This is the discussion:
DS18B20 not working on 3.3v? - Using Arduino / Sensors - Arduino Forum
I have opened the old topic and merged in the new one
Some people have had issues with fake DS18B20 on 3.3volt power.
But don't worry, you can power a DS18B20 from 5volt, and pull up to 3.3volt.
Do keep pull up current below 4mA (so 1k absolute minimum on 3.3volt-logic).
I am running a DS18B20 24/7 on an ESP32, with <10cm wiring, powered from the 5volt USB pin and pulled up with internal pull up of the ESP (no resistor). Works fine.
Leo..
I will try such approach but I think the pullup resistor will have to be something like 500-800 ohm with large network of 30 thermometers span over 50 meters. If I needed 1kohm pullup resistor for 5V logic than I presume I will have to go even below that with 3V3 logic. When I will perform tests I will post results here.
Can't do that, since the absolute max sink current of a DS18B20 is 4mA.
A lower pull up resistor value won't guarantee a LOW on the MCU pin.
If you have studied "(data) transmission lines", then you know termination is most important. Use CAT-5/6 wires for the long lengths, and use pull up or series resistors at both ends.
Leo..
I will make experiments but with 5V going down to 1K with pullup was improving signal rise and did NOT harm signal fall anyhow. I will check different pullup resistors.
Adding terminators will sink extra current down even more. I did not see echo on oscilloscope so I presume termination was NOT a problem.
But I did observe slower and slower signal rise with more and more thermometers. What is weird the slow signal rise was especially visible with beginning of transmission, next bits/bytes were much better than the first. However missing first bits was disabling transmission.
Using CAT-5/6 wires for such low frequencies is way overkill. Even standard alarm or telephone cables are ok with such low frequencies. But I will do experiments with different cables too. But I do not expect any improvement.