Filament bite marks, a bit too shallow?

[IRobertI 520]

It's been so long since I've paid any attention to the bite marks on the filament that I can't remember what they're supposed to look like... Recently I've been battling under extrusion and I finally took a look at the filament and extruder. Don't these marks look way too shallow to you? IIRC they used to be a lot deeper. This is with the newest bolt and extruder.

I also took a look at the passive roller and it looks deformed. But I find it hard to believe this could be the issue? Or could it?

The screw is fully tightened just like I've always had it.

 

[snowygrouch 1]

I cannot say for certain, but what I can do is say that with respect to this I did 2 things.

1) I replaced this black plastic part with a steel ball bearing (that fitted OVER the tiny bearing in the middle and has

in the end the same diameter.

2) To test the bite strength, I just warm the hotend - then pull by hand on the filament coming out the bottom

of the extruder. If you can pull it back through the extuder then its too soft. Its not very scientific but

worked for me. With mine I cannot pull the filament back through the extruder even pulling so hard that

the machine is having its two opposite feet lifted off the table.

I dont think the black plastic roller is a very good idea, because the pressure on it is very high and consequently

as you can see the roller is now deformed. This is probably limiting the "effective" spring pressure because

beyond a certain point its just squidging the roller...and hence all the effort doing that is less effort going into

pushing the PLA onto the knurling.

C.

 

 

[IRobertI 520]

Funny you should mention replacing the delrin with a ball bearing, I had the same idea Unfortunately the one I had that fits is only 14mm outer diameter rather than the (IMHO a bit too small) 15mm of the delrin ring. I might order one that fits just because it'd feel better. And I agree that using a relatively soft plastic as the roller is not exactly ideal.

However, there's no way I can pull out the filament without breaking something so I guess I'm good on that front. On to the next thing

 

[gr5 2,173]

I didn't even check the two links below but they are in my notes to some day buy one of these ball bearings. I think these are snowygrouch links maybe:

http://www.thingiverse.com/thing:35106

http://umforum.ultimaker.com/index.php?/topic/1676-modification-of-upgraded-extruder-drive/

 

 

[IRobertI 520]

Thanks for the links. Also, what the hell, I had already posted in that thread and I have absolutely no memory of doing so... I think I'm getting Alzheimers. I guess I didn't have the same idea, I had actually seen someone do it and then promptly forgotten all about it.

Wow...

 

[gr5 2,173]

lol. You're one of those guys who thinks every idea is his. :-P

 

[IRobertI 520]

I guess I am gr5 Ah, the bliss of thinking all your ideas are fresh and amazing

Right, under extrusion keeps happening and I'm starting to run out of ideas so I turn to you guys. Details:

- Newest bolt

- Newest extruder upgrade

- Newest hotend

- PLA from colorfabb

- Temperature 200C-230C (usually I print around 190C but I wanted to remove that variable)

- Bowden hasn't moved

- Removed the hotend and took a look inside the PEEK, absolutely clean, no sign of leakage (installed it just a few days ago).

- Bed is freshly levelled and shouldn't matter regardless once you get a few layers in

- I can manually feed the PLA by turning the gear

- 0.2mm layers with very conservative speed as you can see in the vid. 30(ish) perimeters and 60(ish) fill. I've printed much faster than this in the past without problem.

- Slicer used was KISSlicer with the same kind of settings I've always used with success before.

Here's a picture showing the problem. The lower part shows the problem that starts happening a few layers into the print. The top part shows what the first few layers look like. As you can see it's over extruding there if nothing else. That overextrusion is my attempt to compensate. I started at 2.82mm for the filament diameter which looked pretty much perfect on the first few layers and on this print I'm down to 2.70mm.

The top print is done immediately after the bottom one. I aborted the bottom one, ripped it off the platform and then immediately restarted the print. So there was no cooling or anything inbetween. That's the weirdest part I think, that it stops under extruding when I restart the print and then after a few layers starts doing it again.

