AREF shows 0 volts at any configuration, whether internal 1.1V or Vcc

Hello, I already googled around about this issue and seems like nobody experience same issue, so I have to post in here:

I have atmega328p on breadboard, all working just fine, very simple setup all it does is running ADC in free-running mode constantly reading voltage on A0, no issues whether I run it using internal Vcc (which is 3.3V) or 1.1 bandgap.

Attached is the schematic, as you can see pin#20 (aref) has a capacitor recommended by datasheet for better noise immunity, I am trying to measure voltage at AREF and it shows zero, I tried everything, 3.3V Vcc, 1.1 bandgap, running ADC, stoppnig ADC, nothing bring voltage to AREF pin and if I understand correctly voltage sources are coupled to AREF pin so there should be a positive voltage there. What am I doing wrong?
schematic.JPG

Not sure I understand, are you saying that the readings from the ADC channels are as expected but the reference voltage does not show up on AREF?

In theory, the ADC uses the voltage on AREF when it is doing a conversion but perhaps if something is damaged then the (TQFP package) pin may not be connected to the chip. If you have a DMM with a diode test then that can be used to see if the pin is connected to the chip. There are two ESD diodes that connect to the power and ground. Testing for one of those diodes will tell if the pin is connecting to the chip inside the package.

AREF is an input. I'm not sure it would ever read any voltage as an output (despite the warnings about it getting shorted to the internal supply if misconfigured.)

AREF is the internal voltage reference, and there is a internal 16k resistor from AREF to GND.

@alexmg2
Is that pin floating ? Can you measure the 16k resistor ? (try to measure it when it is powered).
Are you using the normal Arduino sketch, or did you change the internal mux for external reference voltage ?

There are a few wicked tricks that can be done with the AREF pin.
It can be used as some kind of digital output, since it can alter between 1.1 and Vcc.
It can be used as some kind of analog input, since the AREF voltage can be measured with the Vcc as reference. But the voltage to be measured has to be above 1 V (it might start working at 0.6V or so).

westfw:
I'm not sure it would ever read any voltage...

It does when "external bypass" is part of the selected reference.

Somewhere on this forum I documented how that can be used to calibrate the 1.1 V reference.

IIRC the datasheet clearly states reference voltage may be measured on this pin with high impedance voltmeter but only capacitive load may be connected to this pin. In fact connecting external voltage to AREF pin while internal reference (either bandgap or Vcc) is active may damage the reference and must be avoided.

In theory, the ADC uses the voltage on AREF when it is doing a conversion..

IIRC the datasheet clearly states reference voltage may be measured on this pin with high impedance voltmeter but only capacitive load may be connected to this pin.

Measurements with NANO reveal the following observations:

1. analogReference(DEFAULT) is executed; analogRead(A0) is not executed.
Oscilloscope shows ~0V at AREF-pin of NANO.

2. analogReference(DEFAULT) is executed; analogRead(A0) is executed only once in setup() function.
Oscilloscope shows: ~+5V at AREF-pin of NANO.

3. analogReference(DEFAULT) is executed; analogRead(A0) is executed repeatedly in loop() function at 2-sec interval.
Oscilloscope shows: ~+5V (constant) at AREF-pin of NANO.

Thanks everyone for replies and ideas as well as bench demo, I did run few experiments tonight again and all seems working just fine, I rebuilt the setup from scratch using new chip, looks like two previous chips I was using both had issues with AREF pin as I have experimented alot with them and maybe broke it. This brand new chip worked like a champ and couples both 1.1 or Vcc to AREF pin.