Thermal runaway all of a sudden

Discussion in 'Calibration, Help, and Troubleshooting' started by S33ker, May 4, 2016.

  1. S33ker

    S33ker Member

    Joined:
    Apr 18, 2016
    Messages:
    11
    Likes Received:
    0
    I got my BB Dual online last Friday and have been having a good time printing and experimenting with filament. I got a nice clean print of Benchy using the gcode on the included SD card with the white PLA that came with the kit. I swapped over to some ABS that I've had great luck with on a Polyprinter but encountered some serious curling of the parts. I swapped back to the white PLA and last night printed a couple of things I pulled from Thingiverse that I've never been able to get to work correctly on my Thing-o-matic. Both of the pieces printed out very nicely. I was quite happy with the results.

    On the way home from work tonight I picked up some glow-in-the-dark PLA and tested it against the SD card Benchy that worked so well with the white PLA. After 3 thermal runaways I thought it was the new filament so I loaded in the grey PLA that came with my kit (used up all the white having a good time the previous 3 days). I've now had 3 more thermal runaways with the grey and I'm noticing that it appears to be related to the fans kicking past 80%. The printer seems to be unable to keep the temperature within the range it needs to be. The only thing different is last night it worked and now it doesn't and I'm using grey (or earlier the glow-in-the-dark) rather than white PLA.

    I've read what I could find regarding loose PT100 sensor, heater cartridge, etc and have checked everything I can find in the forums regarding causes for the thermal runaway and yet nothing seems to work. I just don't understand why it worked so well Saturday, Sunday and Monday only to start failing today. Anyone encountered this and if you did and got it dealt with what did it turn out to be?

    Note for any other newbies. From what I was able to dig out "thermal runaway" is a feature that kicks in to protect the printer from damage if the power to the heater stays at 100% for too long. Since the fan is cooling my printhead below setpoint and the printer is trying desperately to bring it to temperature I guess the assumption is that the thermal sensor isn't reading right so it's going into a runaway condition. If this is incorrect would someone please clarify as I had to go to a completely different site to find any information on what it was and this was my interpretation of what was said.
     
  2. S33ker

    S33ker Member

    Joined:
    Apr 18, 2016
    Messages:
    11
    Likes Received:
    0
    Well, it's not a change in the part cooling fans. Thought one or the other may have worked loose or something. I carefully reseated them to just clear the print so that they were as far down and not blowing on the print heads as possible. If I set the temperature to 210 then bring the fans up the unit cannot maintain temperature with the fans running any where above 80%.

    Some may say "Well, just print below 80%". I'd rather fix the problem than kludge around it. I know the printer is capable of printing properly. It did it for 3 days. I just need an idea why it can't do so any more.
     
  3. Old_Tafr

    Old_Tafr Well-Known Member

    Joined:
    Mar 7, 2016
    Messages:
    731
    Likes Received:
    93
    Like you I can't explain why "benchy" printed fine then with anything else (with defaults in s3d) got "thermal runaway" (down) Maybe the way they configured the cooling within the gcode file?

    But two things.

    I lowered the fans to within 1.5 to 2mm of the depth of the nozzles
    and
    I decreased the fans speeds to zero for the first few layers of PLA, then left the max around 50 or 60% at layer 6 (so they don't go any faster for subsequent layers)

    And turned the fans off completely for ABS. I am using the default layer of 0.1mm so prints look good but take a long time. I suspect I may possibly have to have some cooling if I increase the layer height
     
    #3 Old_Tafr, May 4, 2016
    Last edited: May 4, 2016
  4. Falc.be

    Falc.be Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    263
    Likes Received:
    40
    check your cooling settings

    specially the override settings on the right, by default the fan will go to 100% when doing bridging. if the fans are not placed correctly, they will blow on the heater block and cool it down by 20-30 °C. this is enough to fool the firmware into thermal runaway
     
  5. Old_Tafr

    Old_Tafr Well-Known Member

    Joined:
    Mar 7, 2016
    Messages:
    731
    Likes Received:
    93
    On my s3d setting that override is off, I assume by default as I have never messed with it. The intention I presume is to cool more to prevent sagging?
     
  6. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    944
    Likes Received:
    190
    I reckon the way to debug Thermal Runaway - after checking the standard things - is to connect an oscilloscope (if you have one!) and watch the duty cycle (amount of time on versus off) of the heater.

    If it's 100% on and the temperature is falling well the heater just isn't powerful enough for the given situation (with all it's variables of air flow, extrusion volume rate (combination of width, height and speed), nozzle temperature....).

    The recommended solution is to wrap the heater block in Kapton tape. E3d have a silcicon insulator sheath in the pipeline....

