Strange TEMPMON behaviour

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Strange TEMPMON behaviour

Jump to solution
1,822 Views
Davidino
Contributor IV

Hi to all,

I loaded a simple code for monitoring ciclically the temperature junction in two different boards.
The first one is a MIMXRT1064-EVK board with PIMXRT1064DVL6A mounted, the second one is a custom board with MIMXRT1064DVJ6B.
The environmental temperature is 26/27C°. As soon as I turn them on, the temperature shown by EVK board is 18C° while the custom one is 39C°. Clearly both the measured temperatures are strange.

How is it possible that the EVK board temperature is this much below the ambient temperature?

How is it possible that the temperature's gap is around 20C?

After 30 minutes, the EVK board temperature is 30C° while the custom board is 50C°. They keep the difference of 20C°. Besides I've measured the superficial temperature with thermal camera and both ICs are around 35C°.

Thank you.

0 Kudos
Reply
1 Solution
1,411 Views
Davidino
Contributor IV

I report the answer that I got from the ticket:

the accuracy of TEMPMON is about +/-15°C. In addition to that on some parts the temperature monitor calibration fuses have not been set.

Factory has already been notified and they will do more test to improve this sensor.

Therefore, we recommend that you do not use it at this time.

View solution in original post

0 Kudos
Reply
11 Replies
1,412 Views
Davidino
Contributor IV

I report the answer that I got from the ticket:

the accuracy of TEMPMON is about +/-15°C. In addition to that on some parts the temperature monitor calibration fuses have not been set.

Factory has already been notified and they will do more test to improve this sensor.

Therefore, we recommend that you do not use it at this time.

0 Kudos
Reply
1,614 Views
Davidino
Contributor IV

Hi @_Leo_ ,

I write you again since I didn't receive any reply in the last week. I'd like to know why our demo board just turned on reads a temperature of 10°C below the room temperature, and on your side, the same board just turned reads a temperature 10°C above the room temperature (supposing that you left it off enough time for cooling off).
Thanks

0 Kudos
Reply
1,516 Views
_Leo_
NXP TechSupport
NXP TechSupport

Hi @Davidino,

Please excuse the delay in my response.

Since September 29, I have internally requested the technical specifications of TEMPMON, since that is not found in the documentation, however I am still waiting for such information.

As soon as I have an answer I will let you know immediately.

Sorry for any inconvenience this may cause you.

0 Kudos
Reply
1,758 Views
Davidino
Contributor IV

Update:

We did a new test. We create a thermal insulated box with some power resistors inside. We could reach an environment temperature of 60°C.

In this new condition, as soon as we turn on the custom board we read a temperature of 80°C while the demo one starts with a temperature of 72°C. After half an hour, the custom board shows a temperature of 88 °C while the demo board shows 80°C.

The difference between the two boards in an high temperature environement is 8 degrees, while at room temperture was 20°C.

Have you got any explanations?

Thank you.

0 Kudos
Reply
1,751 Views
_Leo_
NXP TechSupport
NXP TechSupport

Thank you so much for your interest in our products and for using our community.

Just to clarify, by "...a simple code for monitoring cyclically the temperature junction..." you mean that you are using the SDK example evkmimxrt1064_temperaturemonitor?

If not, I recommend that you refer to the example, since the appropriate calibration is performed in the TEMPMON_Init function, which is explained in more detail in Chapter 19 Temperature Monitor (TEMPMON) of the reference manual.

Hope it helps you. Please let me know your results and if you have more questions about it.

Have a nice day!

0 Kudos
Reply
1,735 Views
Davidino
Contributor IV

Hello @_Leo_ ,

thank you for your answer. You can find the project that we used in attachment. It's the demo project temperaturemonitor with the exception that we ciclically print the current temperature, we disabled the temperature interrupt and the code part that changes the core frequency with the temperature.

Thank you.

Best Regards,

Davidino 

0 Kudos
Reply
1,713 Views
_Leo_
NXP TechSupport
NXP TechSupport

First of all, thank you for your patient, we are solving cases as soon as possible.

I have run your attached project and effectively get the same results as the SDK example, although unfortunately I do not have a thermal insulated box nor a thermal camera. What I get is a reading of about 42°C while the ambient temperature is 26/27°C.

Now, as it mentions in the reference manual the TEMPMON determine the on-die operational temperature, which, according to our application team, is approximately equal to the junction temperature. The latter, in operation, is higher than the temperature of the case and the temperature of the outside of the part.

Regarding the temperature difference between the custom and evaluation boards, have you taken a look at the processor's power consumption on your board? I want to make sure there isn't actually an issue that is causing them to draw more current which would also cause the die to heat.

And just as an extra fact, we have no report of any errata regarding the RT1064 TEMPMON.

Hope it help you to explain your results.

0 Kudos
Reply
1,708 Views
Davidino
Contributor IV

Hello @_Leo_ ,

thank you for your answer. I want to focus on the following two matters:

  1. In the first message I wrote "The environmental temperature is 26/27C°. As soon as I turn them on, the temperature shown by EVK board is 18C°...".
    Could you check on your side that, with the EVK demo board, which has been left off for 1 hour, as soon as you turn it on, the first two/three temperature readings are below the environment temperature? Why does it happen?
  2. Concerning the power consumption, since our custom board is significant different compare to the demo board, we cannot make comparison. Besides we cannot measure exclusively the processor consumption. Since the project loaded on the demo board is exactly the same loaded on the custom one, and the project basically uses only the tempmon module, can we assume that the processor power consumption is the same?

Thank you for your help.

Best Regards,

Davidino

0 Kudos
Reply
1,672 Views
_Leo_
NXP TechSupport
NXP TechSupport

Hi @Davidino,

Regarding the first matter my results with the same methodology that you are using are the following at an environmental temperature of 25/27C°.

nxf86756_0-1664299085113.jpeg

 

Once it reaches 40.2°C I let it run for approximately 20min and it reaches 42.7°C.

Still, let me check your case with our applications team, to see what comments they have about your reading of 18°C at an environmental temperature of 26/27C°.

Regarding the second matter, it cannot be assumed that even if it is the same example the processor power consumption is the same, it also highly depends on the implemented hardware.

 

0 Kudos
Reply
1,428 Views
Davidino
Contributor IV

Hello @_Leo_ ,

have you got any news?

Thank you.

 

0 Kudos
Reply
1,605 Views
Davidino
Contributor IV

Hello @_Leo_ ,

I write you since I haven't received any replies since last week. What are the reasons why our EVK board, just turned on, reads a temperature 10 °C less than the room temperature, while your EVK board, supposing that it was left off for enough time, reads a temperature of 10°C more than the room temperature as soon as you turn it on? I hardly believe that it get instantly hotter...

Please let me know. Thanks.

0 Kudos
Reply