New beta version of openXsensor (=openXvario)

Development & General Chat for the superb openxvario project.

Moderator: rainer

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Sun Apr 30, 2017 3:38 pm

I checked again the formula (based on documentation on the web) and it seems me OK.

Still I got some thermistors from aliexpress and I made a test.
My thermistors where given as 100k with Beta 3950.
When I measure the resistance with an ohmmeter I get
- 100 k at 25 °c (OK)
- 7 k at 100 °c (boiling water).
When I put 7 k as resistor in my oXs calculation (using a B_COEFFICIENT = 3950), I get a temperature that is much to low (53 °c).
I order to get 100 °c, I have to define a B_COEFFICIENT = 1700.

I presume that the thermistor available on ebay and aliexpress does not match the annonced specifications.

Can you try changing the B_COEFFICIENT in order to get the expected temperature (at about 120 degree).
Note: even if you change B_COEFFICIENT, you will still get a temperature = 25 when the sensor is at 25 °c (assuming that TEMPERATURE_NOMINAL = 25 and that the resistance of the thermistor at 25 °c = the value of TERMISTOR_NOMINAL)

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Sun Apr 30, 2017 4:04 pm

Thanks, i will try that later or tomorrow. These are the thermistors I bought

http://www.ebay.de/itm/3x-5x-10xThermis ... gJSou4RZXQ

The thing is, last night I did remember that my multimeter also has a temp sensor, so i stuck both my multimeter sensor and the sensor I bought from ebay to a piece of aluminum using a piece of self adhesive copper tape, and then used my heat gun to heat them up. Amazingly, the readings were very similar with the radio reading about 10 degrees higher at higher temperatures (over 100 degrees). But then again, i don't think I was running my engine at 150 degrees when the mixture was totally rich. That reading seems a bit too high. My temperature gun was reading way lower than that, giving me a bigger difference between its readings and the radio.
What I'm going also to do is to connect a pot to the oXs and compare the resistance with the readings given by the radio and the sensor datasheet. Then I will measure the resistance of the sensor when immersed in boiling water, and at different temperatures..
The problem is that I don't have a good reference so I can calibrate the sensor :(

Thank you for looking into it..

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Sun Apr 30, 2017 4:26 pm

FYI, I just made a test with the 5 thermistor I got.
Putting them at the same time in boiling water, I got some differences in the resistance (measured with an ohmmeter).
It varies from 6.7 k up to 7.k.
Keeping the same B_parameter, it would make a difference of about 2 °c at 100 °c.
So for best results, it is good to calibrate the thermistor.
If you can measure the resistor of the thermistor at 2 temperatures (e.g. 25 an 100) you can easily calculate the parameters to be used in oXs using following formula:
1/T1 = 1/T0 + (ln(R1/R0)/B_COEFFICIENT)
Where T are temperatures in °K (so °C + 273.15)
T0 = e.g (273.15 + 25)
T1 = e.g. (273.15 + 100)
R0 = resistor at e.g. 25
R1 = resistor at e.g. 100

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Sun Apr 30, 2017 8:01 pm

mstrens wrote:
Sun Apr 30, 2017 3:38 pm
I checked again the formula (based on documentation on the web) and it seems me OK.

Still I got some thermistors from aliexpress and I made a test.
My thermistors where given as 100k with Beta 3950.
When I measure the resistance with an ohmmeter I get
- 100 k at 25 °c (OK)
- 7 k at 100 °c (boiling water).
When I put 7 k as resistor in my oXs calculation (using a B_COEFFICIENT = 3950), I get a temperature that is much to low (53 °c).
I order to get 100 °c, I have to define a B_COEFFICIENT = 1700.

I presume that the thermistor available on ebay and aliexpress does not match the annonced specifications.

