Review on 🌡️ Pack of 5pcs Gikfun Waterproof DS18B20 Digital Thermal Probe Temperature Sensor for Arduino - EK1083 by Billy Berry

We have used them to monitor and control experimental aquariums. Several of these sensors hung in salt water for several months without any problems. There is only one negative - sometimes it gives false readings of -127C (which is likely to happen if the value overflows or the bits are all zeros). I don't think it's a malfunction as it's probably caused by electromagnetic fields or leakage currents from other devices. However, it would be nice to have a way to prevent this. Other researchers have observed the same thing. [cm. Edit at the end] We'll stick with those gauges though. They are reasonably priced and don't seem to have any other issues. If there are many bad values, we find the problematic devices and remove them. Otherwise the Arduino sketch will simply ignore the value of -127 C and use the latest reasonable value. Geek-only update: we're currently using up to 12 of these, some for about 8 months. I ran a calibration test with all 12 sensors in the same insulated water tank at around 20 degrees Celsius. The standard deviation between the sensors was 0.11 degrees. The three expensive HOBO recorders had a lower standard deviation (0.03) and read 0.35 degrees warmer, but I have no evidence that HOBO is actually more accurate. At 36°C the standard deviation for the Gikfun sensors was up to 0.17, but I can't state the temperature difference because the HOBO recorder is very slow and the water has cooled enough that the HOBO readings are unreliable. Edit June 2019 I found the source of most of our sensor errors. We power our Arduinos with a 12V power supply, which seems to be very noisy. The Arduino itself tolerates this perfectly, our problem was not obvious. By simply hooking up a large capacitor to the 12V wires before going into the Arduino, the errors were gone. The right filter with optimal capacitors and inductors would be even better.