Ultimaker 3 not calibrated?

[ATS_DEMUC 0]

16 hours ago, gr5 said:

Okay well that support up the middle of the tube is useless and can get in the way.  I'd use a support blocker to remove that.  And I'd also add some horizontal expansion to the PLA support.  I'll create a project file (based on your file) and post it here.

 

done.  I used 1.5mm for the horizontal expansion and had to block a little bit on all those semi-disk protrusions.

 

 

The file you attached have this "eraser box" (that I've never used before), and I assume it's to erase PLA support. And it works, but since it's not aligned with the buildplate (in Z axis), I still have a bit of PLA inside (attached picture).

 

Why not lower the box (and/or increase its size to have the same height of the tubing) to completely remove the PLA from the inside (last attachment)? Is it because I'll lose support (structurally speaking)?

[ATS_DEMUC 0]

16 hours ago, gr5 said:

Another idea. 

I tilted it 45 degrees to reduce the need for support. 

Used tree supports to hold up some of it (it would otherwise fall over).

Set support angle such that it only supports a few places

 

I would never, ever, do something like this. I don't have that much faith in this printer. That seems a crazy amount of overhangs with PVA (which generally sucks) and even PLA. But if you do, I can try!

 

17 hours ago, gr5 said:

You should look into aquasys as well.  I hear that behaves much better than PVA.

 

Forgot to answer this. I bought Innovatefil Hydrosoluble for that reason. But unfortunately, it's much more expensive, and only found 1.75mm. So just gonna try it in the Prusa. 

 

The advantage of using the UM3 is that I have different materials for piece and support, which greatly increases the quality of my print. Because, as you said, PVA is not very solid, and breaks easily., which might be a problem when I start detaching the supports

Edited May 18, 2022 by ATS_DEMUC

[gr5 2,170]

2 hours ago, ATS_DEMUC said:

Why not lower the box (and/or increase its size to have the same height of the tubing) to completely remove the PLA from the inside (last attachment)? Is it because I'll lose support (structurally speaking)?

Oh - I didn't notice any support in the bottom.  It's fine to make it reach the bed.  That's a good idea.

 

So support blocker doesn't work exactly the way you might think - if you look at your part in PREPARE mode you will see red areas which is where Cura thinks you might need support.  The support blocker needs to cover up those spots.

 

Support blocker doesn't keep support from appearing in it's volume.  It blocks certain areas of the model from needing support.  So usually I don't need the blocker to reach the print bed.

 

Why cura thinks the inside of the tube has overhangs is a mystery to me.

[gr5 2,170]

2 hours ago, ATS_DEMUC said:

I would never, ever, do something like this. I don't have that much faith in this printer. That seems a crazy amount of overhangs with PVA (which generally sucks) and even PLA. But if you do, I can try!

Well I don't usually print PVA with overhangs but with PLA I can do 45 degree overhangs - and well - much much steeper.  Maybe 70 degree overhangs.  They start to get ugly though.  But 45 degree should be nice looking overhangs.  If it's PLA.

 

I don't know about PVA.  But 45 degrees is really no big deal.  I would think PVA can easily print at 45 degrees and much steeper.

 

In fact that 45 degree version should print fine with zero support.  But...  It's so small - it's like a straw.  It will wobble.  It will vibrate.  And it will be delicate.  So I added some tree support.  Really I would probably add support blocker and not have it touch those... bulgy intermittent disks at all.  yes it's steeper than 45 degrees briefly there but I expect it will still print fine.

 

Truly I would try the "straw" sideways.  And block support within the straw by changing support from "everywhere" to "touching build plate".

[ATS_DEMUC 0]

37 minutes ago, gr5 said:

if you look at your part in PREPARE mode you will see red areas which is where Cura thinks you might need support.  The support blocker needs to cover up those spots.

 

Gotcha! Interesting tool. Will come in handy

 

37 minutes ago, gr5 said:

Why cura thinks the inside of the tube has overhangs is a mystery to me.

 

Right? it doesn't make any sense to me either.

