Page 1 of 1

Temperature Measurement with the Sensor Rev. P

Posted: Mon May 18, 2020 2:12 am
by Fabian_B
I'm working with the Windsensor Rev. P5 and did some measurements in comparison with an ultrasonic reference anemometer. For calibration and temperature compensation i used the Method recommended in Paul Badgers Blog Post:


I've got some problemes with the temperature measurement. The temperature measured by the sensor is always aproximatly 20°C to high. Indoors the voltage reading of the tmp pin is 250 (25°C room temp.) and outdoors around 180 (8°C). According to this formula

Tambient = ( Vout*(5/1024) – 0.400 ) / 0.0195

the measured temperatures are 42°C and 24.5°C respectively. Nevertheless i think that the sensors are not broken. I scaled the measured windspeed of the Rev.P5 and the ultrasonic anemometer and plotted the two graphs. They match really good. Only the absolute windspeed value of the Rev.P5 differs, caused by false temperature compensation, a lot.

I'm powering the Sensors with a 5V Powerbank with a voltage regulator (11,1 V output) in between the sensor an the power source. An Arduino Uno is connected directly to the powerbank and the Arduino reads the voltage output from the sensor with analog Pins A0 to A3 (setup with two sensors). Arduino and Sensors are connected to the same ground by the way.

1. Is the Rev. P5 using another thermistor so that the formula above has changed?
2. Could there be another reason why the tmp reading is to high?
3. What is the exact purpuse of the potentiometer? I think its for calibrating the zero voltage but during testing in couldn't figure out how to use it properly.

Thank you for your help!
Kind regards,

Re: Temperature Measurement with the Sensor Rev. P

Posted: Mon May 18, 2020 10:10 am
by oz
The temp sensor is MCP9701A. Its output is not scaled at all.

Tambient = ( Vout*(5/1024) – 0.400 ) / 0.0195
should be correct if Vout is the read from the analog pin ADC on a 5 volt arduino.

Recheck pin connections?
Test your arduino ADC with a simple voltage divider to see if reads ~512

I'm baffled. We do test these parts 100% and check the temp readings, in a temp controlled wind tunnel, but it's also possible that someone forgot to glance at the temp. I've never had a bad temp sensor but I've certainly had a few that aren't soldered on correctly but they usually zero output when not connected. We use the same formula in our tester.

You could try to inspect and maybe touch up the leads on the temp sensor with just a bit of solder if you feel competent. Otherwise I can just replace the sensor for you if you send email to sales at modern device dt com.

Let me know what the temp sensor says on the top - you'll need a magnifier and the right lighting to see. In the old days they used to print on parts, the standard now is laser etch which is much harder to read, plus everything gets smaller because, you know, Moore's Law.

Should read "APxx" where the x's are numerals.

Regarding pot settings - we use these to calibrate wind sensors for piece-to-piece consistency. Leave the pots alone. We paint a little red dot on the pots where they are calibrated. If your sensor doesn't have a dot we did ship a few before we decided to get a marker for the calibration. The pot will change the curve a bit as it governs the temp of the sensing element.


Re: Temperature Measurement with the Sensor Rev. P

Posted: Thu May 21, 2020 8:40 am
by Fabian_B
Thank you for your help.

I think i found the problem. I powered the Arduino with 5V despite the advice from the Arduino Homepage that the Arduino could be unstable with 5V power supply. I changed the power supply to 11V and replaced the Arduino Uno with an Micro for downsizing reasons. After that step three out of five sensors worked. Then i checked the soldering of the two remaining sensors: the outputs weren't zero, however the sensors worked fine after resoldering. Now, all sensors are working with one to two degrees deveation.

Re: Temperature Measurement with the Sensor Rev. P

Posted: Fri May 22, 2020 9:26 am
by oz

Glad you got it working. About +- 1.5 degree variation max, is what I see in the shop too, when testing the sensors.