Need your tips to get better prints

[gr5 2,172]

 

By the way, illuminarti did a study that showed that with typical printing speeds and temperatures, the filament tends to underextrude by 10% of requested due to head pressure.

 

[corngolem 0]

Don't know how you conclude to underextrusion. It thinks the filament is smaller (2,89) than what it is (2,93) so more plastic arrives at the head (overextrusion).

I tried increasing the flow before, the head was already full and the feeder turned empty.

 

[ultifan 0]

Don't know how you conclude to underextrusion. It thinks the filament is smaller (2,89) than what it is (2,93) so more plastic arrives at the head (overextrusion).

I tried increasing the flow before, the head was already full and the feeder turned empty.

 

That's correct, if Cura thinks it's getting 2.89 mm, but in reality you have (measured) 2.93 (average), you should get more PLA mass per length /time, thus get over-extruding, in theory.

Unless, your measuring tool is not accurate....

Imho, the best way to find out if this is causing under-extruding is, lower your PLA ( ! ) filament diameter in Cura.

(I assume you've checked in Cura PLA, not ABS)

 

[gr5 2,172]

My bad, sorry.

 

[illuminarti 18]

When measuring your filament, I'd recommend always measuring in two perpendicular directions at each point, and then calculate the cross-sectional area at that point, compared to the nominal perfectly circular 2.89 mm diameter. Often the cross-sectional area will vary less than the maximum diameter, because often what is happening is that the filament is becoming slightly oval, not actually fatter all the way around. Within reason, a small amount of eccentricity in the shape isn't a problem - mostly you need a constant cross-sectional area.

That said - 3.09mm is probably enough to get your filament stuck in the Bowden tube - that could certainly cause extrusion problems. A 2.89 to 2.93mm variation in average diameter is unlikely to be noticeable by itself, but if the filament is getting stuck, then you're going to have more problems.

 

[corngolem 0]

A small update:

I'm pretty sure that the Silver PLA 4043D ingeo (and any Silver or Gold filament) included in every Ultimaker needs a higher temperature (230 to 240 instead of 210 to 220) to print without underextrusion. It might also be extremely prone to curling at the nozzle (natural 4043D from Makerbot extrudes straight).

I noticed that the filament is not always moving constantly when the feeder's large wheel is turning, and when underextrusion occurs the filament has an uneven trail instead of clear "insect" like marks.

I would conclude that either the drive's bolt pattern is not adapted to plastic (this would explain the tiny parts of plastic found on the uneven trail and around the bolt) or that Silver PLA is too fragile.

I also noticed that underextrusion and bad marks happen when there is a large surface to print, it's very sparse when there are only walls or walls with a narrow space between them. Edit: a possible explanation is that the repeated diagonal movements pull on the filament enough to create that trail which makes it difficult for the drive's bolt to catch the filament.

 

[corngolem 0]

Who can solve my under extrusion issue ?

 

See how underextrusion mostly occurs when there is a large surface to print

(I replaced the double washers by one nut to prevent retraction from tightening the large wheel's nut)

 

[gr5 2,172]

Those pics certainly look like underextrusion to me. I wish you could zoom in a little more as backlash can cause similar patterns.

There's so many possibilities:

1) Illuminarti mentioned that your filament seems like it might get stuck in the tube. Did you address this? Try taking a section that is 3.09 and see if it slides smoothly through the tube. Remember if you don't reply or convincingly refute our suggestions we might not be very helpful in return.

2) Often the "bolt pattern" gets partly filled with plastic and occasionally needs cleaning out. Take a look at it with a flashlight and reading glasses or similar. It's probably tricky to spot gray plastic in a gray bolt. THF is a solvent that might help (the only solvent that seems to work on PLA).

3) I don't know why but the assembly instructions didn't work for me for the feeder so I moved around the washers a bit to get everything to line up nicely. Plus I added a huge washer to keep the big wood gear from wobbling. This is a kit, it's not perfect. Don't be afraid to make your own solutions.

4) Spring tension: If you measure the spring length it should be about 11 to 12mm. Illuminatri's is 11mm. Mine is 11.5mm. Both work fine for us.

5) Speed. Illuminarti did some good work that shows that with a .4mm nozzle, 10 mm^3/sec is about the limit that you can squeeze through the nozzle. He would argue that is over the limit and it's closer to 8.

10 would be the equivalent of .2mm layers at 125mm per sec or .1mm layers at 250mm per sec. Are you approaching these speeds or going over? What happens if you don't go over 5 (62mm/sec at .2mm or 125mm per sec at .1mm)? These speeds are fine at >230C for me but at 190C one needs to go much slower.

