this is why you don't leave your printer unattended

I did a second thermal runaway test today

[click here]

it's quite scarey how quickly i got a good flame and how quickly the fan caught fire , regrettably i have no idea what the actual temperature was when the block melted because my thermocouple has failed

for those who are a little more impatient [4k timelapse]

Edited 1 time(s). Last edit at 07/25/2014 01:44AM by thejollygrimreaper.

---

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

Great test.

I assume that was with a 12V, 40W heater cartridge, can you test a 20W?

Quote
A2
Great test.

I assume that was with a 12V, 40W heater cartridge, can you test a 20W?

12v 40watt heatcartridge running on a 19v supply

if you have a link to a 20watt heater cartridge i'll gladly order one and test

---

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

Probably cost as much as the hot end.

SC25-1, 25W.
[www.hotwatt.com]

Quote
A2
Probably cost as much as the hot end.

SC25-1, 25W.
[www.hotwatt.com]

and the rest, i'llsend them and email, they might be nice enough to send and evaluation sample or somthing

---

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

This makes me want to swap out the MOSFETs on my controller boards with OMNIFETs. These are smart MOSFETs from ST (eg: VNP10N06) with internal thermal shutdown which should help prevent most thermal runaway failures. There are still lots of other things that could fail and cause a fire, though...

One other thing I realized about your test - running a 12v, 40w cartridge at 19v actually ends up drawing about 100w, which is particularly dire. Probably lots of people do this, but 25w would seem ample for almost any hot-end. (My latest prototype hot-end has an average current draw of only about 6w while extruding.)

Quote
LoboCNC
This makes me want to swap out the MOSFETs on my controller boards with OMNIFETs. These are smart MOSFETs from ST (eg: VNP10N06) with internal thermal shutdown which should help prevent most thermal runaway failures. There are still lots of other things that could fail and cause a fire, though...

One other thing I realized about your test - running a 12v, 40w cartridge at 19v actually ends up drawing about 100w, which is particularly dire. Probably lots of people do this, but 25w would seem ample for almost any hot-end. (My latest prototype hot-end has an average current draw of only about 6w while extruding.)

the sweet spot seems to be somwhere between 20watts and 30watts,

it actually gets worse because some people are running 24v powersupplies so in this kind of event the 12v heater cartridge would be drawing around 160watts which is more power again,

i do have a potential solution to this problem and it involves a simple 2, 3 or 4 amp fuse inline with the hotend depending on how much power you want to provide to the hotend , you would get the fuse blow well before the hotend got to any dangerous fire starting temperatures
i do realise however marlins software pwm may be an issue, but thats not a big deal I've done the modification to use hardware pwm which runs a lot quicker

i will test this on monday

---

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

I once built a makeshift bunsen burner out of a bic lighter and some duct tape to see how long it took to melt some aluminum shavings. Not for any scientific reason mind you, I'm just a pyromaniac. But, I'm working on a mathematical formula to govern the power to the hot end. That should let me get rid of the thermistor all together. Trouble is incorporating it into the firmware, which isn't my strong suit.

There seems to be an urge to add more fuses and thermistors to solve the problem. It's a natural instinct, but adding a fallible component to another fallible component just makes things more complicated. Instead of adding fallible components to the system, we should eliminate them. No thermistor, no thermal runaway. Granted it doesn't solve the mofset problem, but I'd guess the majority fail due to the thermistor coming loose.

It wouldn't be a one size fits all option for the firmware, but an option, much like what exists now with thermistors, based on which hot end your using and extrusuion speed. Might take a while to write. I'll have to figure out the whole arduino thing. Just looks like a random jumble of words and numbers to me. Basically, I want to make the closed loop system into an open loop system.

Is there some protection in the software?

Something like:

If hotend is on
and temperature is less then 10C (in case the thermistor is disconnected)
then stop and shut down

though that still doesn’t protect from a thermistor that just doest touch the heater block.

