Beware your 3D printer (They can cause fires) (Page 4 of 8)

To add to this conversation, I had a home (work?) built Prusa i3 that some coworkers and I put together in 8/2013.  I was running with an E3D v4.  The kapton tape method (I recommend fire cement as the only way now) of keeping the thermistor in place was not adequate and I very nearly burned down my office on an overnight print when the thermistor must have come out.

The printer was a write off, the office had smoke damage that had to be cleaned and I was very, very lucky that my printer was metal framed and mounted to a fire retardant surface.

I know that a software change that detected a large difference between setpoint and actual above 100C would cover most cases and should be implemented in the firmware (I don't think any firmwares currently have this specific implementation).  However, it is still possible for a software or hardware failure to cause 100% power to the heater so manufacturers should test the stock setup with runaway conditions to ensure that the worst that happens is a melted hot end.

This is the second E3D fire that I'm aware of.  The poor thermistor mounting and multi part construction caused me to switch to a Pico on my personal printer.  So far I haven't burned my home down but I'm still trepidatious every time I do a long print.

In the meantime, my office will not allow the UM2 to be run unattended until we have created a fire resistant enclosure.

It would be quite possible to make a electrical cut-off on the same principal as a fire sprinkler, and you guys already have the equipment to do it. 

Theoretically, the only place where it should get hot enough to melt the filament is INSIDE or right adjacent to the extruder, so if you were to use a piece of filament as a trigger to hold a switch closed.  place it above the extruder or printer.  If it gets hot enough to melt the plastic then the switch opens and turns off power allowing it to cool.

Or if you want a lower temperature "trigger" use a stick of low temperature hot glue which melts at 120 C.

as I do not yet have a printer I can not actually test this out, however I could do up the electrical design, and a concept device if people like the idea.  an easy DIY deal.

Found this on Kickstarter today
https://www.kickstarter.com/projects/10 … d-printers

The Kickstarter page explains it this way:

I'd imagine it's possible that the failure that produces the initial spark / smoke might not spread if the energy source is disconnected, but if it's left connected a small initial failure (sparking of insulation, etc) might cause a larger self-sustaining fire (e.g. plastic ignites).

Definitely not a mitigation for all failure modes, but I do think it's valid for some.

Hey Greg, as an electrician by trade I can tell you first hand just because a product is "UL listed does not make it safe or hell even to code. Some things that get a UL stamp are a direct contridiction to the NEC, and just general common sense in regards to safety. Unfortunately though litte has been done. My condolences and sympathy to the OP, but just because something has an underwriters stamp does not make it safe. Squid strips are a fire hazard because it allows you to easily overload a ckt by plugging in 10 things, yes a breaker is suppose to trip but many people dont replace breakers like they should (I'll start new post to not go too far off course)...UL listed, just one example but theres hundreds. As far as having a temp cutoff im going to look around into hardware I use during my job and maybe somebody w some code ability will find something sooner but this is a nightmare im sure we all have (and should)

Edit: once printer arrives (soon hopefully) I will do some tests with inline thermal cutoff fuses and general thermal cutoffs for entire printer

"All your base are belong to us." SD4 with a RUMBA, supernight PS, 40mm fan on X motor, lawsey carriages with new better tolerance rods, flanged rear bearings, new NEMA 17 with leadscrew on Z-Axis, and e3d v6 with MK5.


Ward and Jago are my heroes

I'm bumping this thread because I had a close one last night!  Picture:

The printer was running fine all day so I figured it was okay to start a 4 hour print an let it go overnight. I watched about the first 20 minutes then went to bed and saw this when I got up in the morning.

Here is what I think happened: The object (tiki head) has some overhangs, so what I think happened is while it was printing at a high speed, an overhang curled up and the nozzle caught it, causing the Y skip.  Then the plastic continued to curl, causing further skips and the plastic (nylon) kept building up.  Eventually, it snagged the thermistor wire and pulled it out.  Notice the thermistor still hanging beside the heater block instead of inside it.  At this point, I think it was pure luck that the heater did not run away and cause a fire!

The print did finish.  I never knew this, but I guess at super high temps the dye can burn right out of nylon because you can see how the bottom of the piece is brown but it goes to green and then singed black.  Also, the nozzle and heater block used to have some gunk on them, but that was burned clean off and its pretty clean with a hint of singed plastic all around. 

I dodged a bullet on this one.  I knew the risk and decided to go for it, so I'm not blaming anyone else.  I will think a little harder about running the printer while I am away from it.  I can't say I'll never do it again but if the object has any overhangs or I have not printed it before, I will be there to keep an eye on it.  Stay safe out there.

Edit: amazingly, the thermistor was still functional and reading room temp by the time I got back to it in the morning.  I replaced it anyway.  Perhaps it was reading the temp very close to the heater block and that prevented a total run away heating scenario.

Wow, indeed, this is exactly the thing that worries me.  Has anyone seen an inline thermal fuse like they put in hair dryers that cut off at an appropriate temperature for printing?  Also, I think Repetier-Host could add some additional safety to the code easily enough.  Like, if it sees your extruder temperature suddenly change faster than is possible, to shut down.  This doesn't really eliminate the threat since R.H. sometimes crashes, but at least this kind of failure mechanism can be avoided.  What temperature was the open circuit extruder thermister reading?

Yes, the nozzle turned a ghostly color because it got so hot it burned off any debris that was one there.  It is a well used nozzle so the entire underside was caked on ABS and that burned clean off down to bare bronze.  It must have been very hot.

Where would you install a thermal fuse like the one linked?  I don't know if that could be made to work with a 3D printer, at least not at the hobbiest level.  I like the solution of using 2 thermistors and having it cut off if the temp differential between them gets to a certain amount.  Yes, its another point of failure but I think it would be worth it.  I wish E3D would impliment it and firmware was available to do this, because its over my head to write the code.

Marlin already does this. If the PID loop calls for an increase in heat and commands 100% heater but there is no corresponding increase in temperature, the watchdog kills the print. Most likely, mdrVB6 is running a very outdated version or Marlin or the print would have stopped when the thermistor pulled out.

Open circuit thermistors read 0C, which trips MINTEMP. His didn't go open circuit, though.

Maybe not, but it is a good precaution.  In addition to thermal fuses on the extruder and heat bed, I have ordered one of these on a tip from jagowilson in another thread.  My printer sits in a 6x6 equipment closet along with a bunch of other operating equipment/machinery I don't want to have to listen to while I am working so having an overhead fire suppressant makes a lot of sense.  However, since I will probably destroy everything in that closet if that thing ever trips I want to do whatever I can to make sure it never has to.  What I want is a smoke detector inside the printer enclosure connected to circuit breaker on the entire machine but have not found such a device.  Found a  couple of companies with intentions but no actual products.  Wouldn't be a hard Arduino project but I'd rather not reinvent it if I don't have to.  I need to get to a point where I *can* leave a print going while I do other stuff (or sleep) - I can't afford to babysit it all day.

Removing electricity from a fire does not extinguish said fire.

The *only* acceptable solution for unattended printing is a fireproof cabinet.