Hello, I have a plan that I think should work, but I'd like to check before going further.
I am repurposing an existing device as a PC button box using a 5V Pro Micro. This device has a 265 Ohm potentiometer that I would like to connect to an analog input. I could just use this as is, but it will draw 19mA continuously. This isn't really a problem, but if I can reduce it without ill effect I'd like to do that if possible.
If I place a resistor before the potentiometer this will reduce the current flow, but this will also reduce the range of voltages measured at the output of the pot. What I'd like to do is connect the AREF pin after the fixed resistor, so the range of measured voltage is between the top of the pot and the wiper.
If I were to do this with a 1K resistor as in the drawing this would give a voltage range of around 0 to 1 volt from the potentiometer. Mapped across 1024 possible analogread values this would be just under 1mV per value.
Is there a limit on the smallest voltage differences the analog pins can register provided the AREF pin is fed an appropriately low voltage? At what point would I run into noise issues?
I'm aware that the AREF pin isn't exposed on the Pro Micro, I'd plan to solder a wire to it's capacitor. Using a different value pot isn't really possible, this project is more a case of adding the Arduino to the device rather than adding components to the Arduino.
There is no clear limit but I would advice against going considerably below 1V with AREF because the noise issues will increase with lowered reference voltages.
You can eliminate much of that noise by using the internal noise cancelling features (which are not part of the Arduino library interface) but you have to measure in the actual circuit to see how efficient they will be. Nobody will be able to give you a fixed value from where you will have too much noise as that depends on many factors, many specific to the actual circuit.
It is on the Micro, but not on the Pro Micro you're using.
Ah, yes, thanks for reminding me about that. That would certainly work and it's close enough to what the AREF would have been at anyway that I'm only losing 5% of the measurement range.
To my understanding the AREF method should be a bit better in that the pot readings would be blind to any variations in the value of R1, but it's not something that's actually going to vary anyway as it'll be a fixed resistor not subject to any real changes in power dissipation or ambient temperature.
Lets say I wanted to economise and use the same resistor to limit current to something else that may be an intermittent load, (meaning the voltage across it would change) is there a downside to using the AREF pin rather than the internal reference, aside from needing to add the wire for it?
Thanks, I don't anticipate a need to go much below 1V anyway. The circuit will be very simple, it's just interfacing some switches and knobs to the arduino so they can be passed to the PC as a joystick/button box. It'll be the potentiometer, a 12 position rotary switch that I'll probably set up as a voltage divider into another analog input, and some toggle switches that I may connect as a matrix or just to individual pins.
It'll sit on my computer desk around the typical other devices, external HDDs, monitors, switch mode power supplies etc. It's built into an aluminium box.
Oops. Looks like the ATmega32U4 INTERNAL reference is 2.56V, not 1.1 like on the ATmega328P (UNO, Nano).
The ATmeg32U4 datasheet says the minimum Aref value is 2.56V which does not make sense since the internal "2.56V" reference (Vint) may be as low a 2.4V. Even stranger, the chart of Vint vs Vcc shows that it is always BELOW 2.54V! With a 5V supply at 25°C the Vint is about 2.51V.
Interesting... When it says "minimum" does that normally imply it'll be damaged if connected to something less while in use, or just that it won't work properly?
I could test it and see what it actually works at if it's not likely to be dangerous to it.
"Absolute Minimum" would imply damage might occur.
"Minimum" usually implies that it won't always work correctly (like less accuracy or more noise than rated).
I hope you realise that they must be powered from the same Aref voltage, to preserve ratiometric behaviour. Powering a pot or resistive devider from a different source/voltage than Aref is asking for instability.
Leo..
Thank you Leo. I don't really understand why it's necessary, but I have modified my schematic to have one current limiting resistor feeding both AREF, the voltage divider, and the pot.
Is there any reason I shouldn't take one of my voltage divider positions from the 5V rail? It would allow me to use fewer resistors in the divider, which would give wider value gaps between them. It would mean connecting 5V to an analog pin while the AREF would be seeing around 1V, but according to other topics here this is OK, provided I set the reference to external before calling analogRead.
