Thermal Response time for DS18B20 Temperature Sensor

[weststar42]

I just received the DS18B20, High Temp Waterproof DS18B20 Digital temperature sensor + extras -, ID: 642, and have it connected and working with an Arduino. The temperature readings are good and it seems to be working pretty well.

I've noticed that it takes multiple seconds for the probe's temperature to settle in. If I move it from a cold temperature bath to a hot temperature bath it might take 5 or 10 seconds for the readings, which update at less than 1 second, to achieve steady state temperature readings.

This makes me wonder if (1) I have the sensor working correctly and it happens to have a long thermal time constant or (2) if I could change a parameter to make it get to steady state readings faster. Both the OneWire library example and Dallas Temperature library example have the same long thermal response time.

Any experience with how fast the DS18B20 waterproof probe responds to temperature changes?

Thanks,

[adafruit_support_bill]

These do have a fairly long response time. And the one-wire read-times are relatively slow as well. For applications such as vessel temperature control, the temperatures are not changing very fast anyhow. If faster response is required, other sensor types should be considered.

[weststar42]

Thanks for the prompt feedback and confirming my experience. I recommend you add the long temperature time constant as something you list in the description and technical specs. I saw the 750 msec query time but nothing about taking 5 to 10 seconds to achieve steady state readings. Even the Maxim data sheet does not indicate that long.

[Franklin97355]

The problem is not the sensor alone. Any thermal mass (metal can, waterproof coating, thermal coupling compound) will slow down the response time. As Bill says, if you need faster response times you need to choose a different sensor but they all will have a settling time of their own.

[adafruit_support_bill]

There is no single number for that. The time it takes to reach steady-state readings depends also on the thermal conductivity of the material (gas, liquid or solid) that surrounds the sensor, in addition to any convection.

[spork99]

If you want faster stabilization/response of your temperature measurements, generally, you want the smallest possible temperature sensor. The less mass contained in the sensor body, the less mass that has to be heated or cooled to equilibrium.

In my experience, a few seconds is more than fast enough in most applications. When I've needed something faster, I've either used small thermistors or made really small bare wire thermocouples. Omega sells thermocouple wire as small as 0.001" dia, but it can be a bear to work with.

Of course, as others have mentioned, the specific conditions of heat transfer where you're making your measurement also have a huge effect. Keep in mind too that any thermal transients will result in uneven temperatures throughout whatever you're measuring.