Heat creep with PETG

[RobboCop]

Describe the bug So, I have assembled this setup with a Dragon highflow and Orbiter. Works the goods with PLA ~210C, I printed the fasted Benchy that the unmodified community firmware would allow (acceleration is capped at 500mm/s/s) and it is probably the cleanest benchy I have ever printed. Happy, high fives all around. Now, I bought the printed largely to print things for my workshop and racecar, so PLA is pretty and all, but not up to scratch. The printer is still in my house and so I am saving the ASA for when I have finished building the enclosure for it in the workshop. About 15 minutes into my first PETG print, the whole thing clogs, I wasn't monitoring it and so I found it at the end. Clean stop, the last layer of infill that was printed looks quite sharp really. Anyway, I tried to manually feed filament, no good, no good with the nozzle removed... yikes. Cold pull no good. Heated back up and got a acupuncture needle through, started to feel happier, but still couldn't feed in filament. I've removed the hot end now and there is clearly a coating of hardened filament coating the outside of the heatbreak, at least part way up.

I'm guessing that the setup could handle the lower temperatures (and less fluid behaviour of PLA) but PETG not so much...

To Reproduce Steps to reproduce the behavior:

1. Print PETG 250C, 0.6 nozzle, 1mm retraction @ 35mm/s, I was only printing at 60mm/s (maybe printing faster could avoid the problem?)
2. Come back later and find clog

Expected behavior Hot end not jammed after print

Screenshots Looks great, Planning on re-printing the parts in black ASA once I get onto that

Additional context Unrelated, but I had to sand down the bottom of the part cooling nozzle as it was ever so slightly lower (or the same height) as the nozzle and was scraping up the prints as they were going down.

[KoenVanduffel]

I have not seen heat creap before even not when printing ABS at 280 °C in my own setup or had any heat creap reports.

I print PETG between 240 and 250 (higher temperature if I need structural parts) with 0.5 to 0.7 mm retraction @60 mm/s retraction speed. Print speed is typically between 100 and 200 and acceleration 1500 to 4000 (Klipper). The higher speed might make the problem less noticeable but it would still occur imho. The 1 mm retraction you use is on the high side for a direct drive setup and could pull some molten material into the heat break.

The physical setup should not cause any issues as long as its ridged enough. Hence I don't believe the Orbiter to Dragon mount itself causes this.

Where there could be an issue is in the fan mount in combination with the fan used. The fan shroud inlet is 29 mm wide and outlet 18 mm, maybe that is too much a squeeze for some types of fans or fans that are rotating slower. BTW the potmeter you have is that to regulate fanspeed?

I adapted the V2 design (main V2 changes: made it more sturdy and the Dragon sits 1.3 mm lower, that was a placement bug in the original version) to have a 24 mm outlet. If your fan is stalling due to the squeeze this should fix it (I can make one that does not squeeze at all if you like that better). I have uploaded it to the Dragon & Orbiter directory.

I see you are using the original part fan shroud. The latest model has improved placement and fixation and should not touch the build plate any more. The hotend will have to be turned 90° to make the ears fit. Cooling performance is improved a lot. The Dragon you have is the Phaetus one. I dont have a cad model of that one but assume the difference is only hte corners that are chamfered and not the overall dimensions.

[RobboCop]

Thanks for the fast and detailed reply.

I am seriously considering the conversion to Klipper just to reap the full benefits of the parts that I have installed. I could do my own custom configuration of the community or stock firmware, but that's probably 80% of the effort for much less gain.

This is my first direct drive unit and so it is more than likely that I have too much retraction. Hopefully that is the cause of my problems. I attempted to tune stringing based off of stringing tests, but if I am honest, none of the stringing test prints I did had great results. That's how I ended up with 1mm, but I have changed it to 0.6mm now and we'll see what that goes. I would certainly prefer the problem to be my dumb settings than a heat creep issue.

The potentiometer I have on the side doesn't regulate fan speed but is just the CHEP fan mod. It just disables the hot end fan when the hot end is below ~50C. I suppose it's not impossible that it might be stealing a tiny amount of current when the fan is at full song, but the transistor should just be acting as a relay so I am guessing it would be negligible. The fan itself is the factory unit. It is completely possible that it isn't up to the task with the upgraded parts, but I would think that this would only be the case when running the kinds of speeds/accel that you are running, rather than the speeds that I am currently stuck at, which aren't really that much more than the factory unit is capable of. Also, I would have thought that there was more headroom in the PID tuning...

I thought I had the latest version of all of the parts, but I guess I must have missed some updates. As I said in my original post, I am planning on re-printing in black ASA once I build an enclosure for the printer and move it to my shed, so I will get the latest versions of the files then.

I managed to clear the jam in hot end just by using a heat gun and warming up the entire thing. Once it was hot I gently pushed a 1.5mm drill bit through the unit and that cleared the vast majority of the plastic. I believe I got the rest after I reassembled it and heated the whole thing to 260 and extruded a bunch of plastic through. For now I am not going to print any PETG if I can avoid it and will hopefully be able to solve all of the other issues with the new ASA setup and tuning.

Once again, thanks for your advice and continued development of the parts.

[KoenVanduffel]

About the CHEP fan cut-off: I haven't seen the circuit but if its a transistor it steals 0.7V. Thats quite neglectable on the 24V power supply. I use the stock 3010 as well so that should be OK. You might have seen I made a version that doesn't close as much that should be much less sensitive.

As to Klipper you can find my setup in the other GitHub repo. Setting up Klipper from scratch (well the sample files in the main Klipper repo) was a bit of a steep learning curve as those are not quite complete. I tried to have complete setups for all 3 the boards on the GitHub.

[KoenVanduffel]

Hi, After your report I did do some further investigation and found that I can induce heat creep by slowing down the fan just a little (20V instead of 24V power supply is what I tested). This means the shroud indeed closes too much and I was just lucky till now not to get heat creep.

There is already a quick and dirty design with an increased fan outlet which i will replace with a polished up design today.

A little info to the history: In the first iteration i had a shroud opening of 20mm this has not caused issues (at least not to my knowledge), the design you used has an opening of 18mm. i tested that with the stock fan and a Sunon both were working fine. However when the stock fan has a reduced voltage it (partially) stalls. Hence this narrowing of the shroud is too close to what the fan can provide in pressure. To stay far away from the point where issues occur i have made the revised design with a 24mm outlet. The reason to narrow the outlet is not to blow too much air over the heater block and the extended melt zone coming up from the heater block. I will run a few tests with a stator inside the fan shroud to try and eliminate this stalling. There are a few videos that show what can happen to a centrifugal fan + duct, can't find back the one i was looking for but this one shows it too: https://www.reddit.com/r/3Dprinting/comments/7sqzw4/heres_what_happens_when_you_put_a_duct_on_an/ here is another one: https://www.youtube.com/watch?v=rDq4KSlZcTo&t=7s

[RobboCop]

Thanks for investigating further. I am planning on re-printing the supports once I have my enclosure built and so I will definitely print the other version when I get there. Unfortunately, I have to finish the custom lower cabinets the enclosure is going to be sitting on first, so not something I am expecting to get done overnight.

I was also considering replacing the stock heater block fan with something designed for PC watercooling or something to get some more static pressure. I just doubt there would be any watercooling gear that small. Maybe something for cooling like an industrial rack mounted server or something? I mean, next to that 5015 axial fan, noise isn't an issue anyway.

[KoenVanduffel]

Both for the Dragon high flow and Dragon standard flow new versions of the mount block with integrated fan stator to vastly improve air flow have been uploaded. Feel free to upgrade to this version.