I'm going to test how flammable my PLA is. The plastic in your video is that PLA or ABS?

Quote
napalmvictory
I'm working on a mathematical formula to govern the power to the hot end.

What's your feed back, and what is the math you speak of.

Quote
BackEMF
Is there some protection in the software?

Something like:

If hotend is on
and temperature is less then 10C (in case the thermistor is disconnected)
then stop and shut down

though that still doesn’t protect from a thermistor that just doest touch the heater block.

Marlin firmware got this added recently. It checks if there's power to the heater and the thermistor isn't registering any rise in temp. Once that passes a threshold it shuts everything down assuming that the heater and thermistor have become separated. I've had a heater fall out of the hot end block, causing a really bad mess...

Quote
BackEMF
Is there some protection in the software?

Something like:

If hotend is on
and temperature is less then 10C (in case the thermistor is disconnected)
then stop and shut down

though that still doesn’t protect from a thermistor that just doest touch the heater block.

I'm going to test how flammable my PLA is. The plastic in your video is that PLA or ABS?

pla and it instantly bursts into flames

---

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

i have done this test again one the 12v power supply and the wirewound 5.7ohms resistor seems pretty safe it topped out at 330c odd

https://www.youtube.com/watch?v=vVRwHwxNy7s

---

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

Quote
thejollygrimreaper
i have done this test again one the 12v power supply and the wirewound 5.7ohms resistor seems pretty safe it topped out at 330c odd

https://www.youtube.com/watch?v=vVRwHwxNy7s

I too am afraid of those ceramic heaters. Do you have a part number for the resistor? Or a place to buy it? I would not mind having 5-10 of them as spares.

---

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

Modicum V1 sold on e-bay user jaguarking11

Quote
jaguarking11

Quote
thejollygrimreaper
i have done this test again one the 12v power supply and the wirewound 5.7ohms resistor seems pretty safe it topped out at 330c odd

https://www.youtube.com/watch?v=vVRwHwxNy7s

I too am afraid of those ceramic heaters. Do you have a part number for the resistor? Or a place to buy it? I would not mind having 5-10 of them as spares.

these are apparently one of them [www.digikey.com]

the problem I've always had with using resistors is that while they are somewhat safer they aren't as reliable, most of the resistors I've used in the past have failed due to the leads falling off,

Edited 1 time(s). Last edit at 08/08/2014 06:30PM by thejollygrimreaper.

---

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

Quote
thejollygrimreaper
i have done this test again one the 12v power supply and the wirewound 5.7ohms resistor seems pretty safe it topped out at 330c odd

https://www.youtube.com/watch?v=vVRwHwxNy7s

Why is this odd? Its 25 watts instead of 40w...

Quote
greenman100

Quote
thejollygrimreaper
i have done this test again one the 12v power supply and the wirewound 5.7ohms resistor seems pretty safe it topped out at 330c odd

https://www.youtube.com/watch?v=vVRwHwxNy7s

Why is this odd? Its 25 watts instead of 40w...

i meant odd more in the context of roughly 330c ,

---

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

It's not odd at all. Every element has its failure point or peak energy absorbtion. The wire wound stuff seems to be designed in a way that it won't start a fire. Instead it will self destruct. I suspect that has to do with the ignition point of a pcb. If I recall correctly that should be in the 400 to 500 centigrade range..... therefore it's designed to fail instead of ignite. At least that is my concmussion. As for them failing on a hot end. That has everything to do with the mounting.

