Fire Hazard

It was a sad day at my house. I had just completed the building and commissioning of a RepRapPro Tricolor printer and had printed some beautiful pieces of art downloaded from Thingiverse. I was still in the mono extruder mode waiting until all of the "bugs" had been fixed. This evening I was printing a part in our home office while my wife and I were in an adjoining room watching an episode of Breaking Bad. A fire alarm went off and we both looked at each other, wondering if this was a sound effect of the program. Quickly concluding this was not the case, I went around the house checking for a smoke alarm trip. When I opened the office door, I saw my new printer on fire with a flame reaching about 2 feet in the air. I was able to carry the printer outside and sprayed it with water to douse the flame. As you can see from the photos, there was major damage done from the hotbed up through the filament drive motors. Thankfully, all of the damage was confined to the printer. Just yesterday I had been printing an object with a runtime of over 6 hours and had left the house to run an errand. I hate to think of the consequences if it had caught on fire while I was away.

My guess is that the thermistor for the hot end failed causing it to overheat and exceed the flammable temperature of the adjacent plastic parts. It is also possible that the hotbed thermistor failed but I doubt the bed could reach dangerous temperatures. I was using an Azteeg X3 Pro controller. Hopefully it is still operable.

I'm sending this to alert others of the fire danger associated with 3D Printers. This in no way is directed at the RepRapPro printer since the problem could have happened to any printer. In retrospect, my controller has the option to connect two different thermistors to the primary hot end and monitor both readings. If the difference between these readings exceeds a set value, the voltage to the heater is turned off. I will definitely use this safeguard next time. I'm also open to any other suggestions for improving the safety of the design to prevent fires. Has this happened to others and, if so, how did you make the printer safer. Thanks in advance for your thoughts

Jim in Texas

ABS and/or PLA can easily catch fire when you have a runaway hot end. I've said it a million times. 3d printed parts have no business on a 3d printer x carriage. We're exposing a thermoplastic material to high temperature environments for the sake of building a 3d printed 3d printer. We need to have some serious conversations about safety conflicting with ideology. I could not live with myself if someone died using one of my products. It hasn't happened yet to my knowledge, but it will. It's only a matter of time before someone's house gets burnt down.

Quote
goldenmongoose
ABS and/or PLA can easily catch fire when you have a runaway hot end. I've said it a million times. 3d printed parts have no business on a 3d printer x carriage. We're exposing a thermoplastic material to high temperature environments for the sake of building a 3d printed 3d printer. We need to have some serious conversations about safety conflicting with ideology. I could not live with myself if someone died using one of my products. It hasn't happened yet to my knowledge, but it will. It's only a matter of time before someone's house gets burnt down.

I agree try to minimize everything that will burn and they should be in a metal enclosure with thermal fuses because obviously the plastic that your printing with will burn. One of the worst things is when the thermistor becomes unattached to the heat source. Very important to try to minimize that possibility even if it's not ideal for the thermal readout ring terminals and little screwed in standoffs with thermistor glued in seem to be a better choice. When wires short or break firmware can take care of it but there still should be a thermal fuse located somewhere near the source and a fire resistant chamber or guard over whole machine

Edited 1 time(s). Last edit at 06/28/2014 10:54PM by cnc dick.

I'm not so sure about gluing a thermistor onto a hot end. I think rpone did a cheaper hot end with a glue on thermistor, but I think axial components are the safer way to go.
See: [en.wikipedia.org].
There's a body armor called "dragon skin" that can withstand a .50 cal bullet. But when exposed to high heat, the glue holding the plates of ceramic disks together softens and disintegrates; Might be useful information if your in, oh say Iraq, Iran, Ipod? Basically glue and heat don't mix

Before taking any thing apart take pictures and post them here. Your assumption may or may not be correct, some one here might be able to better determine the cause, or guide you with a method to rigorously debug the cause of the fire, but if you destroy the evidence it wont help anyone.

