NTC THERMISTOR MF52-103 10Kohm

Moderators: grovkillen, Stuntteam, TD-er

Post Reply
Message
Author
planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

NTC THERMISTOR MF52-103 10Kohm

#1 Post by planedo79 » 30 May 2020, 13:07

Hi,

I am trying to get temperature value with Wemos D1 mini and Esp Easy Mega web UI. I'm using analog input and I get some value 996. How do i convert this to temperature value? I read few post about this, but all seems to be done with arduino IDE/ and arduino board. Thermistor deails:

-10bit
-10K ohm
-resistor used in circuit 150K ohm (few differences in post should it be 10K or 150K)
-Thermistor is NTC MF52-103
-Voltage 3.3v used in wemos
-Voltage between trermistor legs 0.297 Volts

what equation should I put to the formula text box?
Attachments
thermistor espeasy mega.jpg
thermistor espeasy mega.jpg (119.47 KiB) Viewed 18884 times

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: NTC THERMISTOR MF52-103 10Kohm

#2 Post by TD-er » 30 May 2020, 13:39

The thermistor itself is only 10k, and given the output value, I guess you have placed the thermistor between the 3v3 and the analog input.
This means the value will be higher, if the thermistor value becomes lower.

But the resistor you used is 150k, which is 15 times the value of the thermistor.
So you also reduce the resolution of your measurement by a factor of 15.

I suggest you start with a 10k resistor.
Whether you place the NTC between 3v3 and A0, or between GND and A0 is up to you.
You must calibrate the values yourself (or compute them)
The analog input plugin expects max. 2 calibration points.

So let's say your measured value becomes "997" (as it is now in your screenshot), at 20 degree C.
Then you enter 20 as the "new" value.
So point1 becomes 997 => 20

Then you find another point e.g 30 degree C.
For this you heat up the NTC to this point (validate with another thermometer) and see what the ADC outputs. (e.g. 900)

But first you should make the circuit a bit more sensitive, by lowering the resistor value and see if the ADC value changes when heating (or cooling) the NTC.
And keep in mind the analog input only does linear interpolation, so it will probably give an error outside the calibrated values, because the NTC does not have a linear response.

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#3 Post by planedo79 » 31 May 2020, 21:51

Thanks I Will try this.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: NTC THERMISTOR MF52-103 10Kohm

#4 Post by Wiki » 03 Jun 2020, 15:45

In the past dealing with NTC to get temperatures I used normal physics: One calibration point has been ice water (0°C), the second calibration point has been boiling water (100°C). You have to wait quite a while to get stable readings. Gave me a wide range and worked almost at the point.

Recommendation: Because the behaviour of NTCs is mostly nonlinear, I gave up and went over to BMP280 for normal temperature readings, for higher values (>80°C) I take a thermocouple MAX6675.

[edit]
and in installations where only two wires are present, I take a Dallas DS18B20 in parasite mode
[/edit]

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#5 Post by planedo79 » 06 Jun 2020, 09:08

Thanks for advice. I have few Thermistor laying around and thought I could use these. I used ds18b20 also, but it was some cheap version and I wasnt water proof as promissed this caused trouble and my wemos d1 mini had problems to wake from deep sleep /connection problem which dissapeared after removing the ds18b20 sensor. My goal is to use these Thermistors and built sensor by my own as to monitor my mcu controller box where there is lipo battery and as a protection if it has any warming issues ect.. I'll write info here as I have it done..

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#6 Post by planedo79 » 06 Jun 2020, 14:34

Hi,

this is what I did manage to get out. Bowl +ice cubes +water and approx 15min and I got 0 Celcius value, and after that boiling water for 10minutes.

Thanks for help!
Attachments
thermistor espeasy mega2.jpg
thermistor espeasy mega2.jpg (113.78 KiB) Viewed 18638 times

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: NTC THERMISTOR MF52-103 10Kohm

#7 Post by TD-er » 06 Jun 2020, 17:01

Roughly 7 steps per degree is not bad.
Just take into account the curve is not linear, so there will be an error.
You may want to plot the NTC formula between 0 and 100 degree Celcius to see how big the error will be.

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#8 Post by planedo79 » 06 Jun 2020, 17:26

Hi,

Can you explain this how it is done "plot the NTC formula between 0 and 100 degree Celcius " ?

I found equasion how to determine the actual celsius/temperature reading. Not good at physics, but I get this value... 19.35 ? I know this is for linear, and there can be errors...

I also measured the resistance at zero and 100 Celcius, and had these values.

100 Celcius 0.718 Kohm
0 Celcius 30.2 Kohm

If I look for the tempereture table, I found that the resistance values does not match.. there is about (table value 98.96 and multimeter reading is 29.8)

Here is picture about the equasion..
Attachments
temeprature table MF52 type.jpg
temeprature table MF52 type.jpg (174.46 KiB) Viewed 18623 times
is this right.jpg
is this right.jpg (114.18 KiB) Viewed 18627 times

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: NTC THERMISTOR MF52-103 10Kohm

#9 Post by Wiki » 06 Jun 2020, 21:12

Eeehm, having a look at the table:

-5°C = 42.8 kOhm
0°C = 98.96 kOhm ??????
+5°C = 25.58 kOhm

Following the physics of any NTC, thats not possible. I suppose, you don't have a problem but the guy who produced the table.

Your multimeter reading fits pretty good the overall behaviour of the thermistor following the table, just trust it.

And be aware: due to external influences you will never be able to get 100% matching readings. Using ADC for resistors with a device like the ESP8266 is and will be never more than an estimation (more or less rough) than an adequate reading. Believ me, take it as it is.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#10 Post by planedo79 » 06 Jun 2020, 21:29

I was careless when I didn’t notice that. A clear mistake in the table now when looking more closely. Since I am new to the hobby automation electronics working, I am quite satisfied with this result I got a thermistor. For me, a value of +/- 1 C would be enough. Is it even possible to get one with this sensor?
Actually, exact values are not my goal, but how things should work to get things done right. Before that, I shouldn't do expensive sensors. But good attention from you and thank you for that! :)

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: NTC THERMISTOR MF52-103 10Kohm

#11 Post by TD-er » 06 Jun 2020, 22:30

Depends on the range where you would like to have this accuracy.

If it is a narrow range, you may want to re-calibrate around that range.
If I look at that table, the resistance hardly changes when close to 100 degree C.
But if you calibrate its 'hot' point at around 30 degree, then I think the error will be a lot less severe.

What other resistor do you use? 10k?

planedo79
Normal user
Posts: 23
Joined: 10 Apr 2020, 12:22

Re: NTC THERMISTOR MF52-103 10Kohm

#12 Post by planedo79 » 06 Jun 2020, 22:57

Hi,

Well I thought my range would be between 0 to 35 Celcius. I red about this sensor it might work ok with lower temperatures, but not quite sure are we talking minus celcius or what that range is. I have wide range of resistors that I can use if needed. Not that far yet I could point something precise resistor value. In this project I used 10K, as You recommended it

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: NTC THERMISTOR MF52-103 10Kohm

#13 Post by TD-er » 06 Jun 2020, 23:51

Yep 10k is still fine.
But given the tolerances and the unknown resistors in the ESP module and the shown calibration values it could also have been 12k or 15k, which ever you had laying around.

I expect the 25 degree calibration value to be around 511.
So that would suggest your current value as in the picture was taken when over 30 degree.

Maybe you can calibrate it quite simple for 36 degree, by holding the NTC in your hand for a while.

Post Reply

Who is online

Users browsing this forum: No registered users and 17 guests