Honeycomb infill weirdness

I have a strange thing going on when I print flat pieces with ABS. For instance, a 2.0 mm flat base for an electronics enclosure gets printed in this order on my rig (10" i3v):

1. Base layer, 0.4 mm thick, printed on 45 degree diagonals
2. Layer 2, 0.2 mm, on the other 45 diagonal.
3. Layer 3, 0.2 mm, back to the first 45 diagonal
4. Layers 4, 5 and 6 are honeycomb infill
5. Layer 7, on 45 degree diagonal, has to "bridge" the openings in the infill. This is where the excitement begins...

Instead of sagging down into the honeycomb voids, the layer 7 threads bow upward. My first impression was that heated air was getting trapped and puffing up. But now I don't think this is the explanation because the worst threads for bowing occur before an individual honeycomb cell is even starting to be closed up.

This bowing up is worse towards the middle of an object, where the diagonals are longer, than nearer the corners where the diagonals are short. This layer ends up having the texture of alligator skin, with lots of little sharp peaks.

The ABS cools in the upward bowed position. When layer 8, which swaps diagonals and crosses at 90 degrees, goes over the layer 7 threads, the nozzle hits the hardened plastic. I can hear and feel the collisions. Layer 9 swaps diagonals again, and is not quite as rough as the previous two. But the finished surface does not look very good.

My present working theory is that hot end (0.4 mm hex) is becoming cooler over the course of placing the long diagonal threads. This is based on the observation that the short diagonal threads don't have the issue. There is a brief stop at the end of each diagonal where there is no filament pulling heat out of the hot end. (I have printed the ABS up to 250F and still had the problem, FWIW.)

I'll also mention that the problem is much worse with ABS than PLA. My PLA bases are a little rough, too, but much more acceptable than the ABS.

No doubt someone else has seen this. What is causing it? How do I fix it?

Jim

Do you have a fan trying to cool the extruded plastic? I've experienced the same thing with the bridges over the honeycombs puffing out, and thought it was from my print fan blowing air into the cells. That would explain how they bow out before the cells are closed off.

No cooling fan. There is a fan on the extruder but it is not directed at the printed object.

Jim

That is very strange! I don't think I've had a bridge curl upwards towards the nozzle before.

Assuming you have the makerfarm supplied extruder fan, one thing I might try is to reverse the direction of the fan by unscrewing it and flipping it. My thought is that if your fan is blowing from the front of the extruder to the back of the x carriage, a small amount of air could be leaking down on to the print. I know this happens because if I have my extruder in the middle of the bed with the fan on, the bed will not heat up since the air is cooling the glass. Try flipping the fan so that the air pulls from behind the extruder and exhausts out the front of the printer.

I will try this. But that doesn't explain why the problem only happens on the long diagonals.

If I change the object by adding a hole in the middle of the flat piece to cause the diagonals to be shortened by half, then there is no issue on the shortened diagonals. The full length diagonals on either side of the hole bow up as usual. So that suggests it is not the fan, nor the geographic location on the bed, but rather just the length of a single extrusion thread.

Jim

Hmm, do you notice that the thread being extruded is coming out cleanly without sticking to the side of the hot end tip while bridging?

I have tried to look closely at this, but couldn't really tell. It does behave like the ABS is glopping up as it comes out on the longer threads. But it doesn't do this on the shorter threads??

The cone of my hot end does have some ABS stuck to it (the result of an unfortunate break in USB comm's in the middle of a print). But as best my eyes can tell that stuck-to-cone ABS is not contacting anything on the print.

Jim

Yeah I guess what I had meant is if the thread was sticking to the angled sides of the cone shaped nozzle, which it seems like it isn't.

You could try varying the print speed knob during this problematic layer to see if an increase or decrease of print speed makes this bridging more flat.

Also, if it's possible, maybe a picture of the issue would reveal something we didn't think of before.

Good luck!

I was curious what the infill settings are. If the infill is low, perhaps the layer is not adhering to the lower layer. This would be more evident on a longer run than a shorter run. It might even be a temp issue if it is lack of adherence.

