Need help with worlds largest ultimaker 3d printer!!

[kinginnovator 6]

Hi everyone! I have built an ultimaker printer with a print bed of 4'X4' 5.5' (z) (1220mmX1220mm) .... Its about 8 feet total vertical... Now I'm trying to  calibrate most everything on and having some trouble in some areas atm.. Issue one is after a z movement to go onto the next layer the extruder stops extruding and it will follow a total layer path equal to a single layer across the part with the extruder not extruding as if it were Ironing out the part (ironing is disabled). It will then turn back on and print that layer just fine.. What setting could cause this to happen? This giant format printer is not going to play by the rules on the smaller printers. So just keep this in mind. Settings are soooo different.. Any ideas are welcome! 

[GregValiant 1,115]

Post a picture of that monster.

There is no setting for a "Practice Layer".  Are you sure it isn't just a really long combing move?

[kinginnovator 6]

I could try and disable this? Retraction is off atm.. I'm dealing with  (2) 3mm  filaments that are simultaneously extruding into a dual channel setup with of course a single output head. It is definitely traveling along the entire part's surface as if yes printing a "practice layer"

[kinginnovator 6]

[GregValiant 1,115]

I think you just took the lead from "Godzilla".  That bad boy is pellet fed.  With your dual 3mm lines I'd guess you can retract.  From what I've gathered, that isn't easy on a pellet fed machine.

 

If you'd like you can post a Cura 3mf project file ("File | Save Project").  I'll take a look and I'm sure there are others here whose curiosity would be peaked enough to attempt to troubleshoot the "Practice Layer" problem.

[Torgeir 228]

Hi @kinginnovator,

 

Sure interesting, -but first question; what kind of the Ultimaker's models was your "template"???

This will be a good hint of where to go.

 

Thanks

Torgeir

 

[kinginnovator 6]

LOL! I can turn on retract just not ready for it because it's going to be a challenge because of the amount of thermal mass within this setup. I will make a video and post the 3mf project file by tomorrow. I have plans to make it with not one but yes two pellet heads so that's coming! But I'd rather nail down the path of least resistance with the (2) 3mm filament heads because so much more has been done on this side of plastic flow.  No reinventing the wheel quite yet my friend.. I truly need help tuning this, there are so many more variables. But with this setup I can extrude more than a pellet system atm (800 watt potential)  I'll add a video soon.. I really appreciate your help here..!!

[kinginnovator 6]

Torgeir. I will also supply the model too. Thanks for joining my lonely battle here! LOL

Edited July 14 by kinginnovator

[gr5 2,071]

project file please.  🙂

 

[kinginnovator 6]

Hey there! Here is a link to the video let me know if your able to view it. It is possible it's trying to perform an infill into tiny spots? Not sure. This is a flex shaft drive system on a work feeding 2 simultaneous extruders.  https://photos.google.com/photo/AF1QipMl9G8s5sFJpooJsKSTadyVOIFku2QRTwHhzJ2D

CFFFP_Test frame.3mf

[kinginnovator 6]

15 hours ago, GregValiant said:

I think you just took the lead from "Godzilla".  That bad boy is pellet fed.  With your dual 3mm lines I'd guess you can retract.  From what I've gathered, that isn't easy on a pellet fed machine.

 

If you'd like you can post a Cura 3mf project file ("File | Save Project").  I'll take a look and I'm sure there are others here whose curiosity would be peaked enough to attempt to troubleshoot the "Practice Layer" problem.

Here is a link to the video. Let me know if your able to view it. Flex drive setup on a worm gear. It's possible it could be trying to infill into tiny areas? Don't know... 90 percent infill because I tried 100 percent to try an eliminate such tiny movements instead of long continuous because the start/stop of this setup is very difficult to toon. Linear advance is very difficult atm.. 

CFFFP_Test frame.3mf

[GregValiant 1,115]

If you look at the preview you'll see there are small yellow dots at the corners and intersections.  All that movement is to put those dots down.  They are each around .2mm long extrusions so the nozzle doesn't spend much time on any of them.

You can eliminate the dots by changing your design so that the inner and outer radii of the corners are concentric.

You can also change the Combing Mode from "All" to "None" but without retracting that will leave a lot of stringing.

 

This is the difference when the corners radii are concentric.  That's the yellow dot on your model.

 

[GregValiant 1,115]

I get "link not found" on that video.

[kinginnovator 6]

https://photos.app.goo.gl/UGENqVYuz39u29gZ8

 

This makes a lot of sense! I will adjust most likely the radii.. I don't mind stringing on a giant part.. Least of my concerns, its different when you are simply trying to successfully print a large part.. Let me know if this link works

[kinginnovator 6]

Another question.. What do you recommend as a possible solution to this problem with under extrusion of single line starts?  This is a wall that is only 1 line path.. 

CFFFP_Test ring.3mf

[GregValiant 1,115]

There are two possibilities.  The Cura setting "Retraction Extra Prime Amount" which is by volume but it would only work if you are retracting.  Linear Advance would have the same problem, no retraction, no advance.