 

32 minutes ago, gr5 said:

Truly I would try the "straw" sideways.  And block support within the straw by changing support from "everywhere" to "touching build plate".

 

I'm going to try your first slice, the vertical one. If that doesn't work, I'll try to print the 45 degrees one.

 

I've heated the PVA for ~20hours, hopefully it's in a GREAT shape now 😅

[tomnagel 126]

My approach would be to first be sure that the printer and cores are in good shape, and *only then* start printing difficult print jobs. These straws are challenging prints with PVA I think, cause there is very little material on each layer.

My advice would be to 
- print the layershift model that I posted above here
- with the AA core and the BB core

- with 2 colors of PLA

 

With this experiment, we focus on the printer hardware (incl calibrations), and we take out 1) the PVA (with all its difficulties and challenges) and 2) the challenges of this printed part

(by the way: it's perfectly fine to print PLA with a BB core, allthough the printer and Cura will not recommend it (don't do the opposite; don't print PVA in a AA core)

 

Just my 2 cents, I'm also perfectly fine with you guys going forward as you are.

[ATS_DEMUC 0]

I don't like what I'm seeing.

 

There's small bits of filament gathering in one side of the "tube" and on the opposite side, it's lacking material (you can see a small dent in the video). 

Why would such thing happen? Why is PVA behaving like this? It was dried for 20hours, and it has been open (out of vaccum) for only 2 days or so. 

 

Plus, I'm afraid that line of PLA (in the direction of the core switching point) will mess up something sooner or later.

 

Printing video: https://streamable.com/1u6gve

[ATS_DEMUC 0]

5 minutes ago, ATS_DEMUC said:

it's lacking material (you can see a small dent in the video). 

 

Well, it's becoming a BIG dent (attached picture).

I'm going to cancel the print. If I can't have a tube (closed everywhere besides the top and bottom), it's un-usable.

 

40 minutes ago, tomnagel said:

My advice would be to 
- print the layershift model that I posted above here
- with the AA core and the BB core

- with 2 colors of PLA

Maybe you were right, and I need to check the state of the printer first (?)

Edited May 18, 2022 by ATS_DEMUC

[ATS_DEMUC 0]

Looking at the slice, the "dent" is in the spot where the PVA starts and stops extruding. So.. is he creating a gap? why isn't the first point of deposition right next to the last point of deposition? I just cannot understand this behavior @gr5

[ATS_DEMUC 0]

There is a clear discontinuity between the first layers and the following ones. This discontinuity is noticeable even in the PLA, which makes it even weirder.

In the first layers, the PVA surfaces are smooth. But then that weird stringing (it's not really stringing, it's the nozzle overflow that gets stuck to that side for some reason, I guess) starts happening, the surface quality massively decreases, and the "dent" starts happening. 

 

What the hell happened to create this discontinuity? The piece even broke around that discontinuity when I was inspecting it (attached picture). I wish I could provide better quality photos, but this is just to small to focus.

 

EDIT: It's so weird (I'm repeating myself, I know) that it only starts to happen after some layers printed. The surface quality is so much different before and after the discontinuity that it seems that the PVA lost all the quality from that specific moment (it even looks like it got humid, or over-heated in the nozzle after that specific layer)

 

 

Edited May 18, 2022 by ATS_DEMUC
added a thought

[gr5 2,170]

Please add the prime tower feature.  Search for it in the features.  without the tower you have underextrusion  That "filigree wall" you have that runs across the build plate to where it changes nozzles.  That is all lost filament that means it will underextrude when it starts printing again.  You can fix that by adding a prime tower.  The default settings for the tower are all fine.

 

I think that "dent" you describe is just underextrusion.  When it starts printing again there is not enough filament in the nozzle.

[gr5 2,170]

Oh hell!!!  I just realized that was a video!  lol.  Your travel speed is much too slow.  You want 150-250mm/sec for travel moves.  Please fix that.  Having a slow travel speed reduces quality drastically and is causing some of that leaking issue I mentioned (but again a tower would help even more).

