Thermal constant of thermistor

[Simon]

I am looking at the following datasheet http://www.farnell.com/datasheets/1533980.pdf?_ga=1.210406930.646536181.1456469885 and I'm a bit confused by the specs (yes operating temperature is wider range than storage - yea right). But the thermal constant is hown at 75S, presumably they mean that it takes 75 seconds for an undisclosed temperature change to reflect in the value of the resistance...

[mikerj]

If it's a proper thermal time constant then it's an exponential function exactly like the time constant of an RC circuit i.e. after one time constant the output will be 63.2% of the final value, after two time constants 86.5% and after five time constants you are at 99.3%.

Thermistors drift at temperature extremes, so whilst it may operate accurately at +105C at beginning of life, it will undoubtedly drift if used at that temperature for sustained periods (you'll need to get reliability data from the manufacturer to check this).  The lower storage temperature range likely gives an almost indefinite shelf life.

[wraper]

and I'm a bit confused by the specs (yes operating temperature is wider range than storage - yea right).

You don't heat the whole of the sensor when measuring temperature. Also leads must remain solderable after the storage.

[Simon]

I'm staying well away from that manufacturer. The datasheet is very vague. I don't trust them.

[joseph nicholas]

I can´t imagine any thermistor would take that long to respond to a temperature change.

[Simon]

It is a bit long which is why I called bolocks on the company and moved on. It's an italian company and the datasheet sounds more like they don't want to get sued rather than provide actual functional information.

[mikerj]

I can´t imagine any thermistor would take that long to respond to a temperature change.

This one appears to be buried within a nice insulating plastic housing, so it's possible.

[joseph nicholas]

I can´t imagine any thermistor would take that long to respond to a temperature change.

This one appears to be buried within a nice insulating plastic housing, so it's possible.

The picture on the web site look like any thermistor.  Perhaps you got it wrong.

[Zero999]

It does say:

Classification according to protection against electric shock (sensitive element & cable) Basic insulation for 250 VAC

[Simon]

Yes it is insulated

[Zero999]

By the way, I'm sure you know this but others may not: basic insulation is not enough as a method of protection against shock, when connected to the mains, the thermistor needs an enclosure, which must be earthed, if it's conductive.

[Simon]

It's going to have all of 5V across it, I do need something that is sea water proof though so I'm actuaslly looking at a screw sensor as the metal will hopefully react at least as fast.

[wraper]

I can´t imagine any thermistor would take that long to respond to a temperature change.

As it is stated in air, I don't see anything unusual with it.

[mikerj]

I can´t imagine any thermistor would take that long to respond to a temperature change.

This one appears to be buried within a nice insulating plastic housing, so it's possible.

The picture on the web site look like any thermistor.  Perhaps you got it wrong.

Perhaps you are looking at completely the wrong part.

Both the drawing and the specification in the datasheet clearly show the thermistor is encapsulated in plastic, and this is a picture of the actual device from the Farnell website:

[Simon]

Yes I know, as i said I need something that will cope with sea air

[qno]

Hi,

It is easy to measure the time constant.
You have to use a temperature step.

Use 2 cups with water and a stop watch. One with cold water with ice cubes and one with boiling water.
Then you have a 0 and 100°C ref.

Dip the sensor into the cold water and wait until temperature does not change.
Take the sensor out of the ice water, start the stopwatch and put the sensor in the boiling water.
Stop the stopwatch when temperature readout passes 63.2 °C.

You have now the temperature constant in water.
Air has lower density and will take longer than water. (maybe you can estimate the time constant by difference in density of air and water)

Maybe you can come up with a test by putting the sensor in your freezer and outside temperature.
Calculate the temperature difference and measure the time it takes to heat up to 63.2% of the temperature difference.

[Simon]

Yes it would have to be an air test as I am specifically measuring the temperature of air so the reaction time in water is pointless. some datasheets do specify the reaction time in both.

[Zero999]

Presumably the air will be moving but the manufacturer of the thermistor will test it in still air, so  the specification on the datasheet will be pretty meaningless.

Does it have to be a thermistor? There are other types of temperature sensor which give a voltage or current dependant on the temperature.

There are also temperature sensors with a built-in ADC and give a digital output. Here's one which comes encapsulated on the end of a cable and is waterproof.
https://www.adafruit.com/product/381

[Simon]

Yes I will think the air would be moving as the sensor is to be placed near the outlet. The sensor from as a fruit is not recommended for sea environments as it could corrode. I'm also a bit worried about the communication protocol as is totally new to me.

[Zero999]

There seems to be a huge range of products which can do this. For example, this one does pressure too and uses CAN bus:
http://www.pmc1.com/Products/sts-dtmcan-pressure-and-temperature-transmitter/
http://www.pmc1.com/Customer-Content/www/Products/Files/DTM-CAN_B.pdf

[Simon]

Well I'm avoiding digital busses for the moment as the rest of the project is taking up enough time. I'm mainly after a good thermistor or other eay to use sensor that i can take an analogue reading from. The challenge is getting one already made for the environment. which is why I thought of the sensor type that is just potted into a screw.

[Jay_Diddy_B]

Hi,

Thermal time is analogous to an RC time constant.

Tau = RC

Tau (thermal) = Thermal Mass x Thermal Resistance

The sensor from Farnell, manufactured by Carel, is a cylinder of mostly plastic attached to wire. The sensor will have a thermal mass dominated by that amount of plastic.

The sensor has a time constant of 75 seconds in air (assume natural convection).

If you have forced air, the thermal resistance will be reduced by a 5x and the thermal time constant will be proportionally shorter.

This is independent of the manufacture, any similar sized blob of plastic will have a similar time constant.

A metal-cased sensor will have a higher thermal mass than a plastic sensor, and therefore a longer time constant.

The beta and the resistance at 25C define the resistance versus temperature curve.

Jay_Diddy_B

 

[Simon]

Yes I've not really thought of that although I think what also needs considering is the thermal conductivity of the material surrounding the thermistor. I don't know how conductive the plastic is so maybe the metal would actually heat quicker? It will probably be a case of getting one and trying it.

[Jay_Diddy_B]

Simon,

The thermal conductivity of the sensor, doesn't make a lot difference. The thermal resistance is going to be dominated the thermal resistance of the sensor to the ambient. The plastic is a lot better thermal conductor than still air.

Regards,

Jay_Diddy_B

[Simon]

yes but I was thinking metal versus plastic. would the metal heat faster than the platic as the heat has to move through the material.