Reply
Posts: 8
Registered: ‎11-02-2016

temperature sensor c8051f550

Hi all,

 

i am trying to  read out the temperature of a c8051f550 controller, using the 12 bits adc ,

 

I am using an internal voltage reference of 1.5 volts.

 

When I read the result of the conversio is around 2830 counts, which are 1.0366 (1.5 x (2830/4096))

 

When I apply the formula in the manual:

 

temp [ºC]= (vtemp - offset)/slope, with slope = 3.33mv/ºC and offset = 856 mV

 

i get,, 52ºC !!  which is definitely not the ambien temperature of the room, is this the inner temp. of the chip?

 dosen't feel that hot..

 

 

regards! 

 

 

Posts: 7,913
Registered: ‎08-13-2003

Re: temperature sensor c8051f550

yes, it is the inner temperature of the chip.

 

how much it is higher than ambient depend on I/O

erik
Posts: 269
Registered: ‎03-03-2015

Re: temperature sensor c8051f550

Hi Tommygun,

 

The offset of the temperature sensor (and to a lesser degree, the slope) will vary quite a bit from device to device.  I recommend calibrating the device by recording the temp sensor measurement at a known temperature, and adjusting from there based on the slope (store this calibration data in flash).  For more accurate measurement, you can calibrate at two points, and calculate the slope for each device.

 

Regards,

Joe

Posts: 7,913
Registered: ‎08-13-2003

Re: temperature sensor c8051f550

[ Edited ]

The offset of the temperature sensor (and to a lesser degree, the slope) will vary quite a bit from device to device.  I recommend calibrating the device after it has run a bit for the chip to reach equilbrium by recording the temp sensor measurement at a known temperature, and adjusting from there based on the slope

erik
Posts: 8
Registered: ‎11-02-2016

Re: temperature sensor c8051f550

Hi all,

 

 

Seems that there is a problem with the slope and/or the offset.

 

This morning When I arrived to the office I turned on the device and saw a similar reading ( around 50 ºC)

 

There must be around 24ºC in the office , so it is definitely not a "inner temperature" reading, I do not believe that the chip will increase to 50ºC in less than 5 seconds!

 

Looking in the manual:

 

 table.png

 

I'm confused because a corrected slope would be 3.41 mv/ºc which is far away from reality..

maybe I got one sample reaaally bad.

 

 

regards,

 

Victor

 

 

Posts: 298
Registered: ‎09-22-2009

Re: temperature sensor c8051f550

The temp sensor is typically quite variable from part to part for these devices. So a minimum of a one-point calibration is highly recommended when using the temp sensor, and two-point calibration should be used if you need more accuracy.

<a href="http://community.silabs.com/t5/Welcome-and-Announcements/Community-Ranking-System-and-Recognition-Program/m-p/140490#U140490"><font color="#000000"><font size="2">Hero</font></font> </a> jmg
Posts: 1,077
Registered: ‎04-27-2004

Re: temperature sensor c8051f550


I'm confused because a corrected slope would be 3.41 mv/ºc which is far away from reality..

maybe I got one sample reaaally bad. 


That table applies to Sensor alone, did you also check your 1.5v reference ?

 

eg I can get  a more sensible  (1.4 * (2830/4096)-0.856)/3.33m  = 33.4189ºC

Posts: 2,735
Registered: ‎02-07-2002

Re: temperature sensor c8051f550

[ Edited ]

Alan_S wrote:

The temp sensor is typically quite variable from part to part for these devices. So a minimum of a one-point calibration is highly recommended when using the temp sensor, and two-point calibration should be used if you need more accuracy.


It would help people realize this if SiLabs would provide the Min and Max values alongside the Typ values in the datasheet.

Posts: 2,735
Registered: ‎02-07-2002

Re: temperature sensor c8051f550


jmg wrote:

That table applies to Sensor alone, did you also check your 1.5v reference ?

 

eg I can get  a more sensible  (1.4 * (2830/4096)-0.856)/3.33m  = 33.4189ºC


But the datasheet specifies the reference to be between 1.45 and 1.55.

<a href="http://community.silabs.com/t5/Welcome-and-Announcements/Community-Ranking-System-and-Recognition-Program/m-p/140490#U140490"><font color="#000000"><font size="2">Hero</font></font> </a> jmg
Posts: 1,077
Registered: ‎04-27-2004

Re: temperature sensor c8051f550


vanmierlo wrote:

jmg wrote:

That table applies to Sensor alone, did you also check your 1.5v reference ?

 

eg I can get  a more sensible  (1.4 * (2830/4096)-0.856)/3.33m  = 33.4189ºC


But the datasheet specifies the reference to be between 1.45 and 1.55.


Sure, but when measurements do not match specs, measure everything, assume nothing !

 

If I do a 2 sigma run, it gives

 

(1.45 * (2830/4096)-0.856-0.014*2)/(3.33m+88u*2) = 33.60840ºC

Posts: 2,735
Registered: ‎02-07-2002

Re: temperature sensor c8051f550

And then you didn't even take ADC offset and gain errors into consideration.

Posts: 8
Registered: ‎11-02-2016

Re: temperature sensor c8051f550


vanmierlo wrote:

jmg wrote:

That table applies to Sensor alone, did you also check your 1.5v reference ?

 

eg I can get  a more sensible  (1.4 * (2830/4096)-0.856)/3.33m  = 33.4189ºC


But the datasheet specifies the reference to be between 1.45 and 1.55.


The 1.5V reference voltage seems ok: 1.49 volts..

 

Highlighted
Posts: 8
Registered: ‎11-02-2016

Re: temperature sensor c8051f550


vanmierlo wrote:

And then you didn't even take ADC offset and gain errors into consideration.


No I didn't, but I do measure other adc pin voltages, all with gain of 1, and the measures are ok..

 

It is just an issue with the temp. sensor