[ATS_DEMUC 0]

10 minutes ago, gr5 said:

Please add the prime tower feature.  Search for it in the features.  without the tower you have underextrusion  That "filigree wall" you have that runs across the build plate to where it changes nozzles.  That is all lost filament that means it will underextrude when it starts printing again.  You can fix that by adding a prime tower.  The default settings for the tower are all fine.

 

I can add a prime tower. But the last time I did it, both filaments got entangled on the tower somehow, messing up the print.

 

16 minutes ago, gr5 said:

Oh hell!!!  I just realized that was a video!  lol.  Your travel speed is much too slow.  You want 150-250mm/sec for travel moves.  Please fix that.  Having a slow travel speed reduces quality drastically and is causing some of that leaking issue I mentioned (but again a tower would help even more).

 

I lowered the speed to improve quality, I alredy wrote about it on a few posts above. But my print/printer is so messed up that too many problems arose so far

[ATS_DEMUC 0]

On 5/17/2022 at 6:26 PM, ATS_DEMUC said:

I slowed down the travel speed because I noticed that was causing me unwanted dragging. Since the "tube" is, as you also said, "too skinny", the infill isn't uniform (see the green material on the attached picture). And in-between the deposition of the green parts, the high travel speed makes that it doesn't deposits the material properly, and drags it. If you look closely at the previous sent pictures, the brim always look nice, because it is only affected by "print speed".

This "travel speed" change enhanced my print, but it still isn't ok. I still cancel it after ~150 layers because the tubing isn't uniforms, and lacks material.

 

 Here, this is why @gr5

[ATS_DEMUC 0]

13 hours ago, gr5 said:

You want 150-250mm/sec for travel moves.  Please fix that.  Having a slow travel speed reduces quality drastically and is causing some of that leaking issue I mentioned (but again a tower would help even more).

 

I made another slice, with the tower and 150mm/s travel speed.

 

I thought I understood how the eraser block works, it turns out that I don't. 

I've placed the eraser block all around the "straw", Cura still places support on the inside (image and file attached)

 

EDIT: Nevermind, I'm stupid. The block has to go all the way up, and I missed a bit of the straw. Slice done, going to print

 

 

UM3_actuator_v4C_core.3mf

Edited May 19, 2022 by ATS_DEMUC
I'm not smart

[ATS_DEMUC 0]

14 hours ago, gr5 said:

Having a slow travel speed reduces quality drastically and is causing some of that leaking issue I mentioned (but again a tower would help even more).

 

Unfortunately, as I expected that didn't help. The tower doesn't look good either. 

[gr5 2,170]

tower distance

 

Printing slow definitely helps quality but slow travel means more leaking.  I guess I would put the tower as close as possible to the print.  The longer the travel time between the tower and the print, the more leaking that can happen.  There is an X and Y position for the tower.  Keep both print and tower near the middle of the glass where leveling is more accurate (glass tends to warp downwards at the rear corners and you can get parts that don't stick as well due to less squish on the bottom layer).

 

stringing

 

There shouldn't be any plastic lines between the print and tower.  We usually call that "stringing".  That's a big flag right there.  However I think it's only caused by the PLA and if the quality of the PLA is good enough then it's fine.  But if the PLA quality is worse - if you have that "dent" in the PLA then you can make it retract more by checking this for core 2: Limit support Retractions

 

In cura PREVIEW mode, you can click on color scheme just below where it says PREVIEW and click the checkbox that shows travels and light blue are retracting moves and dark blue are non-retracting.  It would be better if they were all light blue from tower to part.

 

layer height

 

I'd try 0.2mm layers (versus current 0.1).  It's just more solid and you are extruding more and the print will go faster as a side benefit.  I think it might improve the quality.  not sure.

