Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
First attempt at printing a model (from Thingiverse) using PLA+. As Prusa Slicer doesn't have a profile for this I modified the temps for generic PLA by upping first layer and other layers by 5 degrees each. However I have now had 2 failures due to thermal runaway. After the second I resliced the file to print with regular PLA that I have had a multiple successes with. It too failed due to thermal runaway.
To test that it wasn't some issue with the original STL file, I printed a previous successful print file from the memory card in standard PLA. It printed absolutely fine.
So now thinking there might be something wrong with the original STL I ran the fix option in Slicer and resliced for PLA+. Printed again but this time I came back to check right about the time it had failed each other time. I was able to observe the nozzle temp start to drop right before my eyes. I decided to check the fan speed and it was running at 100% (255). I proceeded to lower the fan speed to around 150 and the nozzle temp started to come back up again. The print is now continuing past where it had failed every other time.
Decided to check slicer settings and sure enough, the fan is set to reach 100% at layer 4. This coincides with the Z axis height of 0.8 mm that was on the printer display at the time of each failure. Other settings include: Min Speed of 100% and a Max Speed of 100%. I also have Keep Fan Always On and Enable Auto Cooling checked in the Enable section. This section says that "If the estimated layer time is below 15s the fan will run at 100% and print speed will be reduced so that no less than 15s are spent on that layer. If estimated layer time is greater, but still below 100s, fan will run at a proportionally decreasing speed between 100% and 100%."
So... while I found a solution, I still have questions.
- I used the same profile (fan settings) to print the successful model in PLA but the new model failed as the nozzle temp dropped when the fan kicked in. How is it that 100% fan did not drop nozzle temp for one print but it did for the other? Could it be due to the shape of the model? The successful one is relative small in (x,y) area and the time to print each layer is significantly less than the model that fails...
Appreciate any help!
Steven
RE: Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
Without seeing the model and settings (like if you had attached a zipped up project file) the following is an educated guess.
You are almost certainly right about the model shape. Certain model geometries result in the air from the fan blowing back up towards the hot end. This causes a drop in temperature. If extreme enough then you get the problem you describe. It has nothing to do with layer time.
The simplest solution is to fit a silicone sock. Once fitted do a new PID calibration from the lcd as the thermal characteristics will be altered.
The sock will stop the hot end cooling rapidly from the blow back. It also has the added bonus of it you do get a part detached and causing a blob then its much easier to clean up. I wouldn't run a printer without one anymore as they are so effective.
RE: Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
Thanks for the response - I also run a Creality CR-10 and it has a silicon sock. Always assumed that was more to du with clean up. I am a bit surprised the Prusa would require some sort of aftermarket fix like this. In theory, shouldn't the nozzle thermistor register the temp drop and add heating to the nozzle to compensate?
FWIW, while the print was running I tested it by returning the fan to 255 and sure enough, temp dropped like a stone.
Here's the file with the STL I was trying to print. The phone case is the one that failed. I was printing it flat on the bed. The Saturn Angle Adaptor was the one that worked fine.
Thanks!
RE: Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
The sides of that phone case look very likely to perfectly redirect the fan upwards and in. You basically have a tray that all the air going down then then contained and so goes back up.
You are right that the thermistor should register the drop and it then heats up more. It is doing. That's the problem. The problem is the drop is too much, too fast and the heater cant bring it back into spec fast enough and then you get your thermal run away. You could say the printer firmware is too strict and should allow a greater time for the temp to come back up. That would be one way to 'fix' it. However that starts to get into the whole safety side of the thermal runaway logic. You could also just try doing a new PID calibration loop at your increased temperature settings too but the sock is a sure fire fix.
I've got a Cr10S too and when I fitted a petsfang and BIG blower fan on it originally it used to do the same. No matter the model. All the air going to the bed just used to blow back up. Way more cooling than it could cope with. I ended up tweaking the fan position slightly to drop it and fitting a sock to that printer too. Doesn't do it at all now and that is mainly the sock.
RE: Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
Thanks again.
I get what you are saying about the sides of the case redirecting the air - however the temp drop happens right away at layer 4 which is before the sides start printing. That said, once it was done I tried printing the Saturn file (the one that was successful previously) sliced using the same PLA+ settings and this time I watched it. It too had a significant temp drop from the 215 setting to as low as 203 but never went in to shut down. Looks like a silicon sock is in my future...
RE: Print Fan Causing Nozzle Temp to Drop and Eventually Thermal Runaway
Thanks again.
I get what you are saying about the sides of the case redirecting the air - however the temp drop happens right away at layer 4 which is before the sides start printing. That said, once it was done I tried printing the Saturn file (the one that was successful previously) sliced using the same PLA+ settings and this time I watched it. It too had a significant temp drop from the 215 setting to as low as 203 but never went in to shut down. Looks like a silicon sock is in my future...
RE:
Can someone summarise this I have the same issue and confused what needs to be done. Do I need to do PID calibration? what else?
Why does it work perfectly with old .gcode files on another SD card and doesn't with new sliced files?