    I found that watching the heater LED on the PCB (flashing) gives a pretty good idea of the duty cycle. If it's full on (not flickering) then the heater is probably full on.

    This approach takes doubts about whether the PID is confused right out of the equation.
     
    Old_Tafr likes this.
  7. Old_Tafr

    Old_Tafr Well-Known Member

    Joined:
    Mar 7, 2016
    Messages:
    731
    Likes Received:
    93
    For those without a scope hows about an LED and associated resistor strategically placed near the heater or above the cover with the fan in it? On the cover would be easier as it could be connected directly across the heater wires there.... may even try that today.

    I will look for the existing LED but being pointed toward it would be good too :)
     
  8. S33ker

    S33ker Member

    Joined:
    Apr 18, 2016
    Messages:
    11
    Likes Received:
    0
    @Old_Tafr I'd already thought of the fans so I dropped them as far as I could. I actually screwed up on the first attempt and got them too low and it was scraping the print off the bed. . . Doh! I have minimum clearance between the print and the cooling fans now and it still does it.

    @Falc.be The fan settings are "built in" to the Benchy gcode that came with the unit. As I mentioned I was able to get an excellent print using the code on the SD card last Friday. It was awesome to get it to work once whatever was causing the issues I was having up till then disappeared. I still don't know what I did different as I'd followed the calibration steps over and over and then all of a sudden the 20mm calibration block I was testing with came out almost perfect after which I used the SD Benchy gcode to print so nicely.

    @R Design I pulled the electronics cover so I can watch the heating LED and yeah it's coming on and staying on so I'm pretty sure that it's at 100% for whatever the thermal runaway time period is.

    @Old_Tafr There is already an LED built into the board showing the heater status. It's labeled HE0 and is just above the terminals that you connect the heater to. It will blink as the head hovers around temperature. From what I can tell there is no proportional output it is either on or off and they appear to be using pulse width modulation to control the temp. Something someone else in the forums calls bang-bang control which historically means bang - full on. . . bang - full off.

    The thing that confuses me the most is that for 3 days the printer worked without issue with all the stock settings (other than the weird Estep I had to use to get it to work). It could print the Benchy from the SD card without having to do any tinkering with the fans. The gcode already had all that in place yet it printed a beautiful example of Benchy high speed fans and all. I also used the stock settings in the dual FFF file to print two nice prints that I wanted and then yesterday it will no longer do what is used to. The stock PLA fan settings in S3D seemed to work just fine. If I have to play with settings day to day then this thing becomes useless as I'll be spending all my time and filament trying to find a combination that may work for an unknown period of time before I have to start the cycle over. I don't mean to be so negative but it is frustrating to have it work so well for 3 days then suddenly change for no reason. Well. . . no apparent reason.
     
    #8 S33ker, May 4, 2016
    Last edited: May 4, 2016
  9. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    944
    Likes Received:
    190
    Remember it's enough that you change something like layer height (= extrusion volume rate) OR print a bigger print (so the nozzle spends more time at full speed, increasing the average speed) OR increase the nozzle temperature OR.....

    and you can tip the thermal situation to a point it can't deal with it.

    I kind of am beginning to suspect that Edge prints relatively cool (like PLA?) and that BB went through it's trials always at 0.25mm layer height and so nobody noticed that the the heater was struggling. I don't think in Charlbury they use Volcanos or big layer heights or..... ;)
     
  10. Old_Tafr

    Old_Tafr Well-Known Member

    Joined:
    Mar 7, 2016
    Messages:
    731
    Likes Received:
    93
    "There is already an LED built into the board showing the heater status. "

    We need light pipes vertically upwards from these LEDs up through the cover so we can see these with the cover on.
     
  11. S33ker

    S33ker Member

    Joined:
    Apr 18, 2016
    Messages:
    11
    Likes Received:
    0
    @R Design I could see that if anything had changed. That's what's so aggravating about this. The only change from Friday-Saturday-Sunday to Monday and today is that I ran out of white filament and am now using the grey that came in the kit. Models that printed without issue with existing settings previously are no longer printing for some reason.

    @Old_Tafr Guess we could drill a 1.75mm hole and run some clear filament down to the LED to act as light pipes.
    Just tested the ambient temperature idea as temps have returned locally to where they were back on Friday and it failed again.
     
  12. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    944
    Likes Received:
    190
    btw. one of the first tests to do is to check the resistance of the Heater. People say you should unscrew it from the RUMBA but when I did that I got the same value as when the board was attached (and off, of course).

    Also measure the voltage from your power supply (mine was 24.9v - an advantage here!)