 

retraction distance

 

it's only retracting 4.5mm on the PVA instead of 6.5mm (which it is doing on the PLA).  This is probably supposed to save some time as usually we care less about the quality of the PVA?  Not sure. I'd set that to 6.5.  You want the exact amount of retraction that removes all pressure in the print head without actually pulling upwards on the filament (thus pulling air into the nozzel potentially).  That should be about 6.5mm for most filaments.  for shorter bowden tubes - those printers have less retraction.  For longer bowden's like the S5 you want longer.  I'm not sure what the ideal length is for UM3 but I trust PLA profiles more than PVA.  I don't think PVA is any less springy so I would think retraction should be the same.  I'm not certain.

 

print speeds

 

I hate speed changes.  This causes under or over extrusion.  Slow speeds are great.  20mm/sec should be plenty slow enough.  But you have some things going at 12mm/sec.  And the tower is at 6mm/sec.  This is bad as it will then speed up soon after which will have underextrusion for between 2mm and 10mm of what comes next.  Which is most of your part!

 

So set all speeds to 20mm.  No need to go slower.  Or set them all to 15mm/sec but 20mm/sec should be fine.

 

Except travel speeds of course.

 

travel speeds causing vibration

 

Okay this can be a problem.  I see those green dots you show in the image you posted above in cura.  Those dots create extra vibrations.  We can get rid of those.  Usually not a problem but with your very skinny part it may become a problem as you get higher off the print bed.

 

So I reduced those drastically by setting your line width to 0.4mm (which is really recommended for a 0.4mm nozzle - Ultimaker has reversed that decision but doesn't want to touch the profiles for good reasons as they might break something for a few rare people) and wall width to 0.8 as you can see here:

 

Alternatively play with cura 5.0 which does a better job of this exact thing.  The primary change in Cura 5.0 is that it does a better job of thin walls like this.  I think it will even do one pass walls (not certain)!  There are a ton of things to play with and I haven't used 5.0 yet but you can reduce of those vibrations.  Again - shouldn't be a problem with the bottom 2 cm of your part.  I think that with default settings 5.0 will probably do a better job but you can play with line width, wall width, "print thin walls" and "fill gaps between walls" and who knows what new parameters there are.

 

 

 

[gr5 2,170]

One more thing.  Disable (uncheck) these:

enable acceleration control

enable jerk control

 

These might vibrate things a little more.  unchecking these is recommended in engineering profiles and this is definitely an engineering kind of object.  These two options, when checked, decrease jerk and acceleration which means the printer goes slower on corners or any sharp direction change.  The slowdown causes localized over extrusion on said corners.   You will get better quality prints with these unchecked.  You will also get jerkier vibrations on the printer but I think the difference is pretty small.

 

You will also get ugly "ringing" or shadows when printing for example a large cube with writing on the side.  accel and jerk control reduces ringing.  That's it's whole purpose.  It has to do with the harmonic vibration frequency caused by the weight of the print head (and gantry) and the spring constant of the belts.

 

accel and jerk control enabled is a big part of why quality went down on the UM3 versus UM2.  UM3 has a heavier head and to compensate for ringing they added this feature but it hurts dimensional accuracy.

[ATS_DEMUC 0]

1 hour ago, gr5 said:

One more thing.  Disable (uncheck) these:

enable acceleration control

enable jerk control

 

1 hour ago, gr5 said:

tower distance

 

stringing

 

checking this for core 2: Limit support Retractions

 

layer height

 

retraction distance

 

print speeds

 

travel speeds causing vibration

 

 

 

 

My hope is gone..

 

Lack of adhesion on the straw? The tower looks really bad as well.  

Is it because I completely removed the infill in between walls?

 

EDIT: ticked "fill gaps between walls" and sliced it. Let's hope for the best

 

 

Edited May 19, 2022 by ATS_DEMUC

[gr5 2,170]

The PVA never even gets started does it.  The PLA looks okay.

 

PVA does not print well on top of PLA.  I've had the exact same problem you have above with the tower.  I don't know what I did to fix it. But I definitely got it working. I probably dried the PVA and I may have done the Z calibration (aka leveling) more carefully.

 

I only do manual leveling.  active leveling can get confused by any plastic on the nozzle tip.

