Delamination with PET-G Carbon Fibre

  • I am having some issues with my prints, printing in PET-G Carbon Fibre.
    Everything I have printed is fine, but these ones obviously have a problem and I do not know what to adjust to get them right as they are structural, checked the forum, googled etc but nothing seems to match.
    My Print settings are, Print speed 45mm/s, Bed 60, Hotend 245 Fan 50%
    Just wondering if anyone else could give me some pointers, please.
    I am at a loss as all of the other parts have printed fine, hence the head scratching!

    EDIT: I did put a little force on them as they are structural especially the bottom part.

    0_1560582141344_Part1.jpg 0_1560582149472_Part2.jpg 0_1560582164282_Part3.jpg

    I think it is heat related, but I am not sure!

    Thanks in advance. Paul.

  • Did you try lower temperature? 245 sounds high for PET-G

  • Thank you @dragonn
    I did try, will try again now I have changed the nozzle, thanks

    Did a free air extrude and it was curling as it was coming out.
    Did not bother cleaning it just threw another nozzle on.

    I also use the DWC to drop my speed to 60% on the initial layers and then bring it back to 100%.

  • @paulhew Have you tried a higher temperature? I found using monoprice's petg, I needed 250C to get a strong layer adhesion.

  • Moderator

    @paulhew said in Delamination with PET-G Carbon Fibre:

    Hotend 245 Fan 50%

    Too much fan for PETG.

  • As per previous suspicions around the fan but with the additional question: What is your part cooling arrangement?

    What's your nozzle diameter and layer thickness? I've seen some people push layer thickness right down to 0.1 or less with a 0.4mm nozzle and they appear to suffer lower strength across the layers.

    Does your infill form properly as per the visualisation on your slicer or does it come out dotty?

  • Here is my 2 cents worth of advice.

    Either stop using the part cooling fan or raise the temp. I print with CFPetg all the time. I don't use the part cooling fan at all with it. 245C sounds about right but it depends on the filament supplier.

    Try running a temp tower to get it dialed in right before printing good parts.

    Petg likes higher layer heights instead of lower ones. It bonds to itself very well.

    Also make sure the filament is dry, Petg will absorb water as it sits out.

  • Apologies for my lack of replies.

    I have now turned the fan off. I was printing @ 240 but it struggled to get out of the nozzle but seems fine @ 245.
    Good idea about the temp tower, i will re run it again as the weather keeps changing here at the moment.

    I am using a 0.4 hardend nozzle, 0.2 layer height and a Z offset of 0.1mm

    Thanks to you all greatly appreciated as usual.

  • So I re ran my temp tower for this PET-G Carbon Fibre.
    No Fan, bed 60, nozzle 260 - 220 Speed 45mm/s


    None of it looks good, I had no spitting or hissing whilst printing.
    I fear that I will have to buy another brand of CF PET-G and print again.....

    Would appreciate any advice again, please.


  • @paulhew I haven't tried PET_G Carbon fibre but that surface finish and the poor layer adhesion issues are remarkably similar to problem I had with a reel of eSun PET-G. I wrote about it on my blog at the time and there is another thread on these forums where the topic was mentioned again. The consensus was that the filament must have been wet, despite the fact that it was a brand new reel and had arrived in sealed evacuated packaging. Long story short, I gave the filament to someone else who managed to dry it and eventually get some prints out of it but they weren't great. I've printed plenty of other brand PET-G since and never had any issues like that. So my take on it is that the problem is with the filament and not your machine.

  • @deckingman Thank you for your advice as always Ian.
    I have spoken to who I purchased it from and set them pics of the temp tower and one I did in standard PET-G. Big difference. They are debating if to send me another spool.....

    Also whilst cleaning the dross off of one on the bearing blocks legs it snapped and I was being gentle.

    It is in the airing cupboard, my partner is happy as she gets to have the heating on!!!

  • @paulhew For info, this is the blog post that I did back then (May 2017).
    You'll see what I mean about similarities.........

  • @deckingman Very similar to what I am experiencing! Very good write up.
    I purchase a fair bit from this company and have never had reason to complain until now.

    Might have to take a run out to E3D......