---

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

Modicum V1 sold on e-bay user jaguarking11

Quote
jaguarking11
It's not odd at all. Every element has its failure point or peak energy absorbtion. The wire wound stuff seems to be designed in a way that it won't start a fire. Instead it will self destruct. I suspect that has to do with the ignition point of a pcb. If I recall correctly that should be in the 400 to 500 centigrade range..... therefore it's designed to fail instead of ignite. At least that is my concmussion. As for them failing on a hot end. That has everything to do with the mounting.

looking at the datasheets a lot of them are only rated for up to 250c

---

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

I recommend vitreous enamel 6.8R resistors at 12V, [uk.farnell.com]. They top out at around 300C in a typical hot end simply because as the temperature increases the heat loss through conduction, convection and radiation increases, so at some temperature that equals the power in and it won't get any hotter. And they handle surface temperatures up to 450C so they don't fail even in a fault condition.

They have poor dimensional tolerance, so the hole needs to be quite a bit bigger than a typical resistor. Therefore they must be fixed in with something conductive to bridge the gap, such as exhaust putty.

I have never had one fail that was cemented in. I have had one fixed with kapton and one with silicone fail because they don't have a good thermal connection, so their surface can get a lot hotter.

J-Heads from hotends.com come with 5.6R silicone resistors. Although they are not as well suited I have never known one of those to fail either.

As they are shorter than cartridge heaters they allow a smaller heater block and don't stick out, so they are actually more efficient because less heat is wasted.

---

[www.hydraraptor.blogspot.com]

Quote
nophead
I recommend vitreous enamel 6.8R resistors at 12V, [uk.farnell.com]. They top out at around 300C in a typical hot end simply because as the temperature increases the heat loss through conduction, convection and radiation increases, so at some temperature that equals the power in and it won't get any hotter. And they handle surface temperatures up to 450C so they don't fail even in a fault condition.

They have poor dimensional tolerance, so the hole needs to be quite a bit bigger than a typical resistor. Therefore they must be fixed in with something conductive to bridge the gap, such as exhaust putty.

I have never had one fail that was cemented in. I have had one fixed with kapton and one with silicone fail because they don't have a good thermal connection, so their surface can get a lot hotter.

J-Heads from hotends.com come with 5.6R silicone resistors. Although they are not as well suited I have never known one of those to fail either.

As they are shorter than cartridge heaters they allow a smaller heater block and don't stick out, so they are actually more efficient because less heat is wasted.

I'll see if i can get the same ones here in Australia, bad dimensional tolerance isn't much of a problem in my hotends the hole for the heater is 6.5mm in diameter to accommodate some of the worst heater cartridges available,

---

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

Cartridge heaters are supposed to be inserted into close tolerance smooth holes. When the expand they form a good thermal connection. How do you get that with oversized holes?

---

[www.hydraraptor.blogspot.com]

Quote
nophead
Cartridge heaters are supposed to be inserted into close tolerance smooth holes. When the expand they form a good thermal connection. How do you get that with oversized holes?

it's simple .. you don't , most of these dirt cheap red heater cartridges aren't actually 6mm in diameter, I've measure quite a few in the range 5.6mm to 6.3mm in diameter and there is also a good percentage which aren't really round.

i ended up just making the hole 6.5mm to accommodate the whole lot, otherwise i'd get too many people complaining they can't get the heater cartridges in, but even with this configuration in practical use it doesn't cause any problems,

---

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

Would it be possible to use the PID values to estimate how long it should take the hotend to reach operating temperature?

I'm thinking one safety measure would be to monitor the increase in temperature vs. the power supplied to the heating element, and if it's not heating up as expected, assume the thermistor isn't reading correctly, and throw an error message.

That could limit the problems caused by a a loose, or non-responsive (or mis-configured) thermistor.

It would mean that you would *have* to do PID tuning, but I don't see that as a drawback.

Might be a general safety check, but I have so many changes in environment that I have to change temps by hand just to maintain a single temp. If PID wasn't so solid and was more adaptive, I'd adore it.
Fan/no fan is the biggest changer, between PLA and not PLA. I've gotten down the patterns so I just set a manual temp above/below so when I start it doesn't wait literally forever for a temp that won't be reached.

---

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

Now I want to put a thermal fuse or snapdisk on my hotend. seems like an easy alternative to having mosfet protection ect.