Infill is set at 20%, honeycomb. I have been thinking about trying the rectilinear pattern to see if the behavior is any different.

Jim

Jim,

Just for experimentation, I would try setting it higher to see if that makes a difference. You would not have to print the complete part, just up to where it starts doing infill and see what happens. It sounds to me like you extruded filament is not sticking to the lower level. At 20%, there is not a lot to stick to? I mentioned temperature also because with a 20% infill, there is not much heat retention. Certainly less than with solid infill.

I will also try a higher temp and report back, but to date I have run from 230C to 250C. It might be slightly better at 250C, but you would not notice unless you were watching pretty closely during the print process. The final product, after 2 more layers are placed, looks about the same at 250C as at 230C.

I am curious, though, why the shorter runs of thread in the corners, or where interrupted halfway across by a hole, wouldn't experience the same effects if it is a too-low temperature issue? I don't mean to sound like I am challenging you, just trying to understand!

Jim

Jim,

No worries. My thinking is that if the temperature prevents adhesion to the next lower level, then a longer run might exhibit more distortion than a shorter run would as it cools. You had indicated that the shorter runs seemed ok, but not the longer.

Also, if the issue is adhesion but not a temperature issue, then it might be a question of having enough material to adhere to. With a 20% fill, the lower level does not provide a lot of surface area for adherence. That is why I suggested you might try a higher fill rate to see if that helped. Again, if it is adhesion, I would think that longer runs would exhibit more distortion while cooling than shorter runs.

Having said all of that, I only have worked with PLA and do not have any experience with ABS. Hope it helps.

Quote:

---

Originally Posted by JimG

... I have been thinking about trying the rectilinear pattern to see if the behavior is any different.

---

Personally, I prefer rectilinear since it is easier/quicker for the printer to do, but I'll point out that with rectilinear, you don't have full coverage on every layer. Since layers alternate direction, the only place where you are ensured a full stack of filament is where the layers cross.

One test I'd suggest is rotating the infill 45 degrees in Slic3r, so that the "long diagonals" are no longer diagonal. Probably won't make a difference, but it helps to eliminate as many variables as possible.

I experience this issue and have not yet been able to resolve at 20% infill. I only print in PLA and the issue only occurs with larger prints. Maybe 3" wide and up and only wide open flat surfaces. I've tried honeycomb and rectilinear. Only makes a mess in the middle with long diagonal strands, but is fine when it shortens up in the corners. The strands are adhering just fine to the infill but they bow up like a sail in the middle of the actual honey comb. Very weird. Going to higher infill % solves the problem but uses more material and more time of course.

I have had a chance to try a few of the suggestions I have received. Not discounting the others, but just have not had time yet to try them. Here are the results so far:

1. Rectilinear infill pattern (45 degrees) at 20%, 250F --> significant improvement vs the honeycomb at 20%
2. Change to 40% infill, increase temperature to 260F --> a little more improvement vs #1 above.
3. 40% infill, 250F, increase the flow rate by 10% --> slightly worsened vs #2. Too much plastic extruded.

Next I will try reducing the flow rate by 10%, 40% fill, and 250F. But based on what I have seen so far, I think the real keys are changing to the rectilinear fill pattern and increasing the infill from 20% to 40%. Since the result worsened when I increased the flow rate, I want to see if reducing the flow rate below my standard esteps makes an improvement.

I will also have a look at the suggested 90 degree infill pattern to shorten the thread lengths.

Jim

Jim,

It appears that you are got pretty good results from changing the pattern and some additional from changing the infill percentage. Sometimes, when I have a problem, I do a big change and if that solves the problem, bracket back to get to the tipping point. Why not try 80% infill at 250 or 260. If that solves it, you know it is infill related. Then bracket back to 60 or 50 and try to find the tipping point.

As I stated before, I only use PLA, but I do not generally go below 50% infill.

Quote:

---

Originally Posted by JimG

Next I will try reducing the flow rate by 10%, 40% fill, and 250F.

---

This sample is probably the best so far, but only by a slim amount over the #2 trial.

Next up is a change in infill to 70% and use a rectangular, rather than diagonal, infill pattern.

Jim