tiny lack of extrusion on a UM1

[vincentp 0]

Greetings,

it's been a while since I have posted some news about my UM1. Now news good news as we say here.

I've printed many part with great success, mostly 0.2mm layers at low speed, then I wanted to both improve quality and printing speed, so I tuned the settings to finally get there.

First I'd like to thank illuminarti for his tutorial on max printing *volumes*. It's both "obvious" and also not detailed much elsewhere, most studies focussing on feerate or moving speed. Once I understood that printing thicklayer doesn't match with high speed, I finally warped my head around all this and could conclude on my initial prints failure which had :

- non constant extrusion due to printing support (very early versions of cura)

- printing thick layers with not enough temp and/or too fast

I've now settled to those parameters

- printing speed : 80 mm/s

- infill : 50 mm/s (not faster or I don't have perfect flatness / 100% fill on solid faces)

- moves : 150 mm/s (unchanged)

- temp 230 to 235°

- fan set to 12% (with max to 100%). The fan changed everything. Probably because my stock cooling duct isn't perfect and cools the hotend too but this settings works fine (suggestions ?)

I could clearly see the limits in terms of printing speed vs temp & layer thickness. Of course, you can increase the temp again and again but I think my extruder is definitely limited at some point. Overall I can do 100mm/s for outlines without issues BUT I won't have perferly flat surfaces and infills lack of density of material applied. Perimeter lines tend to not bond as well as at 80mm/s as there's less material extruded so they aren't touching / overlapping 100% correctly (but close).

I've been able to print bigger and more complex parts (VERY HAPPY) such as this drone leg. Now finally coming to my question :

I still have a small lack of extrusion on a very specific section of the part. Retraction is disabled

It's printing nicely until it reaches the moment where it has to travel between 3 different printing location. On each location it's printing small areas.

After the big travel / move from the left side (the lowest section of the pact) to the section where it starts to climb, it's like it's having issues to extrude.

I'm wondering what I could do to improve that. The bed is fully level, solid faces are extremelly well printed. All I can explain is that it's due to the fact it's travelling on a long distance and stops extruding during that time (but it's a really short time). Strangely enough, front and back of the under extruded section are missing matter while sides are correct.

Infill is set to 50%

support 15% IIRC

that is not appearing in other places of the print (even where it's using support).

Overall, this happens for a short time in the curve of the leg, then as soon as printing stays in the leg area and doesn't travel much, it's back in business.

any idea ?

gcode is here

http://www.plecterlabs.com/Media/Ultimaker/ArmTall3H.zip

 

[gr5 2,072]

Wow. Very surprising result. Are you sure it's underextrusion and not simply exposing layers below on this tilted area? I guess it's vertical towards the end and still has the same issue.

I looked at this area in repetier host and discovered a few things:

The area of underextrusion in the photo is on the right side of this rectangle. It traces the inner (yellow) edge first and then does the outer area where you can see the underextrusion. So it's kind of amazing that it traces 4 edges first which I would think would give it time for pressure to build back up for a properly extruded side.

The next thing I noticed is that the Z seam is in the left corner of the above picture or the arrowed corner here:

So I'm thinking somehow the Z seam movement caused enough material to extrude during the brief Z movement pause that the head then didn't have as much pressure for the next area to be printed. The path is counter clockwise from that seam but again it does the inner wall before the outer wall followed by the infill last.

 

[gr5 2,072]

By the way I also looked at the amount of plastic extruded for each line in excel and the volume is constant - it looks fine to me.

 

[vincentp 0]

thanks a lot. For the very least it tought me about repetier host that I have just downloaded. It's a weird issue for sure.

I'm not sure it's the Z seam. What you see there is the stringing (I had removed most of the strings as initial cleanup). After that section, it moves to the other side of the leg

you are right, it starts by doing that small rectangle section, it should be filled up but it's not. Then after that it does the rest of the perimeter and infill.

you might be right, and it could be exposing underneath layers as it's slowly taking up the curve. Sort of "inverted overhang".

So what, in this case ? design issue (not mine). Has to be adapted ? printed with more density (I'm at 50% right now) ? print slower ?

