DS18B20 Temperature Sensor Waterproof

We have been using these to monitor and control experimental aquaria. Several of these sensors hung in salt water for several months with no problems. There is just one negative - sometimes there are spurious readings of -127C (probably what you get when the value overflows or the bits are all zeros). I don't think this is exactly a malfunction, because it is probably caused by electromagnetic fields or current leakage from other equipment. Still, it would be nice to have a way to prevent this. Other researchers have seen the same thing. [see edit at the end] Still, we are sticking with these sensors. They are reasonably priced, and seem to have no other problems. If there are a lot of bad values we find the problem equipment and remove it. Otherwise we just have the Arduino sketch ignore the -127 C value and use the most recent reasonable one. Update for geeks only: we are up to 12 of these in use now, some for about 8 months. I did a calibration test with all 12 in the same insulated container of water at about 20 degrees C. The standard deviation among the sensors was 0.11 degrees. Three expensive HOBO loggers had a lower standard deviation (0.03) and indicated 0.35 degrees warmer, but I don't have any evidence that the HOBOs are actually more accurate. At 36 C, the standard deviation for the Gikfun sensors was up to 0.17, but I can't quote a temperature difference because the HOBO logger are very slow and the water was cooling enough to make the HOBO values unreliable. Edit, June 2019. I found the source of most of our sensor errors. We are powering our Arduinos with a 12 V power supply which is apparently very noisy. The Arduino itself tolerates this perfectly, we the problem wasn't obvious. Just putting a big capacitor across the 12 V wires before they go into the Arduino made the errors go away. A proper filter with optimal capacitors and an inductor would be even better.

These will super glue into any Tee connector with the right glue. Using Starbond Toughened Medium Thick CA. The super glue isn't needed for low pressure, but once over 20psi, I had to glue it in as it is installed in a system at 65psi. Temperature is accurate, though a little slow in responding. The whole metal sleeve needs to get to temp, or more than the 1/8" I have in the water stream, so it lags the water temperature by about 3-5 degrees, but once taken into account, or at temperature for a period of time, the temperature matches other thermometers including a Fluke thermocouple to within 1°F. The downside is conversion takes 1/8 second to 7/8 second depending on resolution. 10 bit is best, as 8 bit microcontrollers aren't so accurate beyond that unless in a critical application. This is due to the Dallas OneWire sensor and not the probe or cable attached to it. They are VERY high quality and professional supple cable with strain relieved weatherproof end at metal probe side. I crimped JST connectors to the wires simply for portability as I glued a few into Tees for RO monitoring, the lack of a connector is a bonus as it allows you to use what is needed. Out of box the wires are tinned and stiff enough to plug directly into a breadboard for a microcontroller. Using with both a PIC32 and Arduino and having zero issues.

Am using for temp sensors in home assistant on a Shelly add-on for a Shelly 1 plus. Have one probe outside, one inside, and another on my hot water heater. Easy and works, cheap, stable, accurate enough, and no battery to deal with.

One of the sensors had a very wild readout (off by +-20F), but I imagine it was a fluke. The sensors are accurate to a ~2 degree Fahrenheit, so you probably don't want them for super sensitive data collection but for an ambient room temperature sensor they work great. You will need to use Dallas OneWire protocol for these, but there are libraries available and if you read the docs it should be a non-issue. Docked one star for the bad sensor, but other than that they are good sensors that work well for a budget.

For me they were very easy to set up. I used the milesburton/DallasTemperature libary which requires the paulstoffregen/OneWire. The example from the library worked fine. My problem requires three sensors so it was nice to discover that all three can share the same Aruino/ESP pin. All three red wires to 5V all three black to ground and all three yellow to my data pin - makes wiring nice and easy. Note the data pin requires a pull up resistor 4k7 or 5k. Just one irrespective of how many devices. The in built pull up is not sufficient. Some users have reported issues if the power supply is noisy so if you get odd readings try adding a capacitor across the power input Each device has an 8 byte ID. The sample code from the library scans the devices and reports the IDs so you can hard code them for future use and know which sensor you are reading from. On my desk all five reported the same temperature. Of course they could all be off by the same amount but I took it as a good sign. They responded fairly quickly to being held in my hand so I believe the sensor is in good thermal contact with the case. Holding each unit in turn was how I figured out which address went with which sensor. They appear to be well made. Nice long cable and a nicely finished stainless steel case.

Delivered on time as a quantity of 5 as per the photo and not as described ‘pack of 3’. Setup using one of the five 4.7kohm pull-up resistors in conjunction with a pi pico 2W microcontroller. Straight forward to implement and readings obtained from the three temperature sensors I setup all work perfectly. I have not checked the accuracy against a typical glass thermometer , however, all readings were within 0.5degC of each other at ambient temperature of 17.5degC.

Got 3 of them with a cable about 1m long. I hooked one up to an arduino with LCD pannel and it works great. The other one if hooked up to a Raspberry pi 3 and also works great. All sensors read the same temperature which is good.

good and accurate. Good for monitoring the soiltemp in the greenhouse and chickenhouse. I used others with heatshrink before but after some time they are rotten.

Very pleased with these. The price seemed very good and they accurately measure the temperature. I've had them running for days on a number of occasions, with samples every minute, to plot temperature changes on a tiny little display.

Seems good I used two of them so far and they meet specs OK, Only issue is that there is no indication what the three conductors are for. One assumes that 1 is black and ground, 3 is red and + supply and the yellow is data.