Hello guys, Its probably a noobie question but I have no idea how to go on about this issue. So Im trying to use the ADC of the RPi pico but its too noisy. In the application notes its recommended to use a 3V shunt reference for a stable reference voltage.
Im using the analog read with a potentiometer which works with 3.3V same as the logic voltage of the board.
The question is if I measure voltage coming from the 3.3V potentiometer and compare it against 3V Vref will i lose a portion from the effective range of the potentiometer?
What would you suggest to prevent that? I thought may be i can use a Schottky diode on the Vin side of the pot so i can reduce the max voltage from 3.3V to somewhere 2.9V or lower. Or use a voltage divider on the wiper output to drop the voltage from 3.3V to 2.9V or lower.
Any help or insight is appreciated
Yes, of course, but whether it is safe for the hardware to proceed with your idea is a question best answered by the data sheet for the processor used in the RPi Pico.
Have you tried a capacitor at the analog input to ground? I would suggest a 0.1µF capacitor as close to the analog input of the board and the ground of the board.
it is noted that input voltage for any of the pins should be equal or less than Vdd IO which is 3.3V and it doesnt mention anything about vref.
yes, seems like atmega chips do way better job at least they are stable. i also tried summing 5 readings and taking an average but it didnt really help. they are aware of it thats why they mention it in their datasheet and suggesting to use a shunt reference of 3V
You would probably be better off using an external ADC module. The ADS1115 is slow, but has much higher resolution and accuracy than the ADCs built into most, if not all MCUs. There are many other choices as well.
For better advice, tell us what you really want to measure, and at what resolution, voltage range and sample rate. If it is just the pot, there is hardly any point in doing any more than you have already.
I agree it shouldnt be complicated. I just dont want to lose any effective range from the pot because its already has only 60 degrees effective range which is just at the limit of what i need and im looking for a stable reading at least like the one i get from an atmega chip like 328p.
May be dividing the voltage on output side of the pot with resistors and feeding it to the analog input pin would work out for me. Normally pot output(wiper voltage) should be about 3.3V max when its at 0 degrees but if I use a divider network to reduce that to 2.9V i can stay in range of 3V ADC. Would that work?
No bad idea, a diode will have a slope and this will mess up any measurements.
Last time I looked at the A/D on the Pi (August 2021) there was a newly discovered issue which involved bursts of noise being generated. It may have been solved by now, or at least a recommended workaround suggested. I recommend you have a look at the Raspberry Pi forum. There is a section for Pico related stuff.
With a common Arduino like the Uno/Mega/Nano you would not use a stable external Aref for a pot, unless you also power the pot from that same reference voltage.
If you power the pot from one voltage and power Aref from another (stable or not), then you loose ratiometric behaviour (error cancelation) and you are rewarded with potential instability.
I assume it's the same for a RPi pico, but not many here are using them.
It's the same for an ADS1115. Bad idea to use them for a pot, because you don't have access to it's Aref voltage. You could mitigate that by also measuring supply voltage of the pot and use that result as a correction factor in the code.
For you pico I would suggest you use some smoothing code to cancel out the noise, nothing else.
Leo..
Power the pot from the reference voltage instead of 3V3.
Can you do that with a RPi pico?
You can't shouldn't do that with an Uno/Nano/Mega.
Leo..
Why not?
Somewhere in the Atmega328 datasheet is AFAIK a warning not to load Aref.
It should only be used as an input, to connect an external reference to if needed.
You can of course measure Aref with a DMM, but that is not a significant load (10Megohm).
Leo..
IIRC you shouldn't load internal reference because it is high impedance and it will ruin its stability.
OP wants to use an external shunt reference which is low impedance. With a constant load (the pot) I can see no issue.
Aref is AFAIK connected internally to 5volt with a mosfet switch.
Connecting a load to Aref will make a voltage divider with Ri of that switch (bad).
Talking about the Atmega328 though. Know nothing about the RPi pico.
Consider my advice useless for that board.
Leo..
(For ATMega328 based Arduinos:)
Only when you use Vcc as the reference. When you use INTERNAL it is connected to the on-chip bandgap reference. When you use EXTERNAL (default for the bare chip; on Arduino until you call first analogRead) the pin is input and you provide external reference here.
For completeness you can apply higher voltage than reference to the analog input; it must stay between Vcc and GND. I expect it is the same for most microcontrollers.
Are you sure noise is from reference and not your input? ADC input are typically high impedance inputs and take noise very easily from environment electric fields. Not one or two times i have seen battery operated device with ADC to pickup mains hum.
Can you power the potentiometer from the 3V reference, not the 3.3V supply?
@Smajdalf @MarkT @Vehasmaa @Wawa @Grumpy_Mike
Thanks for all the help guys. I think it is getting more complicated than it should be. I will try to offset or compensate it from code somehow or take a bigger sample to smooth it out.
How do you know this? What is the ADC's specification, are you achieving anything close to the specification? What noise are you observing?
Is your circuit arrangement in accordance with the manufacturer's suggestions?
Look at the basics first.