    Power = volts * volts / resistance

    When I did this calculation I got basically 30W.

    But some people don't. Our heaters were not made in Germany and occasionally one is outside spec. I suppose it's not inconceivable yours might have "deteriorated".
     
    #12 R Design, May 5, 2016
    Last edited: May 5, 2016
    Old_Tafr likes this.
  13. Old_Tafr

    Old_Tafr Well-Known Member

    Joined:
    Mar 7, 2016
    Messages:
    731
    Likes Received:
    93
    Good call ........... I wonder also how good the regulation of the mains power supply is too, i.e. if the mains voltage falls below 110v how well does it maintain the 24v ?

    Even the possibility that when all motors move, the output (24v) voltage drops?

    I know this is unlikely but it is still worth checking
     
    #13 Old_Tafr, May 5, 2016
    Last edited: May 5, 2016
  14. Steven

    Steven Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    211
    Likes Received:
    19
    I'm getting Thermal Runaway; even when trying to print the Benchy test file.
    First one failed due to the second nozzle (i have a dual) was hitting the slightly curled edge of the benchy around layer 10ish.
    Second one failed due to thermal runaway.
    Third attempt also failed with Thermal Runaway (I say the tmp on the LCD dip to 10c below set mark. At around layer 14... oddly as I watch the z axis, they did NOT rotate a full 360 to go to layer 14, although the LCD said layer 14. It did advance on previous layers...
     
  15. mike01hu

    mike01hu Well-Known Member

    Joined:
    Apr 22, 2014
    Messages:
    957
    Likes Received:
    167
    The regulation of these supplies is very good and you would have to have a serious drop (25%+) in main power voltage to have any effect. My experience of these supplies is that they are quite capable of supplying the required current for peak loads; remember that motor currents are pulses, so there is not a "continuous" flow for long periods and these average out to an overall lower current. Peak current comes when all heaters are on at the same time but that reduces substantially when temperatures are reached and PID takes over.
     
  16. mike01hu

    mike01hu Well-Known Member

    Joined:
    Apr 22, 2014
    Messages:
    957
    Likes Received:
    167
    A couple of points here. You did not say whether you have a direct drive dual or hybrid. If it is the hybrid then there is an issue with cooling being all from one side leaving the front warmer and subject to lifting of edges that are on a sloping edge, as on the rear of Benchy. I had the same problem and have yet to address it fully. As a part solution I always heat the second nozzle, even if not using it, so that any contact with cresting or turned up edges is better countered by melting its way through the obstruction.

    Secondly, the Z-axis motors do not have to turn one revolution for one layer shift. How much they turn depends on how thick the layer is and by the pitch of the lead screw.
     
  17. Steven

    Steven Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    211
    Likes Received:
    19
    I have the stock dual. It's also just the BB-1.0. (What is the difference of the BB-1.1 anyway!)
     
  18. mike01hu

    mike01hu Well-Known Member

    Joined:
    Apr 22, 2014
    Messages:
    957
    Likes Received:
    167
    The 1.1 uses the Titan extruder in a mix of direct drive for extruder 0 and Bowden for extruder 1. For your thermal runaway, it may be due to cooling of the hotend, due in part to the cooling fans being set to high, they should be only just clear of the work piece and it will help if you wrap the hotends in Kapton tape to keep the cooling air from them. Next turn off cooling for the first layer then gradually increase it for subsequent layers, say 30% for the next two and so on. There is a lot about this on the forum if you search for "thermal runaway".
     
  19. Steven

    Steven Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    211
    Likes Received:
    19
    Can you look at this clip of the failed benchy; I wrapped the volcano block woth kapton; that seems to help with the thermal runaway, although the temp was about 5-8c below where is should be.


    Please see my caption under the video.
    I need help getting this dialed in.
     
  20. Chase.Wichert

    Chase.Wichert Well-Known Member

    Joined:
    Mar 1, 2016
    Messages:
    775
    Likes Received:
    208
    Is this your first prints? Typically the runaway error is attributed to the print cooling fans being too high and it starts to cool the bed or heater block so it thinks there is a problem and shuts off to protect itself.

    Also if you could post your settings and filament type.
    Bed temp
    Extruded temp
    Speed
    Nozzle size
    First layer settings
    Cooling settings
    Etc.

    If this is the one off of the sd card, try printing the benchy with your own settings, a lot of people have had problems with the benchy that was included, not sure if the news have an updated print file or not.

    Also do you have anything on the bed to help with adhesion?
     
    #20 Chase.Wichert, May 22, 2016
    Last edited: May 22, 2016

Share This Page