 

When I do manual leveling, I skip skip skip skip through the "1mm above" part.  Then do any sheet of paper for the rear - I adjust a few times up and down and feel the paper carefully - the resistance - I try to memorize that.  then I do the front 2 screws pretty quickly and then when doing core #2 I try to get the same Z resistance.  You want it perfectly at the same height as the other core.  This is critical.  I think I get it within about 0.03mm but not sure how to prove that. (not 0.3mm - not 0.1mm but 1/3 better than .1mm).

 

When it starts printing the first time after leveling, I often turn the 3 z screws CCW a bit (as seen from below) to move the bed up just a little.  Maybe 1/4 turn which would be 0.5mm/4 or about 0.1mm.  This makes the bottom layer squish more.  Instead some people set the initial layer flow to 150% or 200%.  Squishing harder into the glass makes a big difference.

 

Back to the PVA - have you ever noticed steam coming out of the PVA nozzle?  Is it pretty quiet (no crackly, popping boiling sounds) while printing PVA?

 

I would change your model to have a solid - filled in bottom.  The bottom of the "straw".  Just so it sticks to the glass better.  So it has more surface area.  Also in the model I would add about 2mm of "brim" for the bottom layer only.  Increase the diameter.  This will help it be more stable and stick better.  You can cut the bottom of the straw off after it is done printing.

 

In summary - this is a difficult print.  Again - I'd print it sideways.  Printing a vertical straw or a vertical pencil is one of those very difficult prints (there are a couple other's).  More difficult than 99% of models that people tend to print.

 

One more thing.  Adding wood glue mixed with water to the bed with a paint brush and then drying - this makes a HUGE difference.  wood glue is mostly PVA I believe.  Here is a video of me mixing and applying (but this won't help your PVA in the tower issue):

https://youtu.be/t58-WTxDy-k?t=595

 

[ATS_DEMUC 0]

13 hours ago, gr5 said:

The PVA never even gets started does it.

Barely. He made the first layer, but the next ones didn't stick to the former.

 

13 hours ago, gr5 said:

I only do manual leveling.

 

Going to try your step-by-step levelling.

 

13 hours ago, gr5 said:

have you ever noticed steam coming out of the PVA nozzle?  Is it pretty quiet (no crackly, popping boiling sounds) while printing PVA?

 

This PVA specifically, no. As it is a new one, no noises or fumes. The old one, yes, a lot of crackling.

 

14 hours ago, gr5 said:

Again - I'd print it sideways.

 

Like the file attached?

 

14 hours ago, gr5 said:

I would change your model to have a solid - filled in bottom.  The bottom of the "straw".  Just so it sticks to the glass better.  So it has more surface area.  Also in the model I would add about 2mm of "brim" for the bottom layer only.

 

I tried this, it's going to print now. 

About the glue, I'll do it if I really need to.

 

Thank you so much for the help so far, and for not giving up

UM3_actuator_v4E_core.3mf

[ATS_DEMUC 0]

The PVA is not adhering to the PLA on the tower. That's a problem, because once the 1st layer doesn't stay in place, all the others won't either. And then the PVA is continuing to get dragged on the tip of the nozzle, to the "straw" part, creating defects.

 

Also, sometimes when I print PVA, both on Ultimaker and Prusa, I feel like  I should be printing colder. The already printed layers are in a flexible consistency, and when the nozzle pushes down to print new layers, they just wobble around, making the deposition inconsistent, and the surfaces less attractive (or even messes up the print completely)

 

That was happening in the beginning of my print right now. So I have 2 problems - adhesion to PLA and PVA solidification.

 

How can I print PVA and make it solidify fast enough for the next layers have a solid "bed" to be deposited on?

 

Should I remove built plate heating? Would that mess up adhesion?

Should I lower the temperature to 190º from 210º?

Should I increase the fan speed to 100% from 50%?

 

Wouldn't all of that also impact the adhesion to PLA? I have to find a sweet spot, don't I?

 

 

[ATS_DEMUC 0]

11 minutes ago, ATS_DEMUC said:

So I have 2 problems - adhesion to PLA and PVA solidification.

