2

I recently purchased a 10k ohm NTC thermistor from a local electronics store for which I couldn't find a datasheet. By going through the post How to measure temperature using a NTC thermistor I understand how to calculate the resistance. My questions are:

  1. Can I use B constant for my 10k ohm NTC thermistor as 4050 or is there any other way to calculate the B constant?

  2. Will the B constant vary between different vendors?

Community
  • 1
tamil_innov
  • 251
  • 2
  • 9
  • 15
  • 2
    I just wouldn't buy a thermistor without it having a recognizable data sheet. Come to think of it I wouldn't buy any component without it having a recognized data sheet. – Andy aka Jul 16 '13 at 11:40
  • 1
    The term "B value" is hardly standard, so you should define your terms. Whenever I have seen something called a "B value" related to thermistors, it was never dimensionless, so you are definitely doing something unusual. – Olin Lathrop Jul 16 '13 at 12:50
  • 1
    What Olin is saying is that 'B' values are measured/quoted in Kelvin and should have a 'K' unit after the figure. – MikeJ-UK Jul 16 '13 at 13:46
  • 1
    The B value even varies by temperature for a single NTC. – starblue Jul 16 '13 at 20:06

2 Answers2

4

No, yes and yes, in that order.

B values vary a lot, typically between 3000 K and over 5000 K, not only between manufacturers but between different parts from the same manufacturer. Even apparently identical parts from one manufacturer will be subject to a tolerance on their B values.

You can measure the resistance of the thermistor at various temperatures and plot the log of resistance against reciprocal absolute temperature, ie ln(R) against 1/Tabs.

The slope of the best-fit straight line through the points will be an approximation of the B value (which as Olin notes is itself an approximation of the Steinhart-Hart cubic, itself an approximation ...) .

For the benefit of those who have not heard of the B (or \$\beta\$) parameter, there are explanations on Wikipedia and on this manufacturer's website (the first I found with a google search).

MikeJ-UK
  • 5,348
  • 15
  • 20
  • I know the question used "B value" without defining it, which is on the OP. But doing the same in your answer is now on you. I have never seen something in a thermistor datasheet called a "B value" that was dimensionless, so your answer lacks meaning without a proper definition. For all you know, you and the OP are talking about different "B values". At least make your answer correct by properly defining terms you use. – Olin Lathrop Jul 16 '13 at 12:54
  • @OlinLathrop - I'm totally astonished you don't know the concept. Rarely if ever have I seen a datasheet that *doesn't* mention B value. – MikeJ-UK Jul 16 '13 at 13:13
  • I know the concept. The point is your use of it. I see you now added a link, but your quoted values are still dimensionless, which is completely contrary to what you claim is supposed to be the self-understood "B value". Also, "B value" may be common in the relatively accurate high temperature metal thermistors, but the cheap low temperature can use a different approximating formula. – Olin Lathrop Jul 16 '13 at 13:24
  • @OlinLathrop - Better now? Olin, there is really no need to be so obtuse. If you notice something missing, just say so. And the B value is the first-order approximation, common for low accuracy measurements. For high accuracy the Steinhart-Hart (3rd order) is more appropriate. The third sentence in your last comment makes no sense, by the way. – MikeJ-UK Jul 16 '13 at 13:36
2

First, you need to define your terms. I'm guessing that "B" is one of the coefficients of the quadratic approximation of the thermistor resistance as a function of temperature.

Second, don't assume anything not in the datasheet. Don't guess, check.

The temperature to resistance function of thermistors can be quite complex and is dependent on the material used. There are various thermistor materials out there, and there can be differences in the exact formulation within the same basic types.

The quadratic approximation is just that, a approximation. Usually second order is good enough so that the remaining error is below the thermistor's inaccuracy anyway, but don't assume everyone uses this approximation. This usually works well enough for the high accuracy types, like platinum, but the much cheaper lower temperature thermistors used to sense power supply overheating and the like have a different function and are much less accurate. Often a exponential or a software lookup table is used. Again, you can't assume.

Olin Lathrop
  • 310,974
  • 36
  • 428
  • 915
  • In my 10K NTC thermistor it is written as 103.is it wise to serach like 10k ntc 103 datasheet? – tamil_innov Jul 17 '13 at 11:17
  • @tamil: Huh? Try asking in English. "103" is most likely a floating point value, like is used to label resistors and capacitors. The first two digits are the value and the third the power of ten to multiply the value by. – Olin Lathrop Jul 17 '13 at 13:17