Asking for help with PETG and Printer config

[WilkPolarny]

Hello everyone!

As a new owner of a printer (now downgraded FW to 01.04.02.00), I have now ventured into printing with PETG to create more stable, structural parts.

I have always printed with (Bambulab) PLA and never had any problems - even in a fully closed chamber, except for warping on objects at the edge of the hotplate (I think the aux fan is to blame).

I have now bought Polymaker PolyLite PETG (Bambulab PETG is still sold out in the EU store) and dried it overnight to @ < 20%RH.
The spool is currently in an AMS with dryboxes @ < 10%RH.

I used the system profile for PolyLite PETG in Bambu Studio as the basis for the prints and processed it as follows:

• Flow ratio after calibration: From 0.95 to 0.9975
• PA / K: Left unchanged at 0.020 after calibration

I first performed the flow rate calibration and then the flow dynamics calibration (K val).

I left all other settings such as temperature (Nozzle= @255°C, Bed: Textured PEI @70°C), cooling (AUX off, etc.), retraction, Vol.Speed (11.5) etc. unchanged.

Unfortunately, I still have problems when printing large parts, which apparently manifest themselves in blobs and stringing and what might look like heat-creep.
Supports have also collapsed and caused problems (The thin support walls just collapse or curl up).

However, I had absolutely no adhesion problems with the Textured PEI Plate during the prints I made.

Can you help me set it up correctly step by step? There are a lot of settings that I can tweak and I’m not sure what might be useful. According to my research, there are a hell of a lot of opinions, but somehow no concrete plan with a sequence for calibrating new filament.

Since stringing and blobs seem to be a big problem, I will probably adjust the retraction settings first. The retraction length is currently at 0.8mm. The manufacturer Polymaker specifies 1mm for direct drive printers. I’m gonna try that out.
Also, I’m going to enable 1 wall for supports, since I always had issues with supports (even with PLA)

I’m not sure about the temperatures either. I am using a 0.4mm hardened steel nozzle, so the manufacturer’s temperatures (230°C-240°C) will probably not be accurate due to the lower heat transfer of steel. That’s why the Bambu default profile probably is @255°C.

Does it still make sense to play with the flow ratio? As all settings influence each other, I don’t know where to start first (in which order) to get this under control.

Is there a way to quickly iterate over the settings with different test models (other than the stuff the calibrations print)? I don’t feel like wasting half a kilo of filament again to wake up to a failed print.

I would be very grateful if you could give me some input here. I’ve already read a lot of forum posts and websites, but I can’t really figure it out.

Best regards

[rjd]

I’m also in the process of dialing in PETG and it’s been an adventure on this printer. I don’t have a complete list of steps that you can follow, but I’ll mention what I think your main issue is, how to mitigate it, and then some other PETG findings I’ve made.

First, I believe your towers aren’t getting enough time for the plastic to cool before the next layer. There are a few ways to solve this. I can’t tell from your pics if all the towers are part of the same piece, but if not, then you can spread them apart on the build plate to buy time while the extruder travels between.

Or, you could run fans. I’ve found success running the part cooling fan at 0% min - 20% max, the auxiliary fan at 50%, and the exhaust fan at 80% to make sure the chamber doesn’t get too hot. With PETG, it’s always a balance between part cooling and layer adhesion. And cooling may need to be adapted differently to every plate of parts.

Or, some combination of the above.

I think your flow ratio is fine, it’s not a factor in your print issue. I’ve also found 1mm retraction works well.

Now to temperature. You’re right about needing to bump the temps up for your hardened steel nozzle, but at a max volumetric speed of only 11.5, it may actually be hurting you. Your print speed is probably being capped by that 11.5, so if your nozzle is hotter than it needs to be, it may be making the heat creep in your parts worse.

To illustrate my point, here’s some data I’ve gathered for a different black PETG from California Filament. I’m using a .4mm stainless steel nozzle, but I don’t think it’s too much different from the hardened. I ran the Max flowrate calibration test in Orca slicer 6 times at different temps over the range of 10mm^3 to 30mm^3.

