Thermal safety testing: call for failure modes

  • 3D printers contain heater elements and flammable materials, so it is very important that they be designed in such a way that failures of various components do not cause fires. This is particularly tricky as the printers are often built by inexperienced people (e.g. me) and because most thermal safety features rely on the firmware doing the right thing. So I'd like to do some "what-if" testing, with the goal of producing a systematic series of tests like this:

    This post is to list all the ways I've thought of that a 3D printer could suffer failures that might affect thermal control, and to request additional suggestions. I haven't started actual testing yet. Please suggest additional awful but plausible ways your printer could die!

    Possible kinds of thermistor:

    • Normal - room temperature distinct from open, highest temperature distinct from short
    • High-temperature - room temperature indistinguishable from open, fine at working temperature

    Non-thermistor possibilities:

    Thermocouples or PT100 sensors connect to daughter boards that communicate via SPI. What can go wrong with these? How do the chips behave when various sensor failures occur? What happens when the chips themselves fail? (e.g. flaky input power or magic smoke escapes)

    Failures to check:

    • Thermistor shorted
    • Thermistor open
    • Thermistor constant resistance regardless of temperature
    • Thermistor intermittent connection
    • VSSA fuse blows
    • Thermistor falls off heater completely before print
    • Thermistor falls off heater completely during print
    • Thermistor partially falls off heater before print
    • Thermistor partially falls off heater during print
    • Heater control MOSFET fails shorted
    • Bed heater SSR fails shorted
    • Aggressive fan cools thermistor more than heater
    • Thermistor connected to wrong heater
    • Thermistor and heater both connected to wrong output (e.g. bed and hot end swapped)
    • Hot end comes in thermal contact with an inappropriate part of the printer
    • Firmware crashes (e.g. infinite loop on MCU, SD card write hangs forever)
    • User screws up PID calibration
    • User screws up thermistor settings (calibration or channel)
    • User screws up thermal shutdown settings

    Some of these are going to cause fires, pretty unavoidably (e.g. a user determined to defeat the thermal safety features). But let's prevent the ones we can and warn people about the others.

Log in to reply