Can you try changing the B_COEFFICIENT in order to get the expected temperature (at about 120 degree).
Note: even if you change B_COEFFICIENT, you will still get a temperature = 25 when the sensor is at 25 °c (assuming that TEMPERATURE_NOMINAL = 25 and that the resistance of the thermistor at 25 °c = the value of TERMISTOR_NOMINAL)
Sorry but I made a mistake.
In order to check the formula used in oXs, I made an xls sheet to calculate the resistor at T1 based on B_COEFFICIENT
I just saw that there was a mistake in my xls sheet. I used login base 10 instead of log neperien.
When I fixed it, I find that to get a resistor of 7k at 100 °c, B_COEFFECIENT has to be about 3950. So the specifications of my thermistor are ok.
To morrow, I will try to make a real test with oXs putting the thermistor in boling water and see if oXs gives about 100°c.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Sun Apr 30, 2017 9:41 pm

Thanks.. That is a good idea. I will also do that, and will check the oXs (radio) and the resistance..
By the way, if you look at the link I posted above.. are your thermistors the same as mine?? Or are they different?

Thanks

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW


mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Sun Apr 30, 2017 9:48 pm


User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Sun Apr 30, 2017 10:02 pm

They look like the same as mine, but yours have some wire already soldered to them.. Mine have only the wies that come out of the little glass bulb. Apart from that they look the same

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Sun Apr 30, 2017 10:02 pm

Have you seen this?

João
Attachments
NTC-Thermistor_3950_Datenblatt.pdf
(203.2 KiB) Downloaded 7 times
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 6:43 am

I did not saw it, but it confirms that resistors should be about 7k at 100 °C

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 7:11 am

Yes, and that is the chart I used to chose the value of the resistor to VCC. 4K7 for my setup.

Looking at the chart I can also see the the resistance is not linear with the temperature..

João

Sent from my thor using Tapatalk

My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 9:46 am

I just made a test with one of my NTC connected to oXs.
At room temperature, I got 20 °c and in water that had boiled a few second ago (NTC being in a plastic bac) I got 95 °c
In a second test I got about 105 °c but I suspect that the resistance was to low due to some humidity.

So it seems me that oXs code is OK.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 9:59 am

I am just preparing another oXs with less wires :)
Will also test in a minute..
Will measure the resistance and see the values on the radio. not at the same time, of course :)

I think that compared with my multimeter that can also measure temperature, the difference starts to be evident above 100ºC
at 15 degrees it also reads lower than my multimeter..

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 11:16 am

Ok, I confirm your measurements. Please see the video

https://www.youtube.com/watch?v=jzJ1TTCDqUk

At 17 degrees, both measure the same.. At higher temps, the radio reads always 2 degrees higher than my multimeter.
Water started to boil, and oXs reads exactly 100 degrees C.
Resistance at 100 degrees is around 6.86 Kilo Ohm.

The strange is, at higher temps the oxs starts reading up to 10 degrees higher than the multimeter. Will try to make also a video on that. But, which one is wrong then? :(

I can't believe that my engine is running at 150 degrees even when the mixture is rich?? :o that is a bit too hot in my opinion..

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 11:59 am

You can try changing B_COEFFICIENT.
If you increase the value (by e.g. 100), you will probably get a lower value from oXs.
This change will normally has a higher effect for the high temperatures.

Better accuracy in the range 100 / 150 °c is possible if you can measure the resistor of NTC at about 100 and 150 °c.
Then you can thake the resistance at 100°C as the reference and you can calculate B_COEFFICIENT in order to match the measured resistance at the other temp (e.g. 150°C). Note: if the given temperature in range 100/150 will be more accurate, the value around 20 °C will become less accurate but this is probably not an issue.

It would also be possible to implement in oXs another way to calculate the temperature (Steinhart–Hart formula) in order to get good result over a wider range of temperature but it would require to know 3 parameters (A B C). The 3 parameters can be calculated if you know the resistances at 3 different temperatures over the whole range.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 12:05 pm

Thanks..
The problem is to find a "Stable" way to measure the higher temperature and take measurements.
Boiling water is wonderful for 100 degrees calibration. But how can I measure higher temps and know exactly the temperature? I don't have any high accuracy measuring equipment.

