Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 6 of 6

Thread: Ambient Temperature Rating of Teensy 4.0 and 4.1?

  1. #1
    Member
    Join Date
    Jun 2015
    Location
    Palm Springs, CA
    Posts
    55

    Ambient Temperature Rating of Teensy 4.0 and 4.1?

    I was wondering if these calcs are accurate:

    3.3V * 100mA (@600Mhz) = 0.33 Watts of heat

    Junction-to-Ambient, no heat sink, natural convection: 40.8 C/W

    0.33 W * 40.8 C/W = 13.5 C temperature difference

    max junction: 95C

    95C - 13.5C = 81.5C = 178.7F without heat sink, natural convection

    But junction-to-case is listed at 16.8 C/W, meaning case to air is around 24 C/W which seems a bit too optimistic because that's in the range of an actual heat sink. Maybe the copper traces on the bottom dissipate heat as well as pins.

    But I also saw somebody measure 83mA in another thread, so that would improve heat tolerance.

  2. #2
    Member
    Join Date
    Jun 2015
    Location
    Palm Springs, CA
    Posts
    55
    https://www.pjrc.com/teensy/datasheets.html

    https://www.sparkfun.com/products/16...lock%20scaling.

    This 1.1" x 1.1" x 0.6" heat sink appears to be 9 C/W with no airflow:
    https://www.jameco.com/Jameco/Produc...DS/2277871.pdf
    If I cut that into a quarter, then it's about 36 C/W, although that seems very high given the amount of fin area.

  3. #3
    Member
    Join Date
    Jun 2015
    Location
    Palm Springs, CA
    Posts
    55
    Replying to myself again...a quarter of that heatsink is likely lower than 36 C/W, maybe 30 or high 20's because all 4 sides are exposed to fresh air. I'll see if I can find a commercial product that fits...

  4. #4
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    26,573
    Quote Originally Posted by NewLinuxFan View Post
    Junction-to-Ambient, no heat sink, natural convection: 40.8 C/W
    Junction to ambient "thermal resistance" is not a rigid specification. It's an empirical measurement. Often semiconductor manufacturers will give this sort of info with a footnote about the specific test conditions. Some are more honest than others, who perform the test using an utterly unrealistic PCB having huge copper area without solder mask. Sometimes that footnote about test conditions gets pushed to an app note or just omitted completely. But you really can't expect a single authoritative junction to ambient spec. It's just not realistic. Thermal performance depends on the surrounding environment.

    The chip has an internal temperature sensor you can use to measure the actual on-die temperature. Using that sensor to monitor the actual temperature in your application is the proper way.

    If the temperature is getting too high, the first step to consider is underclocking to 528 MHz (use the Tools > CPU Speed menu). The CPU voltage used is lower for 528 MHz, so you get a pretty substantial improvement without a huge loss in performance.

  5. #5
    Member
    Join Date
    Jun 2015
    Location
    Palm Springs, CA
    Posts
    55
    Thanks! Seems like the snooze library can be useful too.

  6. #6
    Maybe I am missing something. Using Serial.printf("%.2fC\n", tempmonGetTemp()) to monitor the T4.1 I get Tjunction = 74degrees C. It is a pretty full application using two TDM interfaces, a TFT display and a few leds and pushbuttons. If I assume a very conservative 50 C/W that would mean my power consumption was about 1 watt. That would mean the T4.1 is using about 300 mA of current.

    What am I missing regarding the assumption that the T4.x consumes about 100 mA?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •