near catastrophic hotend failure

[albert 2]

I just had a near desaster....a few things to learn here and a few questions as well.

I have recently bought a second Ultimaker from someone who had finished it but lacked the time to really run it and get it tweaked out.

He has built it quite well with some minor issues I was able to iron out.

The machine has the 1.5.6. board and came without an UltiController. I have ordered one and connected it to the machine yesterday. It did not work, only showed 2 blank lines - just like when you misconnect the cables. This was not the cause, after I had updated the firmware it worked like a charm.

So I started a print about 2 hours ago. Several seconds into the second layer the hotend suddenly began to smoke, and PLA began to leak out of the threads in the PEEK and dripped onto the metal block. I turned everything off by pulling the plug and started to cool the hotend down with a loose fan I always have lying around. The temperature display on the Controller had not gone beyond 220 degrees but the hotend must have gone way beyond 300 degrees. The Pla oozed out in a liquid state for several minutes. I ran the z stage down manually and let the stuff run out.

When I turned the machine back on a few minutes later the temperature showed 190 degrees. When I started to turn the heater back on after the hotend had cooled to 160 degrees the display started to flicker and showed "Heater failure". So I pulled the plug again.

I thought thats it-I have damaged the heater and destroyed the peek.Then I checked the connections between the thermocouple and the small PCB on top of the print head. I re tightened the screws and also the ones connecting the heater on the mainboard.Nothing was loose but it might have been just not tight enough.

Then I waited until the machine had cooled down to room temp and started it again. this time it worked and is still printing at the moment.

LESSONS LEARNED:

the hotend can withstand a lot of heat and "abuse" -at least for a short time.

If I had not seen the smoke the machine would have gone up in flames within the next few minutes. It might have started a fire. This defeats the whole purpose of 3D printing- If I have to sit by while the machine works I cannot use it. I dont want to buy a rocking chair and learn to knit while I am baby sitting the machine.

So this leads to the question: Is there a safety catch in the software that turns off everything if the thermocouple shows strange behavior or the heater current suddenly goes up in a steep ramp?

If the thermocouple disconnects the heater goes haywire and everything could burn up. Perhaps a "safeguard" could be programmed in to prevent "thermal runaway"....?

 

[Daid 303]

There are a lot of safety features in place. But they are not 100% fail safe right now. The biggest issue is when you disconnect the 3 wire connector. This sometimes gives an error, and it sometimes doesn't. On a reboot of the machine it always gives an error. (It's something I want to improve, as I have some ideas on how to catch this case better)

As for the "fire hazard", while it can cost you a PEEK part, a printer bed, and a lot of bad smell. The printer won't go up in flames. This was actually tested.

If you have a temperature sensor not in the heater block, the printer will try to heat up but stop with a "heating failed" error, as it doesn't see the printer warming up while it should. It also detects odd temperature readings and act on that. The main issue left, a 3 wire disconnect, keeps the temperature reading at about the same level as the disconnect. And thus makes it hard to detect the case.

 

[albert 2]

Thank you for the quick answer!

With me it was a one wire disconnect ....before going into the amplifier circuit on the printhead. IF this was the cause. I can only speculate on that.

Is the current going into the heater monitored? That might be another way to sound the alarm if something is in the woods. The difficulty with these traps is setting them up in such a way that they dont trip on a false alarm and ruin a print....

I' m glad the Ultimaker would not go up in flames - if that was tested it somehow dampens my worries a bit. The smoke screen was quite impressive though. I can only imagine what it would be like once it starts to burn thru the perspex...

Probably the only real remedy for failures like that would be a second temperature sensor in the aluminum block with a separate circuit and wires going to the mainboard. But that might be too expensive and complicated.

Please keep up the good work, I really appreciate the efforts to make the ultimaker a better machine.

[msurunner 2]

I had a similar thing happen. We think we traced it back to a plug, which caused the bowden tube to pop, which pulled the cable (3-pin) loose, which caused the runaway. I'm surprised your PEEK survived! Mine definitely didn't. I haven't checked the build instructions recently, but they used to call for the F-comb on the amp board to not have the fingers over the input. That should be reversed, having the comb over the socket and threading the wires into the fingers. That should help the 3-pin from coming out. Of the things I think could be better with that setup would be some clasp connectors between the thermocouple and amp board, and amp board to umbilical cord. Having said that, I have been pretty much problem free for the past six months without them. This does make me want to check the little screws and connectors again though!

 

[zumfab 3]

I have a similar problem. I disassembled the hotend to replace my teflon insulator. While fiddling with that I already noticed weird readings, but thought it had to do with the disassembled state of things. When everything was put back together, things turned out not to be working properly any more.

Temperature reads okay when idle. Then I heat the hotend and the temperature starts climbing normally. Around 65 degrees it slows down and around 70 degrees it stops. The hotend gets hot enough to push PLA through manually with ease, so that seems to be normal temperature or above. That appears to work like it should. I reinstalled the thermocouple sensor, which seemed to be slightly loose, but that did not change anything. I checked the tiny wires to the thermocouple circuit board and refastened them. I disassembled the small circuit board to check for damage. I used another cable to connect to the main board, but no change.

I feel I have checked pretty much every part of the chain and I do not know what to do any more. Is there any part that can partly break and produce these kinds of results? Because up to 65 degrees everything seems to work, after that it just fails.

Oh, and I must add that sometimes there seems to be a temperature error, but I do not really understand why that happens sometimes, while other times it doesn't.

Update: I fully removed the thermocouple from the heatsink and the readings seem a little high. Ambient temperature is indicated as 30-35 degrees and when I put the thing in a cup of cool water it reads 22. Both seem to read higher than they are, but I am not sure how the thing is calibrated.

When I completely disconnect the sensor the reading is 74 degrees, while this guide says it should be ambient temperature.

http://www.reprap.org/wiki/Ultimaker%27s_v1.5.3_PCB#Add-ons

 

[zumfab 3]

Right, after getting access to another Ultimaker and swapping out parts it turns out the small thermocouple PCB is probably the culprit. Ultimaker sells these, but unfortunately has an expected delivery time of over a week.

The board in question is a TC2 board, but I might get my hands on a TC1 board with more ease. Are those interchangeable?