Temp: 240
Layers failed at: 12mm
Max flowrate: 22.55

Temp: 250
Layers failed at: 14mm
Max flowrate: 24.33

Temp: 260
Layers failed at: 16mm
Max flowrate: 26.30

Temp: 270
Layers failed at: 17mm
Max flowate: 27.31

Temp: 280
Layers failed at: 19mm
Max flowrate: 29.25

Temp: 290
Layers failed at: did not fail
Max flowrate: 31.05

So, at 255C, your filament is probably a lot more capable than 11.5 max flowrate. However, changing this to a higher value without implementing some form of cooling will just make your towers look worse (your printer will speed up if its speeds were capped by the flowrate). If you decide to keep the 11.5, then I suggest lowering your nozzle temp.

You can always test your filament flowrate at a given temp to be sure.

The next steps in my investigation will be cooling related to see how to best balance part strength, print quality, and speed. Although, I was pleasantly surprised the PETG seemed to print fine at 290. The layer adhesion is insane.

Having a flowrate over 30 allows for speeds over 300mm/s at .2mm layer height and .53mm line width. Right now, I’m printing some white matte PETG at 260C going 275mm/s and the parts are looking good.

Again, not relevant to your current issue, but hopefully interesting to ponder.

[WilkPolarny]

Hi, and thanks for the reply!

@rjd

First, I believe your towers aren’t getting enough time for the plastic to cool before the next layer. There are a few ways to solve this. I can’t tell from your pics if all the towers are part of the same piece, but if not, then you can spread them apart on the build plate to buy time while the extruder travels between.

Yup, that’s what I thought too. They all belong to a single part. I’m trying to print Plate 2 (support settings had to be fixed for Plate 2, though) of this one:

X3D AMS Riser Lite 58mm high by DutchWorkX3D

I thought that spreading them apart could worsen stringing, right?

@rjd

Or, you could run fans

The default profile is set to run fans at 40% min and 90% max.
The aux fan is also disabled by default - I’m really not a fan (haha) of the aux one - it always caused warping in my prints (even with PLA).

@rjd

Now to temperature. You’re right about needing to bump the temps up for your hardened steel nozzle, but at a max volumetric speed of only 11.5, it may actually be hurting you. Your print speed is probably being capped by that 11.5, so if your nozzle is hotter than it needs to be, it may be making the heat creep in your parts worse.

I agree - most custom profiles I’ve seen reduce the default 255°C temperatures by around 10°C.
I’ll also try to run some tests. I’ll set the temps to 245°C (250°C first) and move on from there with a retraction test and a max flowrate calib.

Thank you!

[WilkPolarny]

Interesting. Did a Max Flowrate-calib from 11mm³/s - 20mm³/s with .5 steps @ 245°C
I found discoloration starting at 11.8mm, and layers failing at around 15mm.
So that means, that I could push the flow rate to 11+(11.84*0.5) = 16.92 mm³/s, right?

Also, reducing the temperature to 245°C from 255°C, looks promising - at least I’m not getting instant stringing with the first line the printer deposits.

[rjd]

@WilkPolarny

16.92 mm³/s

Yeah, that sounds right. But, keep in mind that once you change it, your printer may start printing faster. This will help in reducing blobs and strings since there is just less time for them to occur. However, the skinny towers in that print will still be prone to overheating.

The reduction in nozzle temp will help, but I still highly recommend setting the Aux fan from 30% - 50%. I’ve yet to have a PETG print warp on me in my enclosure (I keep door closed and lid on). The Aux fan helps cool the part of the print the nozzle (and part-cooling fan) is no longer printing on. You could try just printing the towers for a couple inches as a test. I think there is a way to just slice off a part of the model you don’t want right in Orca or Bambu Studio.

If you can get settings where those towers print fine, then the rest of the print should be good to go too.

