Hi,
Can you post a copy of your circuit please.
CAD or picture of hand drawn circuit in jpg or png would be fine.
Can you give me some analog values for some weights that you aim to apply, go as high as 80 pound if you can, the notes specify that close to limit response becomes saturated.
Also, I dont have definitive weight measurements here at home. I'd need to wait to go back to school in a couple days.
What I can tell you is that if I press with little force I get a readout of about 650, and then it gets really hard to increase that number significantly with just my fingers
I can barely get it past 800 by squeezing very hard, as opposed to an automatic of 650 by barely pushing down on it and 210 from just touching it
I can respond with precise values in a few days if that is what you need
The A/D converter has a sample and hold capacitor on its input. If you have a high impedance source, which you will have with that sort of resistor then it will take a long time to charge from the value given by your pot to the value given by this high impedance potential divider. This is why the pot seems to affect the sensor's reading. There are things you can do to mitigate this like reading the same analogue input twice and only using the second reading. Put a delay between the two readings to allow it more time.
The normal value of the resistor should be the same as the half scale on your sensor so do some measurements.
And is there a way to fix no constant output?
No. This applies to this sensor and almost any low cost weighing solution. Have you noticed how your kitchen scales have a "tare" function for getting rid of that offset?
What is the value of the pot. If more than 10k, it needs a 100n cap from wiper to ground.
I would use the sensor from A-in to ground, and use a pull-up resistor.
Grounded things potentially have less problems. Adjust your code for inverted logic...
MEASURE the sensor with the average load you are going to use, and pick a pull-up resistor with that value.
If that's more than 10k, you have to do one of two things. #1 multiple readings in your code #2 10-100n capacitor from A-in to ground.
Ok, I'll add the LED resistor. The pot is 10k, so I dont think I need to add that.
"I would use the sensor from A-in to ground, and use a pull-up resistor.
Grounded things potentially have less problems. Adjust your code for inverted logic..."
Mind explaining that a bit more? Sorry, I'm sort of a beginner at this. Thanks
Thing is, it is not hash he is picking up. There is no noise pickup but a systematic increase where one reading affects the other. If you still use a 1M you still get the same problem.
Yes, I think I get that. The MUXing of the analogue inputs could move charge from one input to another.
That's why every input has to see a low impedance.
But long wiring with that high source impedance could add to that problem.
Leo..