 

But I must add, the lack of adhesion to PLA looks the main issue on Ultimaker.

 

As you might be able to see on the pictures below (ignore the warping, that was due to removing from plate while hot), there is no PVA on the tower whatsoever. Some of it is mixed in the PLA support structure (the C), the rest just got pushed off in every direction by the nozzles hitting it.

 

On the other hand, the few printed layers have high quality surfaces, better than previous prints. Which kind of surprises me, because I saw the bottom layers wobble a decent amount.

 

 

Edited May 20, 2022 by ATS_DEMUC

[gr5 2,170]

On 5/20/2022 at 8:05 AM, ATS_DEMUC said:

But I must add, the lack of adhesion to PLA looks the main issue on Ultimaker.

It could be the brand of PLA.  Do you have Nylon?  PVA and Nylon work quite well together.  I'm not 100% certain but I think they will stick much better.  Maybe too well?

 

You could design your own wipe tower.  Disable the tower and put in two rectangular towers.  I recommend they be 1/5th as wide as tall (which means they would be big) but they can be cone shaped which will maintain the "current layer is 1/5th as wide as from here to the bottom of the print" rule.

 

If you *do* do your own wipe tower you need to make damn sure it prints the towers first after a layer change.  From my experience you do this by dragging the towers into your project before the main part.  You can assign each tower to a different filament by right clicking and assigning the appropriate core.  the towers don't need infill.

 

On 5/20/2022 at 7:53 AM, ATS_DEMUC said:

How can I print PVA and make it solidify fast enough for the next layers have a solid "bed" to be deposited on?

 

Should I remove built plate heating? Would that mess up adhesion?

Should I lower the temperature to 190º from 210º?

Should I increase the fan speed to 100% from 50%?

I really don't know but if you want it to cool more you can just make the tower bigger which gives the main part more time to cool.  Doubling the layer time from 10 seconds to 20 seconds can really help.  I have found that 100% versus 50% sounds the exact same to me on my UM3 - I have to go down to around 20% before the fan even starts to slow.

 

You can experiment with all of those settings while it's printing in the TUNE menu.  So you can let it go for a few layers and then switch to a new temp or fan speed.

 

The ONLY reason you need a heated bed is to keep the part on the bed nicely.  You can do that without any heat if you use the right surface treatments (like wood glue and water).  You want it warm as it prints the first layer so the PLA and PVA get into the micro pores in the glass - otherwise the cold glass cools the filament too fast for it to flow into these holes.

 

Once it is done with the first layer you can lower the heat (for this print).  The other advantage of 60C print bed (or hotter) is that this is above the softening temp for PLA and can reduce warping forces that would lift the corners of your print.  But that is only a problem for larger prints like say a pack of cigarettes flat on the bed.  That large distance from corner to corner creates more lifting forces and warps the corners of the "cigatette pack" off the bed.  With 60C print bed the bottom few layers are above softening temp and act like clay and will spread the forces better and the parts stays down.  Not a problem for your print.

 

Oh and 40C is plenty hot for the first layer.  If you are trying to experiment with cooler air 40C is fine and then turn bed down more after first layer.

[gr5 2,170]

8 hours ago, ATS_DEMUC said:

Like the file attached?

Yes.  Notice it prints 4X faster (only one hour).

 

I would add a cylinder at each end of the straw (add to the cad file) to sit flat on the bed.  Like a penny on the end of this straw shown and another on the other end.  Then when it's printing the PVA on top of the PLA support it will make long long runs down the length of the PLA trough.  Held securely to the print bed at each end of the straw.  That way the PVA doesn't need to stick at all to the PLA.  It just needs to rest on top. 

 

Also having a cylinder shape flat on the glass means it will be rounded which means it will be less likely to warp off the bed.  When sideways like this you do want higher bed temps (60C) to avoid the part from warping off the bed at each end.  The lifting force is concentrated at the ends of the part.  Having it rounded spreads out the forces to more area.  More info about this topic in my video about warping parts.