These are all my notes that I've collected concerning 3d printer safety:

Warning, and Caution Label all the things.
All electrical components in metal enclosures.
Conduit Electrical Metal Flexible. For special situations, run the wires through a closed loop metal coil spring, (i.e. no gaps between the coils). Wrap the coil spring with Kapton tape or equivalent, (prevent air from feeding a potential fire).
Silicone tubing, use a sheathing over the conduit to reduce the oxygen volume available.
Flame Resistant Braided Fiberglass Sleeve Impregnated Silicone
Carbon dioxide gas for plants with an electrically operated solenoid
If using conduit to exhaust fumes, place a smoke detector inside it.
A thermal fuse that goes off before the temp reached combustible temperatures. The ones used in coffee makers are less than 2$
thermal fuse that shuts down the master relay
Fuse linkages or fuses to the power supply that blow before shorted wires have a chance to heat up.
Add inline Thermal Fuses to the 11A and 5A lines.
Thermistor should have a fail safe in the firmware to guard against it being dislodged.
electrostatic filter
Tyvek® (Type 5, and I think Type 10) can be used as a filter for > 1.0 µm micrometer particles.


PopCorn Connector: Flexible plastic conduit, semi air tight. [www.thingiverse.com]
Firmware kill switch, Repetier working on implementing this feature. Thermistor Failure - Kill Printer [forums.reprap.org]
Short circuit fried my laptop!  [forums.reprap.org]
Static Conductivity: Small synchronous rubber or urethane belts can generate an electrical charge. [www.scribd.com]
RELAY, SPDT, 5A, Double-winding latching, 12V, OMRON ELECTRONIC COMPONENTS – G2RK112DC [uk.farnell.com]
Aico Ei141 smoke alarm and base unit that has a 5A relay output. It has nc and no contacts. I just want it to cut the mains power. [forums.reprap.org]
DYP DYP-ME010-A Smoke Sense Detector / Sensor Module w/ Relay - Black + Green [dx.com]
MQ-2 MQ2 Smoke Gas LPG Butane Hydrogen Gas Sensor Detector Module For Arduino AB [www.ebay.co.uk]
Thermal shutdown for your controller. Airpax power transistor bolt onto heatsinks costs a couple of dollars and is available in a wide range of preset trip temperatures. [airpax.sensata.com] [www.ktmarketing.com]
If you do get indeed a 12.4v on main connector and at same time 11.8 on D8, another probable drop (~0.4v difference?) could be on the big yellow ptc polyfuse, and may check that it does not get too hot because those are reported to catch fire.  [forums.reprap.org]
CO2 fire extinguisher (Flame Defender) [forums.reprap.org] [www.amazon.com]
Design for safety, paper. [www.kodiakconsulting.com]
For ultrafine particles, ionizing filters do a great job. They charge up any particles leaving the HEPA filter and catch them with the other electrode. [hackaday.com]
HEPA filter packed in activated carbon [forums.reprap.org]
CE Standard Certification for RepRaps [forums.reprap.org]


Insurance pitfalls of 3D printing
[insider.zurich.co.uk]

Additive risk: Insurance and 3D printing
[www.hlinsurancelaw.com]

Edited 1 time(s). Last edit at 06/28/2014 11:55PM by A2.

Or you could just apply some common sense.

Its a heater (well actually its two) , you don’t ever leave a heater unattended!

I don't know if common sense applies to a multi axis cnc fused deposition modeling appliance.

Quote
goldenmongoose
I'm not so sure about gluing a thermistor onto a hot end. I think rpone did a cheaper hot end with a glue on thermistor, but I think axial components are the safer way to go.
See: [en.wikipedia.org].
There's a body armor called "dragon skin" that can withstand a .50 cal bullet. But when exposed to high heat, the glue holding the plates of ceramic disks together softens and disintegrates; Might be useful information if your in, oh say Iraq, Iran, Ipod? Basically glue and heat don't mix

what I was talking about is not glued into the hot end its glued into a 3 mm brass stand off with something that looks like a high temperature adhesive. And then the little 3 mm stud screws into the heater block so it is replaceable. If you look at the ring terminal ones that they sell it has something that looks like the same adhesive not saying it can't fail but I think I trust it more than tape

Edited 2 time(s). Last edit at 06/29/2014 05:45AM by cnc dick.

It is about common sense and yes making things out of metal some people advise, but that drives up price, there should easy hacks and safeguards out there to prevent a major fire hazard ie have build area enclosed and a co2 system which actives if detects theres a fire'but best prevention is making sure you check on it ,i work with cnc machines all day at work and i would not dream of leaving that running while i wasnt present although we have coolant and grinding metal i have seen one catch fire and nothing flammible except the oil in the coolant so anything even if the x carriage is made out of metal there still have wire flex and maybe other stuff which is flammable treat it like a oven you wouldnt turn oven on and leave it without regulary checking on it and be in the the vacinity of it

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

Quote
cnc_dick
what I was talking about is not glued into the hot end its glued into a 3 mm brass stand off with something that looks like a high temperature adhesive. And then the little 3 mm stud screws into the heater block so it is replaceable. If you look at the ring terminal ones that they sell it has something that looks like the same adhesive not saying it can't fail but I think I trust it more than tape

Kind of like the MakerBot safety cutoff switch: [www.thingiverse.com]. It used a thermostat: [www.digikey.com] to kill the board at 140C. Unfortunately nobody's producing products like this anymore or even having conversations about it. This forum has a thousand sub forums on every little insignificant thing out there, but no safety forum? Priorities admins.

