Thermal runaway protection available in S3D?

[Farr0wn3d]

I hadnt really considered this as a threat for times when I leave my printer unattended but now that I'm aware of it I would like to see if theres something I can do to mitigate the risk.

Is there a provision in s3d or in the marlin firmware that can be set so that in the event of a MOSFET failure or a thermistor losing contact with the heat block that the printer will shut down?

I saw a reddit that mentions marlin firmware inputs to prevent runaway due to a thermistor detachment but thats about it.

[insta]

It's in the newer Marlin firmwares ... I had it enabled in my custom firmware I posted awhile ago.

It's somewhat poorly implemented though, the first time the temperature is set, it's ignored ... but if you change temperature values again (say, when starting the next print), it'll trip the protection. It works great if you reset the printer between runs. Look back through my post history, I've posted the firmware before.

It works by analysing how long the heater has been on, and how much the temperature is changing per second. The idea is the thermistor popping out and reading ambient air won't heat up quickly even though the heater is red-hot for several minutes. I think I had it set to 20C in 4 minutes for the hotend, and 4C in 2 minutes for the bed.

[jimc]

There should really be a thermal switch or something independant of the printers systems that will shut off the power supply if the hot end reaches over a certain temp. Just a backup safety system.

[ednisley]

a provision in s3d or in the marlin firmware

I'll inject my "why you must not trust firmware for safety functions" rant just once more, then I'll STFU and let everybody get on with firmware monitors. Page down to the Important Warning below the pix of my Thing-O-Matic hot end modified for cartridge heaters:

http://softsolder.com/2011/02/25/cartri ... rst-light/

I built a thermal lockout for the PC-grade power supply in the Thing-O-Matic, but the same reasoning applies to all other things that can get catastrophically hot without warning:

http://softsolder.com/2011/03/14/thing- ... t-circuit/

Yes, a firmware / software monitor can help, but it cannot detect failures that kill the monitoring program. The key question: do thermal runaways happen more frequently than firmware glitches? I think the answer is a resounding "No!", but that's just me.

To be fair, I don't have a thermal lockout on my M2, but it also doesn't run unattended for more than half an hour at time while I'm puttering around in the shop...

Now, I'll STFU...

[lem]

I'll inject my "why you must not trust firmware for safety functions" rant just once more

Ed, that works both ways.

My brother worked on F4 fighter radar systems. Eighty percent of the system failures were in the hardware circuits that were monitoring the radar performance. It makes for a really bad day when you can't fire your missiles because some failure in the monitoring hardware shutdowns a perfectly good radar system.

As a side note, the IoT must scare the hell out you.... (It scares me).

Larry

[ednisley]

F4 fighter radar systems

That's a whole 'nother smoke: I'm talking relay logic, not avionics!

Particularly avionics in a Mach 2 fighter dating back to the late 1950s. Hermetic-can transistors FTW! [grin]