    And I cannot chat with you Ian @deckingman !

  • @paulhew said in Delamination with PET-G Carbon Fibre:

    And I cannot chat with you Ian @deckingman !

    I've changed my profile settings so you should be able to chat with me now.

  • Moderator

    It almost seems to be getting better at 220.

    Definitely similar surface finish to what I've seen from moist filament. Will be interesting to see if the dryer makes any difference.

  • I spoke to E3D today and told them of my dilema after reading deckingman's blog.
    Options are Edge filament or MatX.
    Edge my printer could handle it but the MatX needs a bed of 105deg!
    I do not have the mag bed that came with the ender anymore, I have either Mirror tiles or borascilate glass.

    Do you think the ender 3 heated bed could handle 105deg with either a boro or mirror bed?

    Thanks as always!!!...

  • Moderator

    I've used a few rolls of MatX. It's ASA based, which is basically super ABS. It does like a hot bed and no cooling, just like ABS, so you'll likely need to max out the Ender bed which should be able to hit 100c+ given enough time and an enclosure of some sort. It's also a bit smelly, just like ABS.

    Edge is basically PETG and prints pretty nicely. I used Edge for a few of the parts of my DBot and it works fine. I'd probably give Edge a shot to start with.

    ASA is going to have a high likelihood of warping and cracking on the Ender 3, imo.

    I should probably mention that some of the Edge I've had has needed to be dried before it would print nicely. Not sure if that's a common thing, or just happenstance.

  • Update! You might want to get a beverage of your choice or something stronger!

    I bought a roll of Techna PET-G Carbon loaded filament as I thought there was something wrong with the filament I was using, even though I bought a dehumidifier @deckingman suggested, dried the original filament for 2 hours etc.
    As previously mentioned my prints had NO strength even though it was carbon PET-G 75% infill

    So with this new filament I made sure my E steps were correct, printed the cube, measured the walls, calculated my Flow Rate for Cura, it equated to about 74% to get to the correct dimension.

    Re-sliced one of the parts I need for my build, again it was struggling to stick to the bed, checked my settings etc and 3 or 4 failed prints later I hit the Emvio Engineering website to make sure I had the correct paramters for this filament.
    I noticed George was online or maybe. I sent a message on the chat, basically saying I was having the same issues with this new filament as I had with the other carbon filament.
    Explained that I had made sure my ESteps were correct, nozzle and bed temps. I just happened to mention my Flow Rate value of 74% also and he then asked me to call him.

    He asked why and how I ended up with a flow rate of 74% so explained the process I use from Matts hub, Teaching Tech etc.
    He informed me that by reducing this value it would cause delamination issues, prints not sticking and strength issues, basically everything I was experiencing.
    Next question was, do I have Simplify3D?
    No I use Cura.
    He then explained to me that in Simplify 3D there is a extrusion multiplier for 'tweaking' the amount of material being extruded, to give you dimensional accuracy without messing with the flow rate.

    So most of last night I was printing cubes, 2 walls and they were coming out at approx 1mm thick walls.
    By 9pm, I was getting miffed and decided to print the bearing holder at 100% Flow and see what happens.
    1st layer I had to tweak the baby step lower as in my slicer it was set to 0.09, cancelled tweaked Cura to 0.07 started the print again.
    All the support went down without an issue and no dragging, nothing missing and left it overnight.
    Took it off the bed this morning, removed the supports, I did a strength test too, and it took a lot of force to break the upright parts and I am not a wimp by any means. (They are not that thick plus they are Hollow)
    0_1561285315506_IMG_20190623_111825.jpg 0_1561285321659_IMG_20190623_111818.jpg 0_1561285328393_IMG_20190623_111809.jpg

    Yes, the bearing hole is a little tight 21.43mm for a 22mm bearing, but the strength was incredible definately strong enough for the bearing holder.

    I noticed in Cura 4.1 there is a new feature in Experimental section called 'Flow Rate Compensation factor' so I spent two hours playing with it, could not find any documentation on it or how it works or supposed to.

    I went back to my 'dried' PET-G CF filament, printed 7 cubes changing this parameter and they were all coming out with the same wall thickness, 2 walls with 100% flo so I do not know what this parameter is for.
    More testing required...

