relative difference=large error

[Jørn]

Unfortunately R_input does not have a exact resistance as it is not a "real" component. It is a parasitic resistance equivalent due to current leakage in the ESD structure, and it will vary as a function of temperature. I have spoken to one of our designers and he could inform me that the minimum value in the datasheet is inaccurate and the actual resistance lies above 10 MOhm. Which should bring your maximum deviation down to approximately 0.67%. Also remember to make sure the acquisition time is long enough for the sampling capacitor to settle.

[Farhang]

Thanks for further investigation Jørn. I calculated back the internal resistance seen into SAADC and it was different from NRF52 to NRF52 unit, with no hardware changes in between our boards. I looked back at my notes and I have measured it at 1.15 MOhm for one NRF52832 and another NRF52832 measured 10.26 M! (my method for measuring the resistance was 1. confirming the resistances of my voltage divider (with a meter) with the SAADC input of NRF52 NOT connected to it, 2. then I applied a DC voltage at the high end of my divider, measured the voltage at the middle point with a Fluke multimeter (re-confirming my divider and my multimeter) 3.then I connected the NRF52 to measure the middle point. And as i explained above, different units give different measurements! Hence i posted this question. Are there possibly two different revs of the chip (NRF52832) with two different resistor values?

[Farhang]

And to me more clear i calculated back the internal resistance of the NRF52 from Fluke meter voltage measurements of the middle point of the divider. Here are the raw values of my circuit and measurements:

1. voltage divider resistances 1M Pull up and 68K pull down (both 1% tol)
2. voltage at the high end of the divider = 12.49V
3. voltage at the middle point (before connecting the NRF52) = 0.790
4. voltage at the middle point with NRF52 #1 = 0.753V! ( => Input resistance = 1.15M )
5. voltage at the middle point with NRF52 #2 = 0.789V ( => Input resistance = 10.26M )