6) Clogs. I kind of doubt you have one but you might. They can occur in the nozzle head (remove very gently while very hot, I would avoid using tools and instead use fingers/gloves if possible - then burn anything in there out with low temp fire) or in the peek part (much more work).

7) Friction. Some people have drilled a tiny hole in the top of the bowden tube (at the top of the arc) and add a drop of lubricant every few hours of printing and claim it helps. Not me. If you try this google what to use first as this is going to get mixed into the PLA in the head.

 

[illuminarti 18]

What George said. Plus... it looks like the defects are mostly in areas where you're printing solid over the top of infill? If so, try increasing your bottom/top thickness to add a few more layers, to give you a better chance of filling it in.

Also you could try increasing the infill percentage a bit, so you have smaller gaps to bridge. You could try lowering the bridge speed in Cura (I think that slows down solid fill over sparse infill, as if it is bridging over gaps, but I'm not 100% sure).

 

[corngolem 0]

1) & 7) in my opinion there is way too much friction in the tube, I've replaced it with one of a larger inner diameter and there is no friction anymore (it comes from a reverse osmosis device, it's written "DM TUBE DPE04 >LLDPE< 0/D 1/4" X I/D 0. 170" 70°F 15 BAR WRAS NSF 51 61-G 04A1C21 04 -4"). I just started a print and it works fine, only inconvenient is that it's not transparent. I'll know if this was the problem in half an hour.

2) there is not enough plastic dust or bits on the bolt to prevent the carrying of the filament. I consider that as long that there are clear markings on the filament, the feeder is working.

4) the tension seems already adequate, it's hard to tighten more, there are markings. My earlier test showed that the spring had to be loosened considerably for permanent underextrusion to happen.

5) I printed those blocks (pic posted today) at 200% of 50mm/s 0,2mm Skin 230°C - the UM is supposed to go at 150mm/s so I don't expect underextrusion. My earlier tests showed no underextrusion even at 600% and 240°C.

6) I'm not sure how a clog could occur or what it could be made of if there's only filament. I disassembled the head and only found carbonized plastic between the nozzle and the pipe, it was only a deposit, couldn't have clogged the nozzle. I removed it with a metal brush.

Underextrusion makes lines with about 50% missing plastic, so adding more of the underextruded lines would be like pouring more water in a pierced bucket.

 

[illuminarti 18]

Is the 'under-extrusion' happening on the solid bottom layers of your print, or only on the flat tops? If it's on the bottoms, have you made sure your short belts haven't slipped again, causing backlash, as you discussed with George at the start of this thread? If not, then it may be volume-related under-extrusion as discussed below...

If it's mostly on the top surfaces, then I stand by what I said earlier - make your top and bottom skins thicker, and infill a higher percentage. Printing solid plastic over sparse infill never works very well, as there's no plastic underneath for the extrusion to stick to. You need several layers - I typically go for at least 4 or 5 - for all the gaps to fill in and get a smooth surface again. Your photos seem to show exactly this effect on the top - thin cratered spots.

Regarding speed and volume considerations, the fact that the Ultimaker can print at 150mm/s is irrelevant. It certainly can, and it can go a lot faster than that. In the right circumstances it can print at 250 to 300mm/s. What it can't do is fit more plastic through the nozzle than will be allowed by the viscosity of the plastic, the pressure that the extruder can deliver, and the size of the opening in the nozzle. For the stock nozzle, this limit is somewhere in the 8-10mm³/s range. Furthermore as you approach this limit, the pressure in the head builds to the point where the extruder cannot keep up, and under-extrusion occurs. The only way to print the fastest linear speeds is to lower the layer height proportionately, in order to keep the volume of plastic per second below the limit. Printing at 100mm/s and 0.2mm layer height is asking for 8mm³/s - right around that limit, and definitely where I'd expect you to start seeing under-extrusion. See my blog post for more details, and examples of the degradation in feeder engagement pattern that looks a lot like what was posted earlier in the thread. Even when the feeder is making clean marks in the filament, it may not be working as effectively as it should.

If your filament is the proper standard size, and the extruder isn't improperly set up or grinding, so as to deform it badly, then the standard bowden tube should be just fine. You want as little space as possible in the tube in order to keep the filament path as constant and predictable as possible. If you can't push the filament through the tube by hand (before it gets to the hot end) very easily, then there's probably a problem with your filament, not your tube.