 

[illuminarti 18]

I haven't had a chance to look at the paths in Repetier Host... but I think the problem may be related to what is happening lower in the print...

For a start you mention that retraction is disabled? That seems like an odd choice for a part like this. I wonder if you aren't getting oozing as the head moves around inside the print, and/or between parts. Given that you're printing quite hot, and moderately fast, I'd expect to see some stringing and oozing as the head moves; that empties the nozzle and can cause under-extrusion when the print resumes.

That's probably especially likely at the base of the curve, where there's a big flat area to travel across that sounds like it has a very high infill density inside it (so lots of already-printed lines for any ooze of plastic to catch on, and get pulled out).

So, my theory is that the base of that curved part is under extruded, and then the parts above it continue to be under-extruded because you don't have a solid base to build on top of; the lower layers are a bit crappy, and the head is moving pretty fast, with a fairly thin layer, and the plastic just keeps not having much to stick to under it as that curve builds - until it gets vertical and the greater layer-to-layer contact is enough to help it stick and start building solidly again.

So, I'm just throwing out ideas, but I know I've seen that sort of under-extrusion pattern before in localized areas with fairly high linear speeds, and thin layers, where it's not that there's not enough plastic coming out, just that it never manages to stick to the fluff below it.

 

[vincentp 0]

dang ! It totally makes sense, I've never thought about retraction that way. Yes, it's totally possible that the hotend oozes a bit then it's partially empty.

Two things : it looks like it's sitting on fluff but I've checked. Under is 50% infill, and it's fairly correct. The pattern looks like it's under extrusion, like my earlier prints.

regarding retraction: I disabled it as it was worse with it. The upper part of the leg on print #1 had the curve wrong plus many holes along the way, but not everywhere. So I just removed it to reach my goal and get an overall correct print.

So I suppose I now have to tune retraction. questions :

- is there a part you'd suggest to print to test /calibrate it ? (youmagine / thigiverse #)

- I was using default settings, 40mm/s and 4.5 mm of retraction I think. I haven't tried playing much with the new cura 14 settings to deal with the minimum amount of extrusion etc but the areas where retraction should occur on that print are totally relevant and not "abusive", sections are very distant and it should occur

Is it possible that my filament was not sliding well in the bowden tube, hence having trouble to go back in pressure in the right amount of time ? I've check the "bite" / toothing on the filament and it looks ok to me, but I admit that I haven't played much with retraction so far, so I don't know much about what to do. I've hear that it tends to grind a lot the filament and that print failure could come from that too.

[edit]

I'm reading other posts of illuminarty regarding a bug in the firmware (for retraction). I have updated my UM with the firmware provided with cura 14 but that's it

Do I have to play with a Vmax setting somewhere ?

do I need to compile my own firmware ?

I use the regulard reprap marlin/sprinter code flavor. Do i need to move to ultigcode ?

also, my UM1 has the standard filament drive & gear + the v2 springloaded system (which came with a V3 bolt I think). Is there something I must adjust there ? replace the drive unit to have a better retraction ?

I do have a small added retraction clip on the extruder side. Do I need something else on the hotend side ? (I have a v2 hot end)

 

[gr5 2,072]

If you uploaded the latest firmware then your extruder max v should be at 25mm/sec but I'm not 100% sure. Go into the ulticontroller menu and go to the motion settings and it's in there somewhere. You can edit it and save it if you want but then power cycle and make sure it "took". I personally use 25mm/sec on my Um1. Illuminarti uses a higher value - I think maybe 30mm/sec.

For a typical UM1 with this part printed I found 4.5mm to be perfect for retraction distance - this is the only value that needs tuning (the retraction distance):

http://www.thingiverse.com/thing:46157

 

Without the above part you will need 5.5mm retraction due to the whole bowden going up and down on every retract/unretract.

=====

The Z seam is DEFINITELY right at that corner. I looked at the gcode. You can even see it right in my screen shot of repetier host - it changes layers just at the area I highlighted in the right pane. This means the printer hovers in this area longer than normal and so a bit more oozing. I doubled my Z accleration (but touched nothing else) and it cut my layer change time drastically (maybe 1.4X faster? square root of 2?)

=====

I think Illuminarti's analysis (underextrusion combined with broken threads) is spot on!