Thanks for the comments.

Photo No. 2 is as good as it gets. From everything I saw during the disassembly today, the fire started at the heated block which points a finger at the thermistor coming out of the hole in the block. I had tied the resistance heater and thermistor together with Kapton tape and then inserted both devices into the heater block. The heater, thermistor and block assembly were then taped together again with the Kapton tape. The protruding wires were then fastened to the plastic extruder mounting frame. No glue was used. The printer probably had about 8 hours of good runtime before the fire started. I was running ABS at 245F at the time.

My best guess is that the tape gradually released as a result of time and temperature allowing the thermistor to pull out of the hole in the block. In spite of the photo appearance, the damage is limited to the plastic parts above the hotbed, two motors, 3 linear bearings, and one set of extruder parts. The controller powered up and the hotbed looks like it is okay except for visual damage.

My plan is to...
...install redundant thermistors with software to cut off power if the temperature difference exceeds a nominal value
...install thermal fuses where fire hazards exist
...reduce or enclose plastic parts where possible
...minimize wire connections
...install thermal shielding if possible between the heater block and the plastic support above it
...don't leave an operating printer unattended!!!

Gonetotx

This is a great point and one that should be considered for future designs. The piece being printed under the nozzle will always have the potential for igniting as long as we are printing with these plastics. Therefore, everything above the extruder ( x-axis extruder support, filament drive assembly) should be either made non-flammable or moved out of the way. Another way to put it is to say that everything above the nozzle and within the x-y footprint of the hotbed (plus a 5" safety zone) should be non-flammable or adequately shielded.

3-D printing is a slow process and their are going to be prints that are going to take 24 hours even when a small bed. So watching them for long periods of time becomes pretty much impossible they have to be very reliable or just do small prints to take eight to 10 hours. Possibly doing larger prints in pieces that only take eight or so hours would be safe

I think sensibly matching heater power is another thing that people should be thinking about more.

Sure a 40 watt heater cartridge heats up your hot end in 30 seconds, but you have to wait 10 minutes for the bed anyway. If the thermistor falls off, then 40 watts will very quickly start a fire. For most people, dropping down to a 20 watt resistor will allow the same printing performance, but temperature will top out at 250 - which might melt things, but is much less likely to cause a fire!

I think the best way to go is a 3 watt flame proof 6.8 ohm resistor. Heats up in a couple minutes and very stable if you insulate your heater block. I don't trust those heater cartridges either and I've never had a short like the chinese knockoff cartridges.

This discussion about heater spec is really good info for me. I'm using a 12 volt power supply for the hot end and a 24 volt supply for the hotbed. I have no idea what the spec is for my extruder heater. Will a 20 watt heater be enough for the ABS? What is the difference between the 20 watt and the 3 watt flame proof 6.8 ohm resister? By calculation, I would have to tune down the voltage supply to about 4.5 volts to get the 3 watts. Is this correct. What are your thoughts about the safety of running the hotbed on 24 volt power? I did this to speed up the heat up time.

Agree there is hardly any discussion on safety of these home made printers. Something similar has happened with me once in my case it was PCB of heated bad caught fire not sure about the exact root cause but something to do with short circuit. The damage wasn't that bad as i was around so immidiately put the fire off, lost only the PCB.

As many of you have discussed there are number of parts in the whole machine that could catch fire. In stead of making it fire proof i am thinking of the following alternative solution, interested to know what other thinks.

Use raspberry pi with camera to run the printer controller s/w instead of PC, Octoprint (www.octoprint.org) support Pi and Camera both and many people already using it. I think Octoprint also have Android/iPhone app using which we could see the how the printer is doing and also control the printer remotely. Though I yet to try Pi + Octoprint on my printer. We can write software that could interpret the camera images and decide if printer is on fire or may be could attach some sensor to Pi to detect fire. Once the fire is detected send a message to Mobile or ring some alarm at the earliest possible. Idea is to inform the user as quick as possible to avoide large damage like priter putting a house/office on fire.