Clogging can happen if dust and debris gets dragged into the nozzle, or if molten plastic gets forced under pressure into unheated areas where it can solidify and be unable to move back through the hot end. The former can be general dirt that gets stuck to the filament by static, or impurities melted into the plastic, or occasionally wood fibers and dust from the extruder drive. The latter (molten plastic in bad places) is particularly likely when printing very fast (in terms of volume per second) due to the pressure build up, or very slowly, due to the transmission of heat up the filament causing it to soften in areas where it normally wouldn't. In my experience, using the Ulticontroller or sometimes even a computer interface, to feed the filament when doing filament changes is an excellent way to cause instant blockages, as you can end up feeding at way too high of a rate, leading to very high pressures in the nozzle.

 

[illuminarti 18]

Also, looking at your photos, I see some problems with under-extrusion in some spots in the side walls of prints. This could be due to printing at too high of a volume... (if you're using Cura, what is your wall thickness/nozzle size settings, they can sometimes make things even worse).

Also, do you have the Joris and/or 'Skin' (half-height perimeter) settings turned on? Those can both cause defects in walls.

Perhaps it would be helpful for you to attach a sample gcode file here, or email it to me at gcode@fbrc8.com, so we can take a look at the settings, and see if anything unusual is happening.

 

[corngolem 0]

I'm talking about missing plastic inside the lines that makes the print weak, cushion like and see through - not backlash or other issues.

The previous picture (direct link) shows that it mostly happens everywhere (infill, wall, top) when the surface is large enough.

Here's a large 80x80x20mm block I printed yesterday after checking everything. It shows underextrusion on the entire infill.

Gcode for one of the blocks posted yesterday and the one after checkup :

http://www.mediafire.com/?3ba707udk1dacvk

Trace of carbonization inside the heat block

 

About speed and volume, I printed a 20mm cube at 0,4mm and 200% of 50mm/s, it showed no underextrusion while the same cube with other settings did show underxtrusion (earlier tests).

 

With the new tube I can push filament very easily, with the original one it was quite hard and little less hard when there was no markings.

Molten plastic came out of from above and below the main hole of the heat block. It happened only once before during the first print after startup.

 

[gr5 2,172]

With the new tube I can push filament very easily, with the original one it was quite hard and little less hard when there was no markings.

 

So does that mean your prints are much better now and you are all set? are these pics from before or after the tube change?

That second picture in your most recent post of pictures looks like classic underextrusion. It's postcard perfect.

You mention "skin". Does that mean you have the skin setting turned on? I stongly recommend you don't for two reasons. One, it has a bug that can cause underextrusion, and secondly, the two passes are the exact same path. Which is crazy but so that it doesn't collide with infill.

 

[illuminarti 18]

Try printing the same thing at half speed, and see if that helps. If it does, then your under-extrusion is speed/pressure related. Printing at 100mm/s with 0.2mm layers is marginal to begin with. If slower makes it better, take a look at the assembly of the extruder drive, and make sure everything is good. Maybe post some pictures of the teeth marks on your filament (outside of the Bowden).

Looking at the pictures, the bottom surface is much better than the top in general. This is because, as noted above, infilling at 100mm/s over the top of only 5% infill isn't going to work well. You need to have more infill for the surface to rest on and/or print more than 3 layers of solid to get a decent finished surface.

 

[illuminarti 18]

Additionally, I'd invest in some top notch filament from a reputable supplier and see if that makes a difference - the dimensional variability you posted earlier certainly isn't going to help.

 

[corngolem 0]

So does that mean your prints are much better now and you are all set? are these pics from before or after the tube change?

 

From after and the prints are not better.

 

Try printing the same thing at half speed, and see if that helps. If it does, then your under-extrusion is speed/pressure related. Printing at 100mm/s with 0.2mm layers is marginal to begin with. If slower makes it better, take a look at the assembly of the extruder drive, and make sure everything is good. Maybe post some pictures of the teeth marks on your filament (outside of the Bowden).

 

Why would 0,1mm layer height at 100mm/s have no underextrusion and 0,2mm layer height would, isn't it the same amount of plastic squashed more or less thin ?

I printed the same 80x80x20mm block at 0,2mm layer height and 50mm/s : only few layers on one wall showed underextrusion.

The same at 0,1mm layer height and 100mm/s : some underextrusion on one wall, very little on another and none on the remaining 2, underextrusion on the entire infill and some on the top.

To show the markings I would have to turn it back on etc and I don't wanna do that now.

I mentioned before that the model of my UM is revision 4, maybe its feeder just doesn't work properly, who here has experience with it ?

 

[illuminarti 18]

If you're printing at the same 100mm/s linear head speed, then a 0.1mm layer height is extruding half as much plastic per second than a 0.2mm layer height. The fact that reducing the volume seems to be producing better results makes me think that probably the problem, to the extent that there is one, lies in your extruder set up.