The overall fix for your entire problem is simply to print slower. Cut your printing speed in half and quality will go up drastically. Or just be happy with the way it is. Personally I think it looks pretty good! The reason cutting print speed in half will help is because you have a more constant flow of plastic through the head - every time you stop extruding you get a bit of over extrusion and every time you start back up you get a bit of under extrusion - if you can reduce difference it will come out better.

Retraction should help quite a bit also.

Travel speed can also help - it should be at least 150mm/sec if not much faster - maybe 250 or 300mm/sec (do experiments first to make sure your printer can do it but really it should not be a problem if things are oiled). This is because it reduces the time that plastic is not flowing.

 

[gr5 2,072]

you might be right, and it could be exposing underneath layers as it's slowly taking up the curve. Sort of "inverted overhang".

But there is another section that is even more level and it looks fine. Sort of near the upper shin. So I don't think this is it exaclty. I like Illuminarti's theory.

 

[gr5 2,072]

Another thing!

Having the leg tilted at 45 degrees makes the infill not work properly. It's supposed to alternate layers so that each layer above is sitting on a layer in the other direction.

Cura is really smart and puts extra infill under the next section above for extra support over to one side but if this infilll isn't on top of another infill at 90 degrees to this infill then it won't sit on top properly, right? I guess I would have to go back to RH to check. :(

 

[vincentp 0]

all that is very helpful thanks a lot. I'll print that new clip and experiment with retraction (again). I'll just make a fake part that simulates the travel between 3 little rectangular piles / poles and see how it goes. I might increase the travel speed too.

I don't want to print slower in the sense that MANY people are able to achieve at least that speed without insane tricking or custom firmware. I went to the point where the extruded volume seems to be compatible with the printing speed and it works great even on the infills (which are indeed printed slower). Oozing sees totally spot on as the rest of the part is just perfect, even with support active and all. It's a very contextual problem where, as mentioned, retraction should be used.

Overall I want the good stuff (like everyone). I own another (expensive) CNC and when I got it, it was moving at half speed, previous user had tons of problem and instead of actually fixing them and getting deep in the rabbit's hole, he took the cheap bargain, which is again understandable when you need to just get things done. I was happy to go the extra mile to understand and find out what the issue was, then be back in business at full speed and above. Same here. All the above is very logical and makes sense to me, so I'd rather fix it. Same if I need to upgrade my extruder drive to have it more retraction friendly (but the V2 springloaded should work fine)

my bad I guess for not understanding implication of not using it. I'll give it a shot. I'm now at a FAR CRY from what my pathetic situation was back in the day lol. Now I can't stop printing !

regarding gcode flavor, which one should I use with my UM1 ? marlin or ultigcode ?

thanks again

 

[illuminarti 18]

Marlin. UltiGcode is UM2 only.

[vincentp 0]

all right, thank you !

so it seems I'm close to have a totally tune UM1. I'll work on retraction when I'm back to the lab on monday

 

[vincentp 0]

allright,

fine tuned the parameters, my Extruder Vmax was already set to 25 in the firmware settings, adjusted speed to 25 in cura but I got the amount of retraction down to 1.8mm, now I have a good trade off between having the hotend back in pressure and oozing / stringing. Travel speed set to 200 mm/s, no issue there.

the new retraction clip suggested above is also a must, it's really strong and with the ziptie, it won't go anywhere.

thanks everyone !

 

[gr5 2,072]

but I got the amount of retraction down to 1.8mm,

Really? That's enough to get the filament off the top of the bowden (at the top of the arc) and resting on the bottom of the bowden (at the top of the arc)? I need much more. Closer to 4mm.

 

[vincentp 0]

that's correct. If I use 4 mm, the extruder has trouble getting in pressure again and I have underextrusion as a starting point, which make my part above even worse. Around 2mm seems to be right, with very little stringing (nothing I cannot deal with, and the hotend isn't really oozing).

I suspect the filament isn't sliding well enough in the tube, or hot end, hence it has issues with retraction even with the clip. It's been like that since day 1 but I'm happy I have working settings now. For most items I print, stringing isn't an issue, those are functional part, not beauty parts.