i.MX7D Temperature Monitor accuracy

[hiroshi_kimizuk]

Hi,

The temperature monitor function implemented in the i.MX7D is being used in our products under Linux+Debien environment.
There exists some variations in temperature value among our products.

Is there any tolerance at the measured Temperature monitor value since I could not find any specification about this in the i.MX7D data sheet ?

And I could find the answer for the  case of i.MX6S as below;

 Solved: i.MX6S die temperature sensor accuracy - NXP Community

Is this answer also applicable to the case of i.MX7D ?   I mean that the +/- 10 deg  tolenrance exists on the Temperature monitor value ?

Thanks and Best Regards,

KIMIZUKA

[Yuri]

@hiroshi_kimizuk 
Hello,

   i.MX7 temperature sensor accuracy is practically the same as for i.MX6.
"i.MX6 Temperature Sensor Module"

https://www.nxp.com/docs/en/application-note/AN5215.pdf

Regards,
Yuri.

[hiroshi_kimizuk]

Hi, 

Thank you for your reply.

OK, the description of the "Table 2. Accuracy results" in the AN5215; https://www.nxp.com/docs/en/application-note/AN5215.pdf is more correct and accurate value rather than +/- 10 Deg.  Is my understanding correct ? 

However, I have one thing to be clarified. This time, I could find the "Temperature sensor accuracy (Tdelta)" specification is specified as +/- 3 Deg. at the Table 9 (page 24) in the i.MX7Dual Family of Applications Processors Datasheet( IMX7DCEC.pdf (nxp.com) ).  (Though I mentioned I could not find this specification last time, sorry.) 

Is this specification; +/-3 Deg. still valid ?  Or we should follow the evaluation results on the AN5215 ? 

Thanks and Best Regards,

KIMIZUKA

[Yuri]

@hiroshi_kimizuk 
Hello,

  I expect the accuracy of +/-3% in the Datasheet concerns the recent Linux BSP
drivers using.

Regards,
Yuri.

[hiroshi_kimizuk]

Hi, 

Thank you for your reply. 

But I'm llittle bit confused by your explanation about the Tdelta deviation made up with Linux BSP.

1) What is the exact point to reduce Tdelta accuracy value to +/- 3 deg which is contributed by Linux BSP ?

2) As for datasheet, the characterized prime performance should be described.   Otherwise, someone who does not use the device with Linux, can not design any more. 

3) In the application note(AN5215), it is described that this new equation is applied to the thermal driver of Linux BSP.  Is this evaluation also using the updated Linux BSP ?  

Thanks and Best Regards,

KIMIZUKA