Ancestors

Toot

Written by Electronic Eel on 2024-05-20 at 10:27

Trying to understand a datasheet of a NTC, Murata NXFT15XH103FEAB021. They specify the beta/B value, very good, I understand how to calculate temperature / resistance with it.

But they also mention other B-Constants as "Reference Values" for different temperature ranges. I don't understand what they want to tell me with them. Can anyone explain?

https://www.murata.com/en-eu/products/productdetail?partno=NXFT15XH103FEAB021

[#]electronics #datasheet

=> View attached media

=> More informations about this toot | More toots from electronic_eel@treehouse.systems

Descendants

Written by F4GRX Sébastien on 2024-05-20 at 10:39

@electronic_eel actual values fitted on restricted temperature ranges?

=> More informations about this toot | More toots from f4grx@chaos.social

Written by Electronic Eel on 2024-05-20 at 10:57

@f4grx how would that work?

the B-constant of 3380K is given for 25 to 50°C with 1% tolerance.

when I extend that to 25 to 80°C, would the B-constant (and most probably the tolerance) then increase, like they show in their table?

If so, it wouldn't be a constant anymore. and if they can specify a value like this and not just say the tolerance increases, there must be some known and calculable part missing in the temperature/resistance formula. why don't they tell you that part then?

=> More informations about this toot | More toots from electronic_eel@treehouse.systems

Written by Raul on 2024-05-20 at 11:06

@electronic_eel @f4grx My guess again: the B is constant for a fixed temp range. It might simply not be the same constant across ranges that have different max values.

=> More informations about this toot | More toots from raulinbonn@treehouse.systems

Written by F4GRX Sébastien on 2024-05-20 at 11:09

@electronic_eel my hypothesis is that B is a parameter estimated by curve fitting. If the fitting range is changed, result of model (probably some kind of regression) changes. Et variations are not that large.

=> More informations about this toot | More toots from f4grx@chaos.social

Written by Raul on 2024-05-20 at 10:46

@electronic_eel I'm gonna guess chose the right B depending on the max temperature in your target application. So if your temp range is 25-50º , you should assume the firts B they list there at 3380K.

But if your target temp range is 25-100º, the B you should work with should be that last, largest one at 3455K?

=> More informations about this toot | More toots from raulinbonn@treehouse.systems

Written by miek on 2024-05-20 at 11:10

@electronic_eel The B-"constant" isn't really constant, it varies with temperature range because it comes from fitting an exponential function to the real characteristics.

I think what they're saying here is that between 25 to 50°C they guarantee the best fit within that tolerance and, beyond that, they give some typical values for reference but don't test or guarantee it.

=> More informations about this toot | More toots from miek@mastodon.social

Written by jaseg on 2024-05-20 at 11:22

@miek @electronic_eel That would be my reading of that too.

=> More informations about this toot | More toots from jaseg@chaos.social

Written by A.M. Rowsell on 2024-05-20 at 11:16

@electronic_eel Real devices aren't as linear or perfect as ideal, modelled devices.

=> More informations about this toot | More toots from amrowsell@social.sdf.org

Written by blackhornfr on 2024-05-20 at 21:52

@electronic_eel use the resistance/temperature values they give on the product page in order to calculate the parameters according to your needs: https://www.murata.com/products/productdata/8796836626462/NTHCG83.txt?1437969843000

=> More informations about this toot | More toots from blackhornfr@pouet.chapril.org

Written by littlewing on 2024-05-22 at 23:30

@electronic_eel How accurate do you want your temperature measurements to be? Use the beta that best matches your intended temperature range. Your results will be good within ±1C.

=> More informations about this toot | More toots from littlewing@indieweb.social

Written by Electronic Eel on 2024-05-23 at 07:05

@littlewing my thinking was that when they are able to give different beta values for different temp ranges, they must obviously have a better model of the temp/resistance curve than the regular beta-based formula. probably something including more terms.

when they already have that, why don't they publish it?

for me that would be extra accurracy over a broader temp range (i'm targeting 0-70°C with my current application) essentially for free. i'm already using a lookup table based approach to calculate the temp on the MCU, so using a more complex formula to create that table once during build would be easy.

just changing to a different beta value would sacrifice accurracy in one area.

=> More informations about this toot | More toots from electronic_eel@treehouse.systems