Temp Sensors

Hi,

Have used NTCs as off board aqaurium \creature\plant temp sensors to micros and the DS18B20 seems to be a good digital alternative, but wonder if they are as reliable in terms of total failure?

Being a much more active device does the DS18B20 seem to be as long lasting as a ntc can be ?

Would be interested to hear other folk views; partic from any professional users as its noticable many commercial temp based devices still seem to use a ntc.

Also while a genuine Dallas 18B20 purchased from one of the major suppliers should be first quaility and give the best service life , many folk including myself , seem to buy the far east steel encapsuled types with ready made lead attached, so again wonder if they really contain a genuine 18B20 or possibly low spec devices ?

I currently have two applications in my home where I make use of the DS1820 with some success.

Note! I said SOME SUCCESS. I found that the value's returned from the DS1820 sometimes either goes to full max (85) or as if I loose the signal and then the value returned is "-127"
In my Arduino Nano code I simply now compare if the RAW value received from the DS1820 goes either above 80 or below -100. In either those cases I know the value I'm getting is incorrect, and I simply do not update my Temperature register that I use in further down in my code.

I have one DS1820 submerged in my solar water heater system. That one gave lots of error readings the closer it got to the maximum measuring range of 85. In a bright sunny South African day the water often goes greater than 85DegC.

The other application I use three DS1820 sensors to measure the temperature of my swimming pool solar water panels. Once the temp goes over a set-point limit (55DegC) I switch on the pool pump. Then I monitor the DeltaT between the solar Panels inlet line an outlet line. If the DeltaT is less than 0.5DegC I switch the pump off again. In this application too I found that the three probes randomly gives error readings. Again, by just doing some software I monitor the RAW data and only use the data if it is withing my usable limits.

I have some DS1820 sensors in a piece of equipment that has been running 24/7 for over 18 years with no problems at all.

I've designed National and Dallas temp sensors into company products, and have not noticed any reliability issues with either. But I use them only where a digital output is required. For out of the box accuracy, nothing beats the o.1C interchangeability of a good NTC thermistor.

ak