Here's a couple of videos showing first the bottom part of the print and then when it starts going down hill. Please excuse the focus hunting, I was using a macro lense and didn't have enough light so I had to keep the aperture open.

 

 

Any ideas on what to check next?

 

[gr5 2,173]

You've got some play (aka backlash) on one of your axes. Probably the belts. Probably the "short belts". Try sliding down both stepper motors a little tighter. Feel all your belts. I know what you are thinking: "this isn't play, this is underextrusion". Well maybe it is a little, but the major effect I see here is due to backlash. See the explanation here:

http://umforum.ultimaker.com/index.php?/topic/1872-some-calibration-photographs/?p=14396

In the above link, first look at the photo of the 5 cubes and read the text. Then scroll down 2 more postings which explains why loose belts can cause this pattern that is similar to underextrusion.

 

[IRobertI 520]

Nah, I don't think that's it. I don't have any short belts as I'm doing direct drive. The long belts feel and sound the same way they have for the past year. Grub screws are all tight. Holes look nice and round, infill touches the perimeters (except after a while when it starts under extruding).

If it was simply the belts you'd see the same error from the very first layer, but it starts happening after it reaches around 1.5mm or so. It's all very confusing.

 

[gr5 2,173]

Well you definitely have some play/backlash somewhere. Grab the head and jiggle it a little. Maybe the bearings are a little loose? You are looking for around .2-.3mm of play on one or both axes.

 

[IRobertI 520]

The head feels solid. It only moves once I put enough pressure on it to make the rods flex slightly. Just gently jiggling it around I don't feel or see any play.

Again, if this was play/backlash, shouldn't I be seeing it on all layers, all the way from the very first layer (or at least the second or third once any bed irregularities are cancelled out)? As it is the problem doesn't show up until after it has put down a few layers.

I'll be doing some more testing during the week, I haven't had time to do any more investigation over the weekend as I've been out for most of the day.

 

[gr5 2,173]

In the first picture above, on the lower half, I can clearly see two filament passes touching, then a gap, then 2 filaments, then a gap. This can only be caused by play. I don't know where it's coming from. Maybe the nozzle is loose? If the part is rectangular, then sometimes the diagonal infill only touches left and right walls during part of the infill. Other times it might touch the top and left or top and right or top and bottom walls. If the play in the 2 axes are different (only one axis has play) then the pattern could change as you switch. In a square, both switch at the same time (the diagonal from corner to corner) so it's harder to guess which axis it might be. But it really looks like play.

In addition you may be underextruding a bit. If you have a UC you can try increasing "flow". This menu item only appears after you start printing.

 

[illuminarti 18]

I'd guess that the problem with under extrusion on higher layers is due to pressure build up in the head. When you stop and restart the print, the pressure has some time to normalize as excess plastic oozes out. In the second video when the head does its travel move to start a new layer, you can see that it kinda globs out plastic. So I think maybe the excess pressure is causing under-extrusion (similar to what I discussed in my blog post). What extrusion width setting are you using in Kisslicer? Also what do you have set as the minimum and maximum extrusion rates?

 

[IRobertI 520]

Interesting theory. I took a look at the latest piece of filament that I sent through the printer and I'm not seeing the pattern change. That might be because I switched back to the old extruder pressure plate which can put much more pressure on the filament though. It shows really nice, clear and equidistant bite marks but of course the issue still persists.

Extrusion widths are set at 0.4mm to match the nozzle. Extrusion rates are left at the defaults which are way too high (max 75). I never really considered changing them since I never had any issues before. For the test vids I was printing at maybe 70% of the speed I usually print at.

I haven't really touched the printer the last week though, at the moment it's just a source of frustration and bad mood... I'm waiting on parts to rebuild this thing. New pullies, new belts, changing to the "banana blocks" which I just barely managed to print out before this whole mess started (my wooden ones are starting to fall apart) and new couplers for my direct drive. Hopefully that will get rid of any backlash issues.