So I woke up this morning to about the worst failure I can really imagine for a printer, short of a full on fire. At some point during the print, the thermistor got pulled out of my hotend, subsequently the hotend did its job and got hotter and hotter and hotter. I dont know how long it ran like this, it was an over night print, but Im guessing 6-7 hours.
At some point the block heated to the point that the metal around the nozzle actually melted and the nozzle fell out.
This is not a complaint toward e3D they make an amazing product, but more a warning. This printer has ran for the last 8 months without fault, for hundreds if not thousands of hours. But things fail. Check your machines, check your equipment, before and after. I know I will be from now on.
All of my hotends will be getting new v6 heaterblocks as well from http://www.filastruder.com and @Tim_Elmore1
The new version of Marlin has protection from this. Basically, the firmware says, “if the heater is on, and there is no significant rise in temperature in x amount of time, shut everything off.”
I often wondered why there weren’t dual thermistors in hot ends… but i guess that redundancy is costly… here it would have come in handy to at least save some filament!
A recent “thermal runaway” commit to Marlin would have prevented this failure. But, really, we shouldn’t be relying on the microcontroller to ensure the safety of the printer.
You are damned lucky you didn’t have a fire.
Failures like this are so easily preventable using passive systems. Thermal fuses in hot ends should not be considered optional. At some point, someone is going to start a fire and someone is going to DIE.
This is exactly why I always say that people should be using axial-lead thermistors that pass through the heater block and thus can’t be removed without disconnecting the wires.
@Shauki: While you could have a hot end that has a “terminal temperature” of around 300°C, it would take a long time to heat up. Having more power available is about being able to get the hot end from room temperature to ready to print in as little time as possible.
The solution is thermal fuses. They make the hot ends safe even if you connect the heater element directly to a 12v source indefinitely. It heats up to 350°C, the fuse melts, and then it’s done. Replace the fuse and you are back up and running.
And, of course, I can’t find a thermal fuse even rated at 300°C. The highest one I can find at the moment is 240*C, which just might be barely acceptable for PLA.
This type of fuse could be made to work for higher hot-end temperatures if it isn’t integrated directly into the heater block. In that case you might just want to go for a lower temperature fuse (110°C, perhaps?) and integrate it into the wires feeding the heater element.
However there is still the chance of thermal runaway if the thermistor is damaged or malfunctioning. A thermal fuse is the only thing that is a true failsafe (there are also resettable thermal fuses).
@Ashley_Webster , that might be a 3w resistor, but it isn’t operating at 3 watts if you’re getting anywhere near 240C. 5.6 ohms at 12v is 26 watts.
E3D needs higher output heaters since they design for 300C+ temperatures.
@Ashley_Webster I am pretty sure that it depends on the hot end design, not the resistor. The failure temp of the resistor may be higher than the failure temp of the hot end. And regardless of resistor power (IE 1 W or 100W) if you have a well insulated run away the heat will keep increasing until the resistor fails. after that it may even keep increasing depending upon how the resistor fail (if it fails by shorting you could have even bigger problems)
I don’t think you are necessarily solving the problem by using a 24 watt heater instead of a 40 watt heater. Maybe your hot end won’t melt, but it could still potentially get hot enough to fail in a spectacular fashion. Imagine a software bug where your heater is on at full blast and the E3D cooling fan fails. You won’t be dumping out sufficient heat fast enough and it will get very, very hot.
I’m uncomfortable with relying on resistor burnout as a safety mechanism, unless it is a feature that is explicitly built into the heater and is well tested.
I don’t think anybody is claiming resistors burn out safely.
The simple solution is to choose a heater wattage that gives a maximum temperature when full on of about 300C. Then nothing fails catastrophically and the filament does not catch fire.
It only takes two or three minutes typically to reach normal printing temperatures, which is nothing compared to the time it takes to print something.
If you need more power for higher temperature filaments, or want to print very fast, or are impatient, then you need more power. It will max out at a dangerous temperature so you need some form of thermal fuse. You can’t rely on the firmware as you could have a shorted MOSFET, or the cause of overheating might be a crash or a bug in the firmware itself.
To be classed as safe a single fault should not lead to a dangerous state. A second thermistor wired to an independent circuit that shuts down the PSU would probably class as safe.
The simple solution is to choose a heater wattage that gives a maximum temperature when full on of about 300C. Then nothing fails catastrophically and the filament does not catch fire.
This is more complicated with actively cooled hot-ends, like the E3D. You could find a wattage that stabilizes at 300°C with the fan on, but if the fan fails your stabilization temperature will be considerably higher.
Yes but that is two faults. Something that stops the heater control and a faulty fan.
If the fan is under firmware control then the firmware could be a single source of failure but it if is wired directly to the same PSU as the heater it would need a double failure to become unsafe.
@nop_head It is not trivial to "spec a resistor that only gets to * C. That depends on allot of things. ambient temp, air flow, movement speed, insulation, line resistance, ETC.
No but it easy to find empirically. Just wire the hotend to a PSU and manually adjust the voltage until it stabilises at 300C under worst case conditions.
I didn’t set out to do it but various accidents with my early extruders taught me nothing terrible happens with a 6R8 vitreous enamel resistor.
