Getting thermistor settings right



  • I have a random cheap thermistor attached to my hot end. It's not completely random, since I have access to the datasheet, but it's not one of the recognized types. So I'm trying to take the resistance-at-a-given-temperature table and find thermistor settings that give better results over the temperature range I care about (say 200-250 °C). Does the following make sense?

    The thermistor settings are: B, which is a thermistor beta value, and T, which is nominally the thermistor's resistance at 25 °C. So with two parameters, I can make the thermistor calculation come out right at two temperatures. It seems I have two options:

    • Make the thermistor calculation come out right at 25°C and 250°C
    • Make the thermistor calculation come out right at 200°C and 250°C and don't worry about the value at 25°C.

    If I take the first option, then I should just look up the resistance at 250°C and at 25°C and plug it into the formula b = Ln(R t1/ Rt2) / (1/T1 – 1/T2) given at (among other places) http://www.ametherm.com/thermistor/ntc-thermistor-beta. For my 3950 thermistor this comes out to 2563 K.

    If I take the second option, then I look up resistances at 200 and 250 °C and plug them both into the formula to get beta; but then I have to turn the formula around and calculate a value of R at 25°C that makes the 200°C and 250°C measurements come out right. Doable; I get 4238 K and 1121978 ohms (instead of the 100k I actually observe at 25 °C) . This has the annoying side effect of making room temperature read as about 80°C, but if the manufacturer's table can be believed, I should get the right temperature at 200 and 250°C, and something closer in between.

    Does this make sense? Should I expect ill effects from having room temperature read as so high?

    On the wiki, I see a B value for the "E3D hot end thermistor" of 4388, which (using their data sheet) is about right for the 200-250°C temperature range, but needs a different T value to work there. Am I confused, or is the value in the Wiki actually going to give people worse results than 3950 K or the 2806 K that I calculate given 25 °C and 250°C?



  • Anne

    As far as I understand it the Beta value gives you the range and is set to be as near to the truth at 2 figures and I think the std Chinese ones are quoted at 25 degrees and at 50 degrees:

    DC has a spreadsheet that will calculate it for you.

    The E3D one is 25/100 hence the reason the beta value is increased to give better accuracy at the 200-220 range at the cost of 100.

    You don't want to be changing the Resistance at all as that gives you the 25 Degrre point and the beta adjusts at what temp you want the accuracy, I used DC's sheet to calculate the the 3950 Chinese ones needed to be 4008 to get it about right at 100 degrees (I had it on my Bed).

    If you PM Me your E-Mail address I will forward the sheet to you unless of Course dave get here first.

    Doug

    ps I have just piped the figures into the Calculator and reckon the B value needs to be of the order of 4175 to get accuracy around the 220 degree heat point. HTH


  • administrators

    I suggest you use 25C and 220C, then the room temperature reading will be about right and the errors at typical PLA and ABS temperatures will be small. That's what I did to get the value of 4388 for the GT104-2 thermistor used by E3D hot ends.

    There is a calculator on the Heaters page of the online configuration tool at https://configurator.reprapfirmware.org/. For now you should leave the T3 and R3 boxes blank.



  • Ah, I see the problem - I was misreading my data sheet by a factor of ten. So the numbers made it look like there was a huge advantage to using two high temperatures; turns out it helps a little bit and only throws the room-temperature reading off by a few degrees.


Locked
 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.