Are you locking the extruder pressure plate properly, with the lever on the bolt all the way at the bottom of the slot, and the lever rotated past horizontal? How long is the compressed spring on the back of the extruder?

 

Why would 0,1mm layer height at 100mm/s have no underextrusion and 0,2mm layer height would, isn't it the same amount of plastic squashed more or less thin ?

 

 

[illuminarti 18]

The infill is vulnerable to under-extrusion because it's only printing every other layer - the stripes go one way on one layer, and at right angles on the next. As a result the infill is mostly printing into thin air, except at the crossing points. As such, you're going to get poor infill when printing fast and/or with thin layers - again it's a failure to bridge properly, just as with the tops.

The problem with the top - as I've said a bunch of times now - is that the infill is too sparse, and you aren't printing enough solid layers to cover it up well. Make the top and bottoms thicker, or increase the infill percentage.

In general, when you take those things out of the equation, it sounds like you're actually getting fairly good results at slower speed - so again, double checking the alignment and tension of your extruder will probably get you the rest of the way.

 

[gr5 2,172]

Why would 0,1mm layer height at 100mm/s have no underextrusion and 0,2mm layer height would, isn't it the same amount of plastic squashed more or less thin ?

 

Did illuminarti's answer claify this? It's really obvious when you spin the extruder gear by hand. Feel the force against you at 190C and then at 240C. Huge difference. Time it - do a complete spin and see how fast you can do it. Now watch how fast that thing prints when you are doing 100mm and .2mm layers! Then it all becomes clear -- the pressure at 100mm and .2mm layers is pretty impressive. I can't do it by hand - it's too much. When printing I can do that and much faster but only at 240C and if I'm willing to live with a little underextrusion.

 

[corngolem 0]

The feeder's mechanics (not the extruder) seems as it is supposed to be, the spring size is the same as mentioned before.

I don't see any problem with the top other than underextrusion and strings getting caught between layers.

If 0,2mm at 100mm/s causes underextrusion I wonder why 0,4mm at 50mm/s doesn't.

I tried PLA from reprap.me but it simply doesn't stick to the tape or the acrylic plate, so I'll do more tests when I have found proper cheap filament (and when I have time).

 

[gr5 2,172]

If 0,2mm at 100mm/s causes underextrusion I wonder why 0,4mm at 50mm/s doesn't.

 

I've never printed that thick. Have you?

I have printed .2mm at 100mm/sec just fine but it's getting to a point of mild underextrusion. I only print that kind of volume at 240C. But if I need to print at cooler temps I don't go that fast.

 

[pablobell 0]

If 0,2mm at 100mm/s causes underextrusion I wonder why 0,4mm at 50mm/s doesn't.

 

The flow depends on the difference in pressure between the inside of the nozzle and the outside. If the nozzle is put closer to the previous layer (from 0.4 to 0.2mm), the melted PLA finds increased ressistance outside the nozzle, which increases the pressure outside the nozzle, and reduces overall flow. It's like the difference between extruding a big blob in mid air and spreading the same amount of material over a long distance. The latter would have more ressistance.

Or at least that´s what I think.

Also, in theory when printing a wall at 0.4mm layer size the newly deposited PLA would have a circular cross section and touch the lower layer only in the bottom center. In reality, the melted PLA will be pulled by gravity especially if very hot, so it will deform and accumulate in an oval rather than circular cross section. So actual maximum layer size is 0.32 mm (80% nozzle size).

And about the path optimization problems, I agree with you. The slicers are not too smart in planning the shortest print head paths for the shapes involved. I have seen laser cutter paths that are very well optimized, even having in account the head's inertia in each axis. The slicers have to calculate 500-3000 times more layers than a laser cut but not take that much longer. Routing is in itself a very hard problem (https://en.wikipedia.org/wiki/Traveling_salesman_problem). Also there are many variable involved, like keeping the head inside the boundaries of the shape, raising the Z at the same point, etc.

 

[corngolem 0]

I noticed that if when underextrusion occurred I pushed the filament into the feeder (which required strength) I could reduce underextrusion. I interpreted this as the nozzle pressure required being too high for the torque of the feeder motor, aggravated or not by the bowden system, reducing the force of the motor compared to if feeder, filament and nozzle were on a straight line.

Then recently I tried some new PLA (which I'm pretty sure is not ingeo) that has the same diameter as the default setting (2,89mm). I decided to tighten the feeder's bolt (to spring length of 13,48mm) and now at 100mm/s underextrusion is very sparse. I didn't do it before because the instructions said to stop when it got harder to screw and it happened "3mm before".

 

[gr5 2,172]

My spring is at 11.5mm. But maybe we have different springs.