 

If you still have the retraction turned off then you are losing nozzle pressure during a travel move.  I don't see why that would happen on that particular model though.  A trick would be to print it in Spiralize so the nozzle never has to stop.  That won't work on all models of course.

[gr5 2,071]

There is a setting - something like "retract on layer change".  Having that checked could both cause this problem or having it unchecked could cause this problem.  So try it both ways.

[kinginnovator 6]

gr5 and GregValient,  I turned off retraction for this very problem because when you are extruding so much more material (think 10X the volume within the extruder heated pathway)  retraction is pointless at this point if you want to print 10X faster with so much more material. It is actually very responsive to not extruding when the extruder does not move. This type of setup would be similar to a pellet system. Or a Bowden that is 10X less responsive. Is there a way to not shut off the extruder completely in between stop/start moves,  like 1/2 movement or speed? I think this would definitely solve this. 

[GregValiant 1,115]

There is a post processor called "Retract Continue" that does the opposite.  It has an initial retraction and then continues to retract across the combing moves.

It sounds like you would need the opposite and continue to extrude during a combing move.  Does that sound like I'm on the right track?

 

I have no idea how much work it would be.  I am familiar with Retract Continue.  The script is only about 125 lines long.

[gr5 2,071]

Can you just pass "retract continue" a negative number?

[GregValiant 1,115]

That sounds like a plan.  When I slice that ring with retractions turned on I only get 2 retractions.  There might be something else going on.

• Enable Relative Extrusion to make things easier to read.
• Turn on retractions.  The Retraction Amount doesn't matter.
• Enable "Retract at Layer Change"
• Enable the post processor "Retract Continue" and set the retract amount to a negative value. (-0.10 to start).
• Enable the post processor Search and Replace
  • Search = G1 F(\d*) E(.\d*)\n
  • Replace =
  • Regular Expressions = Checked

So you get retractions at layer changes.  The Retract Continue ends up being positive amounts.  The Search and Replace removes all the retractions and primes.

 

You end up with something like this during travel moves...

;MESH:NONMESH
G1 F600 X379.388 Y378.324 Z1.20 E0.07000
G1 F4000 X378.854 Y378.854 Z1.20 E0.07524

 

Edited July 16 by GregValiant

[kinginnovator 6]

I can certainly try this.. I appreciate all your help here so much!! I wonder how fast this could react after a stop or end to a line path before starting a new line. I also wonder if I could adjust acceleration at the beginning 🤔  bit one for a split second while the extruder independently turns on and does not adjust itself relative to acceleration and speed? Just ideas.. if I get results from what you say I should try I will certainly let you all know.. This could also be part of a future problem to solve do to how much faster one could print with the idea of having more pre-melt on demand within the potential volume of a heated housing... 🤔 

[kinginnovator 6]

Here is another tough one to solve.. I'm somehow gaining movement in one direction which is causing skewing. I can print a perfect cylinder with no skew. But when I print that test part (square frame) I get a predictable skew based on the parts orientation. Obviously this eliminates all mechanical scenarios based on the ultimakers symmetrical design. I'm predicting some sort of movement within the program that is causing this. The ring is a perfect symmetrical print that does not have the start/stop wall features of the square frame.  The square frame has a  symmetrical design also just start/stop walls. But what could cause this skew based on path movement? 

[gr5 2,071]

So when you have layer shifting like that there are a few common causes.  All of them mechanical. 

 

You could be losing steps (full steps).  That is fixed by increasing lubrication and decreasing acceleration and jerk.  On a new machine you should be pushing accel harder and harder until you lose steps so you have an idea what the limit is.  Then back off maybe 50%.  Same with jerk setting.

 

Another cause is slipping pulleys or other slipping hardware.  So you can mark the shaft of every stepper and everything attached to that pulley to make sure nothing is slipping.  You mark the rod and pulley with a sharpie, take a phto, then do a print, then check your markings.

 

I know you say it's not mechanical because it moves with the part rotation. That just probably means there is some high-acceleration move related to (for example) changing layers or to the central bar where cura does 2X the speed or 2X the acceleration of typical moves.  The problem is still with the hardware, not with cura.

 

If there was a skew created by cura you would see it in the PREVIEW mode of cura.

[kinginnovator 6]

Thanks for responding.. I'm extremely aware of mechanical skewing. This is not it.. I can print a perfect cylinder over and over and over without any problems. I'm pretty sure if I orientate the part 45 it would go away based on the x,y moving at the same rate just like in a circle they would.  The problem is specific to movements based on geometry and how cura does the paths. I do believe there is a possible problem with acceleration somewhere and steps being lost or gained. My acceleration is set to it's lowest 80mm/sec.. I bet if I print a square part without that center geometry and just a frame it will print perfect. I will try this and get back to you. Jerk is off because I can print perfect cylinders..  How would changing layers be related to acceleration? After a layer change does it do something special in it's acceleration? What do you mean also by central bar where cura does 2X?  I've limited my acceleration withing cura to be around 120 for travels. I will print this at a 45 and see what happens. Will send pics shortly.