Idea behind the thought is there are too many things in this printer that could go wrong and too many things that are flammable (wires, plastic part etc...). If we try to fix all of these printer cost would definately increase many folds.

You don't have to turn down voltage for a 3 watt resistor or anything like that. It just uses less amps. They can reach temps of 300+ C. More than enough. I've never run anything on 24v, but as long as your controller can handle it, it should be fine.

Quote
Gonetotx
My plan is to...
...install redundant thermistors with software to cut off power if the temperature difference exceeds a nominal value
...install thermal fuses where fire hazards exist
...install thermal shielding if possible between the heater block and the plastic support above it

I have the same printer as you. I wonder why RepRapPro are not designing such safety features into the printer kits? The thermal runaway problem has been around for years now, so it's not a new issue.

Andy

Edited 4 time(s). Last edit at 06/30/2014 02:56PM by ajayre.

Honestly, I cant seem to find a proper thermal fuse for ABS 300c preferably. However this one should work nicely connected in series with the resistor or heated cartridge. It looks to be 240c thermal fuse. Which should work nicely with pla. Just tape it to the hot end and connect it in a series with the resistor. Not a bad price either. IF a thermal runaway happens, it trips and kills the hot end, damaging the print and possibly chewing through the filament, but no fire.

[www.ebay.com]

This brings me to the point of having a proper metal chamber on a 3d printer. I think its beneficial for the print and beneficial to contain any sort of hazard. If done properly you could put a 70c thermal fuse connected to a relay and a heatsink to smooth out the temperature inside a heated build chamber. Ensuring that if any fire does start it kills the print and then the chamber will smother out any fire. Could be an effective means to keep people safe and have a heated chamber as well.

---

My Personal Blog. Build blog.
[engineerd3d.ddns.net]

Modicum V1 sold on e-bay user jaguarking11

Quote
Gonetotx
This discussion about heater spec is really good info for me. I'm using a 12 volt power supply for the hot end and a 24 volt supply for the hotbed. I have no idea what the spec is for my extruder heater. Will a 20 watt heater be enough for the ABS? What is the difference between the 20 watt and the 3 watt flame proof 6.8 ohm resister? By calculation, I would have to tune down the voltage supply to about 4.5 volts to get the 3 watts. Is this correct. What are your thoughts about the safety of running the hotbed on 24 volt power? I did this to speed up the heat up time.

3W is the nominal power these resistors are made for. As you said, it is reached under 4.5V. But using 12V, you reach ~20W (U²/R).

About the safety, a simple solution may be to put a thermal switch at the top of the hot-end, where it is attached. Using a NC (normally closed) contact, opening when reaching, say, 80°C, can save your house.

[fr.farnell.com]

This one need to be relayed. A good solution is to make an emergency circuit, with a big relay on the main power supply. This relay is maintained on through a contact (big red button).Then, you can put different other contacts in this command loop, and the thermal protection can be one of them. Using a latched command circuit ensures that the power won't be on again even if the contact closes back.



Or you can use this one:

[fr.farnell.com]

directly on the hotend supply (it can handle 2.5A).

This one is OK too:

[fr.farnell.com]

---

Frédéric

One way to avoid the thermistor to detach from the hotend is to wrap the heating block - including thermistor and heating resistor - with a few turns of "self fusing" silicone heat resistant tape :


There are a few links in that topic : [forums.reprap.org]
Here are some specs of that kind of tape : [www.rescuetape.com]


Some people used to wrap the heater in Kapton tape, but I believe it can detach after a while.
When I was using kapton, I secured the tape with copper or steel wire :



Added benefit for both solutions is to lower heat waste through air blowing on the heating block, and diminish radiation transmission of heat to the part being printed.

---

Most of my technical comments should be correct, but is THIS one ?
Anyway, as a rule of thumb, always double check what people write.

Quote
DeuxVis
One way to avoid the thermistor to detach from the hotend is to wrap the heating block - including thermistor and heating resistor - with a few turns of "self fusing" silicone heat resistant tape :

+1!

Quote

Added benefit for both solutions is to lower heat waste through air blowing on the heating block, and diminish radiation transmission of heat to the part being printed.

...which also allows to reduce the cartridge power, and so reduces fire risks...