Also, after you slice your model, double check the actual speeds the slicer is printing the model at. They may be wildly different than what you have set, especially when the 11.5 max flowrate was set.

Another thing, the default 40% - 90% for part cooling fan for PETG is just way too high. If you can easily crack and split your prints, you know there’s too much cooling. 30% is typically the max for decent strength. Since you’re putting an AMS on this print, you want it to be pretty strong. You can double check fan speed with the slicer too.

[WilkPolarny]

Thank you!

I’m currently printing test cubes showing a few defects using the last settings (reduced temp):

Yes, I’ll try to reduce the cooling fan speed to 0min 30max -ish and I’ll turn on the Aux fan. Because you’re right - stuff is splitting easily.

I’ll also try to increase the max vol. speed as mentioned in my post above - but one after the other.

Are you aware of any decent test models I could print to verify the stuff I tweak? All I currently do is to print cubes (Orca Slicer has the built in voron cube).

And regarding the towers - I’ll start tinkering once I have dialed in settings that generally work.

Thanks!

edit:
@rjd , I did a few more experiments and adjusted the cooling according to your input. While the too high temperature was a major issue - it seems that I was going way too fast.

I printed a Bambu Cube and had to abort mid print due to blobs messing up everything:

But upon LOWERING almost all of the printing speeds by 20%, I get this:

…which is the best result I’ve gotten so far with this PETG filament on my P1S. Not great, but not terrible either. Gotta have to figure out whether slowing down more helps with the overhang sections on the test cube (bambu logo)

[rjd]

I’m glad it’s improving! I haven’t really settled on any good all-rounder test models yet. The stuff I print is generally pretty big, so I avoid using cubes and benchys since the cooling performance looks bad on the benchys, but works fine for the parts I print.

However, I’m trying to improve that and am actively looking for/making models to help test cooling performance (particularly in relation to temp and print speed).

My goal is decent overhang performance, great strength, while printing at 300+mm/s. If I figure out some good settings, I’ll make a separate post. Good luck!

[WilkPolarny]

@rjd ,

thank you very much for your advice, your detailed posts and the time you have taken for me!

I’m now starting to get a feel for the parameters and how they interact and influence each other.
I’ve now bumped my PA a bit and at least the cubes are coming out perfectly now.

I wish you good luck with your test series with your printer - make sure you post your results. It’s always interesting to see.

[WilkPolarny]

Ha,

it looks like things start to get weird once I reach a specific height. This time on a less thin model I aborted mid print:

I don’t really have a clue what might be causing this. Since it all starts to happen at a certain height (further away from the heatbed), I’m guessing it’s temperature related.

The overhang continues with the same angle at all times, but for some reason it starts to get ugly at some point (exactly at the point where the structure splits into the two towers and overall travel is reduced). But unlike with my first PETG attempts, I’m not able to break off the ugly parts - the layers seem to adhere somehow (maybe because I use 4 wall loops on this one). Its only the outermost wall that’s not straight but wavy - and only on overhangs (where the printer even slows down more).

At first I thought it might be the aux fan blowing on the left side, but the right side also has issues with a few of the outer layers splitting/curling up.

I’m currently not sure whether to bump the temperature again, or whether to reduce the cooling fan oven more (currently min 10, max 20, 30 overhangs).

Since overhangs get the 30% treatment, I guess that might be too much.
I’ll also reduce PA a little, I bumped it to .0475 on this filament according to the PA tower I printed (greatly reduced blobs, but might be underextruding now).

Also - is there a way to prevent the nozzle from gunking up with PETG? This happens on long hour prints. Stupid thing is, I can’t clean the nozzle when pausing a print, since the head will park above the poop chute where I can’t get to. Stringing and blobs are currently minimal, but the nozzle still manages to collect filament from rubbing over parts and accumulate PETG over time (which could mess up a layer if a blob releases onto the print).

Any ideas? x)

[rjd]

@WilkPolarny

the layers seem to adhere somehow

I think this is a sign that the 30% cooling on overhangs isn’t too much. It actually looks like the plastic is overheating over getting over-extruded. There is a chance that the printer is slowing down too much. What do your speeds look like in the slicer preview?