    Conclusion, Changing the 'Flow' under Material section in Cura is obviously causing the issues in regards to bed adhesion, de lamination and weak parts, it seems right and a lot of documentation out there refers to it, but IMHO it is not the correct parameter to change. Maybe I need to play with hotend temps?
    I have first hand witnessed the difference between 74 and 100 'Flow' and what it does or does not do.

    Finally - at this moment in time I am seriously considering purchasing Simplify3D, yes I know it is £120 ish, but if it has a proven flow rate compensation, it will be worth it, plus it has a trial for 2 weeks.

  • Moderator

    Flow rate and extrusion multiplier are exactly the same thing. Same behaviour as the slider in the DWC. Don't waste your money on Simplify3D over a silly comment like that, using a 74% multiplier in simplify would lead to the same problem.

    A flow rate of 74% would indicate something is very wrong indeed. Anything more than 10% in either direction from 100% would seem to indicate a separate issue exists in the extrusion system. With properly calibrated E steps it should be pretty close to 100% flow/extrusion multiplier/whatever you want to call it.

    And while it can vary between filaments and plastic types etc, it shouldn't be enormously different.

    I want to focus on the Titan Aero for a moment. This is where the rubber meets the road as it were. The pinch roller mechanism and the drive gear. The tighter the tensioning arm is tightened, the more it compresses the filament and changes the effective diameter of the hobb, leading to different e steps. This can also lead to some odd behaviour if it's too loose/too tight.

    I think you might be having a similar issue to one I've had in the past. Basically the tensioner arm became loose over time, and rather than extruding less, it was actually extruding more. The roller was pinching enough to still grip the filament, but it was loose enough that the effective diameter was increased and the steps per mm values was too high, causing over extrusion. I started compensating by reducing the flow rate, but it kept getting worse. Then I noticed how loose the tensioner arm was. The thumb screw must have worked it's way looser and looser when changing filament because there's no pressure on it at that point.

    I retensioned the arm so that the teeth leave a nice clean bite in the filament, but doesn't crush it or deform it. Then I marked the position of the thumb screw and nut with a marker so I can keep an eye on it.


    With everything retensioned I redid my e steps calibration and found that it had changed quite a bit. Then I did a flow rate test and discovered that it was pretty much perfect at 100% for PLA (which I had used for calibrating the e steps) and for PETG 92% worked best. For some of the filaments with rough texture or lots of fill material like StoneFil/WoodFil a flow rate of 105% worked best.

    Another thing to check is the hobb gear teeth. First, Is the filament lined up with the middle of the hobb? The Aero is known for needing some adjustment in this regard. The large black gear needs to be moved slightly on the hobb and in turn the drive gear on the motor shaft may need to be adjusted accordingly.

    If the filament is lined up with the sweet spot you could be getting unexpected results as the effective diameter changes. It's a fairly wide sweet spot, but it's still totally possible to be far out of alignment.

    Second, are the teeth starting to get worn? Carbon filled filaments are rather rough and E3D has recognized that the standard hobb can get worn quite quickly. Similar to hot the nozzles get worn. They sell hardened steel hobbs now as well.

    As you can see from the photo, my teeth are starting to get worn as well. Many many hours of printing and quite a bit of abrasive materials.


  • @phaedrux Thanks you once again for assisting.

    So I have tightened the pressure on the filament screw, the nut visable on your picture which gives you a guage of how much pressure was nigh on off.
    Measured 110mm, extruded 100mm with the PanelDue, measured 10mm, so E-Steps have not changed.
    Yes - I am following your Ender3 Manual. Absolutely invaluable.

    So did the rest of the guide, adjusting DWC and 100% seems right or maybe my glasses were dirty! Filament diameter was 1.72mm.

    Also did part 2 and my Flow calc was 83.7%. Does that sound right?
    I am going to print 2 motor mounts and see how good they are in the morning.

    Thanks again.


  • Moderator

    @paulhew said in Delamination with PET-G Carbon Fibre:

    Flow calc was 83.7%. Does that sound right?

    Sounds like it could be. I guess we shall see how the prints look.

Log in to reply