NHS3100 Temperature sensor not sensitive ?

[xiangxuejiao00]

Now ,I encounter a question .

when I put the 3100 from indoor temperature 25 into fridge  temperature  -17，and set the measurement interval 1 second ,the 3100 temperature value decline to fridge temperature -17 takes about 15 minutes .The detail of the data 

in the below .

I want to know :

1 . why the temperature changes so low ?

2. is this normal ?

3. is there any config in the project can short down the time ?

[driesmoors]

Hi,

Measuring every second is quite extreme. When the chip is active, it will generate some modest heat. This self-heating is innocuous, as typical measurement intervals are in the order of several minutes. With 1 second, I'm fairly sure that the self-heating plays a role. I'm confident that a measurement interval of, say, 10 seconds all but eliminates this effect.

When a freezer is set to -17C, it does not mean the inside temperature is at -17C, nor that it is constant. Variations of several degrees are very well possible. The positioning inside the freezer also plays a role. Was the freezer motor activated at some point? Was the temperature sensor touching something else? Humidity and subdivisons in the freezer also influence temperature behavior. And of course, the thermal resistance of the temperature logger and monitor is important: a demo PCB will adapt to temperature slower than a label.

Validating temperature accuracy is surprisingly hard for such a simple concept. We know, we have been there ;-)
Without a carefully set up test environment, it is hard to make decisions on accuracy. The temperature as measured by the NHS3100 is accurate to 0.5C or better at -17C, as specified in the datasheet.

Kind regards,
Dries.

[xiangxuejiao00]

thanks for your reply 

As you said ,I do the test again .I also have other questions to ask 

1.the temperature from 25 to 0 is so quick ,it takes about 1 minutes ,but why it takes more than 10 minutes from 0 to -17?

2.can you offer me the Thermal response curve of 3100?

As you recommend before ,if set the interval =1 S,the IC will self -heat and will influence the test result ,

so I do the experiment below 

Two demo 3100 :smileyshocked: one set interval =1S,and the other set interval =10 S,then put the 3100 into fridge ,so I get the result below 

from the temperature 22 to temperature 0，these two experiments are all take about 70 seconds ,and from 0 to -16，

they all take about more than 8minutes .

why this happens?

is there anything wrong ?

thanks 

[driesmoors]

Hi,

I don't have the thermal response curved of the NHS3100, I'm sorry. Per my knowledge, I suspect it will look similar to http://kentchemistry.com/images/links/matter/salolCC.gif - of course, with different scales, offsets, durations, tipping points, etcetera.

I also can't access the link to vechain.com you provided. Is it matching the image I linked you to? The exact curve is of course heavily influenced by your setup, by the substrate, by your handling and measuring. Per my understanding, such a graph is unavoidable as it just boils down to physics.

I'm sorry, I can't provide you with more information about this; I can only give you the guarantee the IC will measure accurately the temperature.

To increase your trust in this (checking is always better, so I fully understand you), you can have the results verified in a specialized lab, where they have the equipment and the knowledge and insights to test and compare, avoiding common pitfalls in the interpretation of the data.

Kind regards,
Dries.

[xiangxuejiao00]

thank you for your reply