---

Frédéric

Quote
jaguarking11
Honestly, I cant seem to find a proper thermal fuse for ABS 300c preferably. However this one should work nicely connected in series with the resistor or heated cartridge. It looks to be 240c thermal fuse. Which should work nicely with pla. Just tape it to the hot end and connect it in a series with the resistor. Not a bad price either. IF a thermal runaway happens, it trips and kills the hot end, damaging the print and possibly chewing through the filament, but no fire.

[www.ebay.com]

This brings me to the point of having a proper metal chamber on a 3d printer. I think its beneficial for the print and beneficial to contain any sort of hazard. If done properly you could put a 70c thermal fuse connected to a relay and a heatsink to smooth out the temperature inside a heated build chamber. Ensuring that if any fire does start it kills the print and then the chamber will smother out any fire. Could be an effective means to keep people safe and have a heated chamber as well.

If you look it up, you'll find the Thermal Hold temp of that is only 200C. I searched everywhere and could not find a fuse with a thermal hold high enough to be of use.

[www.sellifuse.com]

[www.digikey.com]

Edited 1 time(s). Last edit at 07/02/2014 08:23AM by tjb1.

i solved this problem many moons ago by using axial thermistors and insulating the leads with ptfe sleeve:

click here

using this method there is nearly no easy way to get it into a thermal runaway mode, as the thermistor cannot physically leave the heater-block,

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

Quote
thejollygrimreaper
i solved this problem many moons ago by using axial thermistors and insulating the leads with ptfe sleeve:

click here

using this method there is nearly no easy way to get it into a thermal runaway mode, as the thermistor cannot physically leave the heater-block,

My Kossel Mini now has dual axial thermistors with 5c maximum difference I believe, I went with dual after having an E3D kill a thermistor at 300c and then run a 24v catridge at full power for 2-3 minutes before I caught it. I'm guessing it hit 450-500c as the kapton tape turned black and crispy. It had the glass bead (tadpole?) thermistor and it failed, readings jumped all over and it actually reported a lower temp of around 180c the entire time until I killed it. Trying to restart did give me a MINTEMP error I think which is how I actually found out it had broken. Polycarbonate is not fun

I don't see power limiting working for many, my J-Head with a 40w cartridge has trouble maintain temp with the blower fan. I have to slow fan down to 180 to maintain 200c, should make a duct so not as much hits the heater block but hard enough to find a place for a fan on a Kossel Mini, let alone hard mount it and make room for a duct that doesn't affect hotend temps.

I've got the E3D v5 and I've hit 380C with the stock thermistor. Maybe it's a hit and miss quality thing. Still works weeks later. Reports samish temp of bed thermistor at cold, so I assume it's temp reporting hasn't changed.

---

Realizer- One who realizes dreams by making them a reality either by possibility or by completion. Also creating or renewing hopes of dreams.
"keep in mind, even the best printer can not print with the best filament if the user is the problem." -Ohmarinus

Quote
MrDoctorDIV
I've got the E3D v5 and I've hit 380C with the stock thermistor. Maybe it's a hit and miss quality thing. Still works weeks later. Reports samish temp of bed thermistor at cold, so I assume it's temp reporting hasn't changed.

There has been talk of a new thermistor being able to withstand temps up to 500c I believe but I doubt E3D has it or has been including it yet. The one in the kit is only rated for 300c, it's not a quality issue, it's the limit for the thermistor and you going to 380c is just luck.

Quote
tjb1


I don't see power limiting working for many, my J-Head with a 40w cartridge has trouble maintain temp with the blower fan. I have to slow fan down to 180 to maintain 200c, should make a duct so not as much hits the heater block but hard enough to find a place for a fan on a Kossel Mini, let alone hard mount it and make room for a duct that doesn't affect hotend temps.

based on what I've seen , i now have serious reservations about a few things around what is being done with power supplies, the type of power supplies used and style of pwm used and to an extent how the Max_pwm values are used in Marlin but most of it seems to fall on deaf ears.

i'm surprised if anything we don't see things like what has happened to Gonetotx much more often especially with so many people leaving their printers running unattended and even while they are at work!

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

My ideal auto-extinguisher would be four CO² cartridges placed around my printer that blow all over when fire is detected, all electronic sources being shut off as well.

Edited 1 time(s). Last edit at 07/02/2014 02:37PM by MrDoctorDIV.

---

Realizer- One who realizes dreams by making them a reality either by possibility or by completion. Also creating or renewing hopes of dreams.
"keep in mind, even the best printer can not print with the best filament if the user is the problem." -Ohmarinus