I’ve got Slow down for overhangs turned off. I’ve seen others in this forum have this fix a variety of surface finish issues with PETG.

@WilkPolarny

is there a way to prevent the nozzle from gunking up with PETG?

Turning off Reduce infill retraction will stop the nozzle from dragging across your parts. I’ve heard there are coatings you can apply to the nozzle, but haven’t tried any.

[WilkPolarny]

Hi and thanks again for the reply.

I have slowing down for overhangs enabled (default).

The speed for the slope wall is constant and at 57mm/s, wheras the rest is > 120mm/s. Only the first layer prints slower @ 50mm/s.
It’s just weird, that it only starts to act up at a certain height.
Also, “Slow printing down for better layer cooling” is enabled and with a default min-speed of 20mm/s. Layer-times are also kept at 30s/8s

[rjd]

Yeah, I’m honestly not too sure. It has something to do with the transition from the larger area of the print to the skinnier parts, but I’m not sure what. My only idea is that too much heat is still trapped in the skinnier areas, so then the overhang droops/doesn’t lay straight. This could be avoided toward the bottom since the overhangs have more time to cool.

You could try bumping up the part cooling fan for overhangs to like 50-75% and/or increase the Aux fan.

Even though the top of the part is “skinnier” it’s still pretty big, so I’m surprised the overhang is having any issue at all. Let me know if you figure it out.

[Areson]

Following with interest as I am also seeing the same issues, particularly on gentle sloping overhangs.

[PureRandom]

Found this post after having similar problems, I wish there was like a “dumb mode” you could put the slicer in that makes it print everything the same speed and temp and stuff so that one could troubleshoot better to see if it was the printer, filament or slicer causing the problem cause there is just too many variables.

[Areson]

I’ve been playing around with a test print that has both linear and curved overhangs. Linear overhangs seem to generally have no issues, while curves show poor results. Rotating the curved sections to be facing the aux fan vs. facing away doesn’t seem to make a meaningful difference.

I can’t post images yet, but generally speaking slower yielded better results, but it never went away. Even at 5mm3 for volumetric flow I was seeing issues, but they tended to disappear once the print was high enough. Layer height is likely a factor but the top of my prints where it curves back on itself in the other direction don’t show the issues that are present on the lower parts of the overhang.

I’m trying some more tests. Less cooling, as well as playing with the infill/wall overlap percentage. I saw in a different thread that someone suggested that PETG can bulge while it cools, causing the interior fill to put pressure on the outer wall. Adjusting the wall overlap might help with that.

[WilkPolarny]

@Areson

Adjusting the wall overlap might help with that.

There’s btw. a new feature in OrcaSlicer that lets you turn off overlap between the outer- and inner walls.

Precise wall

@PureRandom

I wish there was like a “dumb mode” you could put the slicer in that makes it print everything the same speed

Yea, like for example - to set the minimum time a layer has to take, one has to mess with the cooling settings. I wish there was a global setting or multiplier.

Also for the speeds … there are certain parameters that will limit overall speeds like the maximum volumetric speed of the filament. The only “dumber” way is to set the printer speed on the printer itself to silent, which simply halves the speeds.

[Areson]

I did try setting “Precise Wall” on as well. Didn’t change anything.

[WilkPolarny]

@rjd

Let me know if you figure it out.

OK. There’s some news. Not great news, but not terrible news either. I did a few more experiments until I ran out of filament on the current spool. I even reprinted the second part from above.

@rjd

It has something to do with the transition from the larger area of the print to the skinnier parts, but I’m not sure what.

It has something to do with the layer times. They are usually quite consitent from the bottom up to the top. There are a few breaks in between (green lines on the screenshot) - this is also where further layers like to fail - or in my previous case the outer walls started to get fuzzy/bulge.