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 12:29 pm

I checked the datasheet of your thermistor.
In fact the B_COEFFICIENT is not constant over the whole range of temperature.
The value of 3950 is for the range 25/50
When I calculate it for the range 25/200, I get 4150

For best results over the range 100/150 you could apply following parameters (based only on your datasheet):
#define TERMISTOR_NOMINAL 6710 // nominal resistor of NTC (in Ohm)
#define TEMPERATURE_NOMINAL 100 // nominal temperature of NTC (in degree Celcius)
#define B_COEFFICIENT 4208 // B coefficient of NTC

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 12:39 pm

mstrens wrote:
Mon May 01, 2017 12:29 pm

#define TERMISTOR_NOMINAL 6710 // nominal resistor of NTC (in Ohm)
Thanks Mstrens, Shouldn't I use the resistance I've measured at 100 degrees? 6860 Ohm instead of the datasheet value?

By the way, this data might also apply for your thermistor, it seems to be the same as mine.

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 1:00 pm

If you use the resistance you measured, you should normally get a better accuracy because you avoid already the 1% tolerance of the manufacter.

In the mean time, I also calculated the A B C parameters based on your datatsheet for 25, 100 and 150 °c.
Those parameters could now immediately be recalculated if you can provide 3 sets of resistance/temperature.
If I implement the other formula with those 3 parameters, you would have no error at those 3 temperatures and the errors would be minimized over the whole range.

If you want I can provide a new code (it is about 10 lines to be replaced at a very specific place); so it is easy.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 1:06 pm

If you let me know where to put it, of course I can..

The problem is that I have no reference to measure it accurately :( :(
At lower temps I can compare it with my multimeter, as both seem to agree there.. The problem is the higher temperature.
I am just about to flash the test oXs with the values you posted above..

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 1:37 pm

Currently at the end of file oXs_voltage.cpp you have some thing like

Code: Select all

        float media =  SERIE_RESISTOR /  ( (1023000.0 / voltage ) - 1 ) ;
        
        float temperatura = media / TERMISTOR_NOMINAL;     // (R/Ro)
        temperatura = log(temperatura); // ln(R/Ro)
        temperatura /= B_COEFFICIENT;                   // 1/B * ln(R/Ro)
        temperatura += 1.0 / (TEMPERATURE_NOMINAL + 273.15); // + (1/To)
        temperatura = 1.0 / temperatura ;                 // Invert the value
        temperatura -= 273.15 ;                         // Convert it to Celsius
        voltage = temperatura ;
                
 
#if defined( DEBUG) && defined (DEBUGNTC)
You could insert a few lines in order to get

Code: Select all

        float media =  SERIE_RESISTOR /  ( (1023000.0 / voltage ) - 1 ) ;
        
        float temperatura = media / TERMISTOR_NOMINAL;     // (R/Ro)
        temperatura = log(temperatura); // ln(R/Ro)
        temperatura /= B_COEFFICIENT;                   // 1/B * ln(R/Ro)
        temperatura += 1.0 / (TEMPERATURE_NOMINAL + 273.15); // + (1/To)
        temperatura = 1.0 / temperatura ;                 // Invert the value
        temperatura -= 273.15 ;                         // Convert it to Celsius
        voltage = temperatura ;
                
        #define STEINHART_A 7.00111E-4   
        #define STEINHART_B 2.1644E-4
        #define STEINHART_C 1.0619E-07
        // T = 1/(A + B* ln(R) + C * ln(R) *ln(R) *ln(R)) 
        float steinhart = log(media) ;
        steinhart = 1 / ( STEINHART_A + STEINHART_B * steinhart + STEINHART_C * steinhart * steinhart * steinhart) ;
        steinhart -= 273.15 ;
        voltage = steinhart ;    

#if defined( DEBUG) && defined (DEBUGNTC)
The values for A, B and C are currently calculated based on the theoretical values from the datasheet.
I did not test the code but at least it compiles OK.
If it works, I could remove the lines about temperatura and adapt the documentation.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 1:46 pm

Would i have to comment something??

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 1:50 pm

No need to comment something. New code overwrite the old one. I could cleant it up later.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 1:53 pm

Ok, I just did it and it compiled

But what about the other settings in the config file that i just changed as per your suggestion? Can they remain like that?

