Thermal Runaway due to the slice?

[LFG 0]

I have been trying to solve a thermal runaway issue for some time and I have finally narrowed it down to the sliced g-code from Cure 3.6.0. I know, it doesn't make much sense but let me see if I can describe it in sufficient detail.

 

I first noticed it during printing of a DnD hex dice case. I got he model off of Thingiverse a while ago and printed numerous prints of both the base and lid. I recently loaded it in TinkerCAD and added a design to the top which I then sliced in the top down orientation with the recently updated Cura 3.6.0 and the print failed at the point where it was starting to cap the infill and lay the base layer for the magnet holes. At the time I didn't note that as I assumed, like I am sure most of you are, that it was a printer hot end issue. 

 

Through the next month or so I replaced thermistors, connectors and wires but alas it kept trowing a thermal runaway. That is when I started noticing that it was happening at the same spot every time. Shortly there after I got another printer running and was able to check the print on that one. Again it threw a TR error in the same spot. Being suspect of a bad slice I re-sliced it and moved the model from center. Still errored. I then tried rotating it thinking the hex design was causing a bug issue. No luck, still TR errored. 

 

At this point I started trouble shooting the printers again. Swapped out more thermistors and even re-flashed the printers. By this point I finally got my 3rd printer running and again tried the g-code on that one. Grr... yet another fail at the same spot/layer.  

 

Well as frustrated as I was I decided to abandon my plan of printing this new Tinkercad revised print and went back to printing the plain Jane lid that was sliced with Cura 3.5.1. It printed fine on all three printers. I then proceeded to design card trays, battery boxes and other items for personnel use and discovered that if I designed a wall thickness between 1-mm and 2-mm in my part it would cause issues. basically, during the printing of the wall the hotends ability to maintain temperature will slowly degrade and will eventually cause a thermal runaway.  

 

I have uploaded a STL file of one of the items that has caused the TR error as well as a screen shot of the temp graph showing the degradation. The important thing here is I am still printing with these machines and getting successful prints. The only thing I am changing to get these is the wall thickness of the base model. Another thing to note is after discovering the issue I checked the hex box prints that failed and I think it was because at the same time its capping the infill its running the tapered ends of the interior particians which are at that point only about 1-mm to 2-mm thick.

 

Anyone else seeing this?

 

LFG

Battery Case Base AAA.stl

Battery Case Base AAA.gcode

[burtoogle 513]

Hello @LFG, I cannot see anything obvious in the gcode that would badly affect the hotend temperature. What printer do you have? Do the part cooling fan(s) blow onto the hot end much? If so, perhaps they are causing the temperature to drop and, for some reason, the heater can't supply enough heat? Just guessing, really.

[gr5 2,167]

I've certainly never heard of this.  It seems to take only 4 minutes.  maybe your "I" value in your PID is too low.  Did you try autotuning the PID values?

 

It doesn't seem to be oscillating too much.  My first thought was your wall thickness wasn't a multiple of your wall thickness.  For example if nozzle width is 0.4 and line width is .4 and it's printing a wall that is 0.6 it might print a 0.4 width trace followed by a 0.2 width trace.  You need more heat for the 0.4 trace and less for the 0.2.  But still, make sure your wall width is a multiple of your line widths and make sure the inner and outer line widths are the same and the "wall line count" makes sense.

 

I like smartAvionics fan theory.  Sometimes air from the fan bounces back.  Sometimes it's not a problem when starting out because it's printing slower and the bed may also be heated.

[LFG 0]

Smartavionics,

 

I didn't see anything in the code either. I I haven't tuned the PI but as I said I have no issue with most other prints.  And my current 3 printers are basically custom home builds. Two of them started life as Anet A8s but I have completely change to my own frame design and mounting system and have swapped out the control boards with Ramps 1.4s; basically their just a custom printer using old Anet parts.

 

gr5,

 

I think our on track with the wall thickness multiples. I have am slicing with three walls for a normal outer wall thickness of 1.2-mm. But since I am modeling a 2-mm wall that may be causing the slice issue. My abilities at modeling aside the Hex DnD case is what I am concerned with. Because that used to print no problem. Its been a stable print for many people but now at the 90% complete mark its failing. 

 

I added the stl so you can try if you want.

Hex Top DnD.stl

[nallath 1,120]

The PID controller doesn't like extremely frequent temperature changes. It could be that that is causing some issues?

 

Other than that, you could also just tune your PID to be less aggressive.

[LFG 0]

Well slowing down the print speed from 60mm/s to 45mm/s and no more Thermal Runaways. I am still seeing a lack of hot end control but ... I ma printing again.

 

LFG

[LFG 0]

I solved this by rolling back to 3.5.0. Not sure what 3.6.0 changed but it is either overloading the Arduino processor and it cant keep up with the nozzle or something is bugging the nozzle controls. I have not had an issue since going back. 

 

On a different note, anyone want to purchase 30 or so partial prints?

Edited March 3, 2019 by LFG

[Martinm 0]

Wow, I've got exactly the same problem. My printer(tevo tarantula) works fine with prusa slicer, but with cura 4 the hotend temperature drops after a few layers and triggers the thermal runaway protection. It worked with cura 3.5.0 so what did they change. Did you find a solution to get newer versions of cura working.

[Buzz_L 0]

In my case it seems to have been the fan. It switched from off for the first two layers to 100% at the third. This caused a steady temp drop which the hotend wasn't able to counter. When i tuned the setting to "fade in" the fan over several layers, the problem was gone

[rocketeer 0]

Yes, having exactly the same problem; sometimes a hotend thermal runaway, sometimes a bed runaway, always at the same part of the print.

I'm printing Colorfab lightweight PLA at 1.2mm shell thickness, 0.2mm walls times two and 30% infill = 0.4mm walls with a 0.4mm infill.

 

I've renewed the wiring and checked the MOSFETS.

 

I know Cura hates slicing thin-wall prints which is why I'm not printing single-wall only. Using Cura version 4.80

[rocketeer 0]

P.S. had to switch the part-cooling fan off completely, but the LW PLA prefers fan-off.

[bikeslider 0]

I have the same issue with my printer. It seems to be a cura slicing error that causes the hot end to turn off causing thermal error. Then when it moves to standby position the hot end turns back on and heats to set temp and ejects the filament. This repeats over and over.

 

[bikeslider 0]

My printer has two nozzles. On first layer (M104 s240) this command shut the hot end off causing a thermal error. Not sure why but it might have something to do with repeating the same temp setting. At start of the first layer I manually tuned temp changing it to 241 and the hot end turned back on and the print continued normally. Just to see I tuned temp back to 240 and the print continued normally. Next in cura I will change the temp of the initial layer so that it is at least 1 degree off from the print temp.

 

 

Edited December 10, 2021 by bikeslider

[GregValiant 1,248]

When the printer receives a gcode command it just does what it's told.  If there is a setting in the firmware that limits speed (for example) then you can enter 1 million mm/sec into Cura and it will be over-ridden by the printer speed limit.

Temperature is different.  If I tell my Ender to heat to 500 it will try to get there and when it gets above 265 or something it throws the thermal runaway error and shuts down.  So the error doesn't occur because of the command but rather from being above the firmware setting for max heat.

 

What printer do you have and do you know what the Max Temperature is?  If it's 240 then yes, you would get an error.  At any setting below your Max Temperature the printer should heat to the temperature.

 

When you make a post like this you need to provide the problem gcode file at least.  A Cura project file would be good as well.  Knowing what printer you have is important as they are all different even though they all do the same thing.  Nobody around here likes to guess.