The cooling settings we’ve been working towards are almost the exact same cooling settings and layer times the Bambu Basic PETG uses.
I don’t know why the Generic PETG and the Polymaker Polylite PETG system profiles are just inexplicably wrong for most use cases.

• Based off the Bambu PETG profile, I also increased the layer times from 8s to 15s, forcing the printer to slow down even more to provide enough cooling.

• I did not touch the min speed and kept it at 20mm/s in the Filament setting (this the Bambu PETG profile uses 10mm/s).

• I bumped the max fan speed to 40% from 30%.

• I disabled “Reduce infill retraction” - this helped a lot, but not completely with Gyroid infill going rogue - splitting off (tiny strings curling up) and messing up adjacent walls or layers on top if caught by the nozzle.
  I might choose a different infill type here, but I am only aware of Gyroid as the only non-path-crossing infill for structual parts.

• I bumped the AUX fan to 50%.

• I reduced the overhang fan speed from 90% to 50%.

Fan speeds:

Speed (now uniform across outer walls):

For the weird overhang situation - it ultimately seems that consistent printing speeds and cooling are important.

Disabling “Slow down for overhangs” did not really help, as the overhang/slope sections started to split off when printed with the regular wall speed.

What I did instead, was to additionally set a general SLOW outer wall speed:
In my case for my profile and model to around 60mm/s. This helped a lot.

I managed to print the model from before. It is clearly visible where the layers received more or less cooling based on the surface finish. The skinnier tower was printed super slow at 50% fan and is super shiny, whereas the rest has a matte apparence (Finding A). I don’t know whether matte or shiny is good or whether I should aim for matte OR shiny - the part itself is super stable no matter the surface finish. I was able to stand on it (70kg) and it did not crack/break with 20% Gyroid infill.

The surface defects with layers splitting affect the outer wall ONLY - especially the slopes and overhangs (Finding B). B is also the section that previously turned out hairy/fuzzy. I also don’t really know what to do here to further improvove surface quality. There’s still some bulging and gaps, but they are greatly reduced. I am using 4 walls, this could also be an issue considering the different print speed and cooling times with plastic contracting/expanding.

There are some outer curved sections that are just … white (Finding C) or have white stains that look like residue that super glue would leave on plastic. I don’t know what caused that.

I had to pause the print once because some infill became loose and caused the adjacent walls to split off. I managed to clip off the rogue strands and file down the surface and the printer bridged the hole for the layers above (Finding D).

I am currently drying another spool of the same filament and I’ll try to print the rest of the parts (which still require another 1,5kg of filament - wish me luck).
I aim to make the printed parts less ugly - for these prints now, I don’t care about the surface - as long as they do their job and don’t crumble.

[Areson]

I’ve been working from the Generic PETG profile but have not made a lot of progress. I’ll try the Bambu one.

I’ve printed almost a dozen versions of the test model I made, which is a partial box hose walls are sloped on two sides and curved on the other two, in an attempt to specifically test overhang quality. Generally sloped walls (sloping out) seem to have no issues printing, while curved walls do. I’ve tested a lot of various temperature and cooling combinations and those seem to have done very little. Slowing down the print speed in general improved things slightly (done via volumetric speed) but not enough to be what I’d consider to be clean. I’ll try variations with more consistent speeds given your results.

One thing I did notice is that the layer thickness changed the quality. .20 layers resulted in more clean looking results while .12 was more poor, but again only on the curved sections.

Both of these samples were printed at 265 degrees as that is what my temp tower indicated printed best. In both cases the quality of the curved sections improves visibly above a certain height. Generally it improves where the floor inside the box stops printing, but even above that I generally see some defects, particularly on corners but even on the faces. Various settings make this worse/better.

[WilkPolarny]

Another funny thing I noticed today: I ran the bed tramming gcode to check the bed level with a .4mm feeler gauge (without ever touching the tramming screws). Turns out, the back is on point, the front had between 1mm and 1.2mm offset. I set them to the same level. Let’s see how that affected my prints the whole time (It came like that out of the box a month ago).