Thanks

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 1:57 pm

If later you want to get A B C parameters for your device instead of using the data from the datasheet, you could make 3 sets of measurement.
- put the NTC in water at normal temperature and measure resistance with multimeter and temperature with a normal thermometer.
- put the NTC in boiling water and measure resistance with multimeter and assume that temperature = 100 ° C (or less in altitude)
- put the NTC and the temperature probe of your multimeter on an aluminium plate (ir order to increase inertia) and put in the oven in the kitchen with the thermostat e.g. on 150 °c. Let temperature stabilise and measure resistance and temperature with your multimeter.

Based on the 3 sets of data, A, B, C can be calculated as explained here:
https://en.wikipedia.org/wiki/Steinhart ... efficients

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 2:00 pm

Ok, will have a look at all this later or tomorrow.. Now i only have time to test the code with the other changes you've posted..

Thank you

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 2:01 pm

jhsa wrote:
Mon May 01, 2017 1:53 pm
Ok, I just did it and it compiled

But what about the other settings in the config file that i just changed as per your suggestion? Can they remain like that?

Thanks

João
The 3 settings about
#define TERMISTOR_NOMINAL 100000 // nominal resistor of NTC (in Ohm)
#define TEMPERATURE_NOMINAL 25 // nominal temperature of NTC (in degree Celcius)
#define B_COEFFICIENT 3950 // B coefficient of NTC
are just discarded by the new formula.
You still must keep them as long I did not remove the old code.

The other settings (e.g.)
#define FIRST_NTC_ON_VOLT_NR 3 // uncomment this line when thermistor are used; specify index of first voltage being used for conversion to temperature (e.g. 5 means VOLT_5)
#define LAST_NTC_ON_VOLT_NR 4 // specify index of last voltage being used for conversion to temperature (e.g. 6 means VOLT_6)
#define SERIE_RESISTOR 4700 // resitance connected to Arduino Vcc (in Ohm)
are still used.

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 2:12 pm

Ahh, ok, thanks.. All can remain as I had it then..

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 2:40 pm

Ok, I have just done the test with your first suggestion of changing the coeficient number..

The resuslts are

Mult --- Radio

19 ____ 23
49 ____ 53
65 ____ 67
99 ____ 100

As you predicted the lower temperature was not right anymore, but at 100 degrees both ther multimeter and the radio matched..

Will try your latest suggestion. i think I have time for that.. But only until 100ºC

Thanks

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

User avatar
jhsa
Posts: 16441
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: New beta version of openXsensor (=openXvario)

Post by jhsa » Mon May 01, 2017 3:16 pm

Ok, here is some interesting data for you with the changes you did to the code..

Mult --- Radio

19 ____ 18
22 ____ 23
29 ____ 31
38 ____ 41
49 ____ 50
59 ____ 61
69 ____ 71
80 ____ 82
89 ____ 90
95 ____ 95
99 ____ 99

The radio never reached 100 this time, just like the multimeter. It looks like it is getting more accurate as the temperature get closer to those key points, and a couple degrees difference in between. This might mean that as the temperature approaches 150 degrees, the reading would also be more accurate? That would be fine..
Sometimes I saw 3 or 4 degrees difference, but that could also be because one sensor might react faster than the other.. The radio was always reading higher than the multimeter though..
Form about 90 degrees untill 99 both were reading the same. perhaps later I will try to put them again on an aluminum plate, and heat it up with the heat gun, then check the readings near 150 degrees..
I hope this information helps

By the way, the NTC from my multimeter seems to be a very low resistance one.. About 15 ohms at room temperature, and then it goes up with higher temps..

Thanks

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1056
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: New beta version of openXsensor (=openXvario)

Post by mstrens » Mon May 01, 2017 4:06 pm

As long as your NTC fit the datasheet, with the new formula you should get 150 °C when temperature is 150 °C because the parameters are just just calculated in order to get exactly 25, 100 and 150 when the resistors are equal to the values given by the datasheet.

Errors at 25, 100 and 150 should be small and generated by tolerances on NTC, on resistor to Vcc and on aduino ADC.
Errors between 25/100 and 100/150 can still exist due to way of interpolation but it should be quite low.

I expect that your multimeter does not use a NTC but another component as resistance goes up.

Post Reply

Return to “OpenXVario - an open source vario supported by the open source firmwares!!”

Who is online

Users browsing this forum: No registered users and 2 guests