Thermal tests and measurements (SBCX)

From DAVE Developer's Wiki
Jump to: navigation, search
Info Box
SBC-AXEL-02.png Applies to SBC AXEL


History

Version Date Notes
1.0.0 11th November 2016 First public release

Introduction

This document describes the tests that have been performed on SBCX board to characterize thermal behavior of some configurations. The results here presented can be used by integrators building systems upon SBCX board as a basis for comparison. In fact, as stated in the SBCX's Hardware Manual, systems integrators have always to perform thermal analysis on the application's definitive configuration. Generally this differs significantly with respect to the use cases here described, in terms of both hardware and software.

In the rest of the document:

  • Tj denotes the junction temperature of the iMX6 processor
  • Tamb denotes the ambient temperature of the space in which the DUT (Device Under Test) has been placed
  • ΔT denotes the difference between Tj and Tamb (one the steady-state has been reached).

Test beds

Five different test beds have been considered. From the electronics point of view, all of them are based on the following parts:

  • SBCX equipped with industrial iMX6 Dual Lite processor in plastic package (maximum frequency = 800 MHz) and 1GB SDRAM
  • LCD Ampire AM-800480SETMQW-TA1H.

Settings common to all test beds:

  • SBCX supply voltage: 12V
  • Linux governor: performance
  • LCD backlight intensity: 100% (in terms of power, this is about 1.6W)
  • still air flow
  • software load [1]
    • Iperf test on 100M ethernet link
    • StressAppTest with the following parameters -M 340 -s 86400 -C 4 -m 4 -W. For more details please refer to StressAppTest README
    • Write/read/verify 100MB binary file on Fat32 partition on uSD
    • Write/read/verify 100MB binary file on Fat32 partition on USB memory stick
    • QT Vector Deformation Demo

Test bed #1

  • Tamb = 22.5 °C
  • system layout and configuration
    • SBCX and LCD lie on a table horizontally (see also the following picture)
Test bed #1

Test bed #2

  • Tamb = 22.5 °C
  • system layout and configuration
    • SBCX and LCD are enclosed in a metal box that lies on a table horizontally (see also the following picture)
    • iMX6 processor is thermally coupled with an heat sink (Wakefield-Vette 624-60AB-T4E) through the provided adhesive tape
Test bed #2

Test bed #3

  • Tamb = 22.5 °C
  • system layout and configuration
    • SBCX and LCD are enclosed in a metal box (Stainless Steel AISI430) that is positioned vertically (see also the following picture)
    • iMX6 processor is thermally coupled with an heat sink (Wakefield-Vette 624-60AB-T4E) through the provided adhesive tape
Test bed #3 front view
Test bed #3 back view

Test bed #4

  • Tamb = 22.5 °C
  • system layout and configuration
    • SBCX and LCD are enclosed in a metal box that lies on a table horizontally (like test bed #2)
    • Axel Lite SOM is thermally coupled to an aluminium heat pipe
      • heat pipe dimensions are: 40x22x16.5mm [1]
    • on SOM side, heat pipe is placed on the iMX6 processor and top side SDRAM chips and it is coupled with 40x22x2.03mm gap filler (Hala Thermal TGF-FUS2030-SI)
      • in turn, heat pipe is thermally coupled to the metal box with Wakefield-Vette non-silicone thermal joint compound Type 126


Test bed #4: thermal couplings

Test bed #5

  • Tamb = 22 °C
  • system layout and configuration
    • sytem layout is the same used in test bed #4, apart from the orientation: in this case the box is vertical like test bed #3


[1] Original dimensions. Heat pipe has been tailored to fit enclosure shape.

Results

  • test bed #1
    • overall power consumption (steady state): 6.27 W (max 7.47 W) [1]
    • SOM power consumption (steady state): 2.85 W (max 3.68 W) [1]
    • Tj = 83 °C
    • ∆T = Tj - Tamb = 60.5 °C
  • test bed #2
    • overall power consumption (steady state): 6.5 W (max 7.72 W) [1]
    • SOM power consumption (steady state): 2.9 W (max 3.8 W) [1]
    • Tj = 87 °C
    • ∆T = Tj - Tamb = 64.5 °C
  • test bed #3
    • overall power consumption (steady state): 6.5 W (max 7.68 W) [1]
    • SOM power consumption (steady state): 2.9 W (max 3.72 W) [1]
    • Tj = 83 °C
    • ∆T = Tj - Tamb = 60.5 °C
  • test bed #4
    • overall power consumption (steady state): 6.37 W (max 7.55 W) [1]
    • SOM power consumption (steady state): 2.8 W (max 3.65 W) [1]
    • Tj = 66 °C
    • ∆T = Tj - Tamb = 43.5 °C
  • test bed #5
    • overall power consumption (steady state): 6.39 W (max 7.48 W) [1]
    • SOM power consumption (steady state): 2.8 W (max 3.61 W) [1]
    • Tj = 66 °C
    • ∆T = Tj - Tamb = 44 °C


[1] Even if the software load is the same for all of the test beds, overall and SOM power consumptions may differ across the test beds. This is due to the silicon static power consumption that is a function of the temperature itself.

Useful links