Infill calibration, May have an issue still

[cjdavis618]

I printed these today at the outer corners of the bed using the Zinfill testing model that Mhackney posted. Below you will see a photo with the Z axis at the top and then Y and X in the normal orientation. (X is on the left, lol). These are laid out just as they came off the printer for each axis.


I have an issue with one axis, but possibly 2. Where do you think the issue lies here? Thinwall shell looks great, measures out good, (although one wall is thicker than the rest. Think that is a level issue). Running at .90 extrusion multiplier.

Here is the prints hot off of the bed. Some small gaps, but not too bad. What do you think?

[img]http://i5.photobucket.com/albums/y176/c ... cf6ad5.jpg[/img]

I have better closeups of each if I need to upload them.

[mhackney]

It does look like Delta Arm Blues. Based on the gap pattern and the experiments I did, I would predict the Z tower joints are the ones that are stiff. But you should check all of them just to be sure. One thing I did not mention when I posted this was once I solved my problem I artificially tightened joints (by putting a blob of silly putty on the joints) and tested the fill patterns. The gaps typically appear OPPOSITE the tower with the stiff arms. In your case, it is pretty cleanly isolated so I would predict a single tower.

Let me know what you do find when you do them, it helps me refine predictions!

cheers,
Michael

[Broose]

While we're on the subject...
Those look better than mine.

The infill gaps on mine between the three locations look fairly similar, even though they are were produced on opposite sides of the bed and oriented at different angles as mhackney specified. I sanded and deburred and greased until each of the arms fell under their own weight from the u-joints in both directions. Could this be an extruder issue, or do I need to take the arms apart once more?

-B

[MorbidSlowBurn]

Broose- if your arms can fall straight down from gravity you may need to calibrate your extruder.

Check that when you send a feed command of 50mm it takes 50mm of filament. Measure this prior to the extruder. Then do polygonhell's extruder calibration. After you have done that try again and let us know what you get.

[cjdavis618]

I have looked over the machine a bit more and I think rather than more sanding on a joint, I think I need to fill one. Some of my joints came in with some mill marks on the side and I didn't file them smooth. I think the machine marks acted like a cutter and ate material off if the tower joints. After some use, one joint has about .002 of play from side to side and that column was also very tight on the belt. This column had the belt rubbing on the extrusion down one side so the bearings up top were not level and favored one side.


I loosened the belts up all around and it it made the machine much more smooth on the tower movement. About to run some more test prints and see what the result is.

Wish me luck...

[cjdavis618]

Found some interesting things with the Bowden tube and retracts.

I have been running retract testing to get rid of blobs between sections of extrusions. Z lift so far has made things worse when I use it so I have been focusing on retracts. Testing this retract thingi http://www.thingiverse.com/thing:22263.

I have been able to get my best prints on small items with 7mm retract and 300mm/s. But by placing lines on the filament, the retracts are working fine at the extruder end and are right on for distance. At the other end next to the extruder, the 7mm I retract is now closer to 1-2mm. I'm sure this is the tube stretching but dang what a difference.

What do you guys run your retracts at and what speeds? I can actually see the extruder mount bow during feeding to the hotend. Not loosing steps that I can tell though.

[Polygonhell]

I have been able to get my best prints on small items with 7mm retract and 300mm/s.

There is no point in trying to retract that fast, one it's clamped to much less than that in the firmware.
Two I suspect with ~590 steps/mm if it weren't capped in the firmware you'd likely exceed the arduinos ability to output a reliable pulse train.
I think the cap in marlin is set at 60mm/s I can't remember what I set it to in repetier, but it's similar.
With high retract rates it's also really hard to see the extruder skip, and it can do it intermittently.

[mhackney]

A black sharpi mark on an extruder gear can give a visual indication of retract, the mark should go back and then forward to where it started. If it doesn't go back all the way, you are missing steps and the extruder will eventually starve filament to the nozzle. You would back off on acceleration and retract speed. I run retract at 60mm/s and accel at 1250.

[cjdavis618]

I will drop that and try more prints then. Started using kisslicer due to the wipe feature and it does much better. I do have lines marked on the gears and even across the stepper shaft to make sure it doesn't slip there either. Infill looks better now that I loosened the belt also. Still have more tests to run but I ran out of weekend. Lol

Happy easter to everyone and bless you all.

Thanks for the help you all offer.

[cjdavis618]

More printing showed that my issues were getting worse.

Took the afternoon and fund a few things.

1. The nut on the hotend was working loose and causing the head to drag. Tightened that back down and it is solid now.
2. Mhackney, you were right, Z axis was still a bit tight as was one arm on the X tower. I chamfered the inside of the bearing holes to clear the material left by the lathe tip that turned them.

Note to other users. When they say to have the arm bearings freely moving, they mean it. If you take an arm, with a bearing on the other end and swing it from the bearing on top while it is hanging down, it should go back and forth like a good clock pendulum until gravity stops it with no friction. If you can feel any "Sticktion" (Nice term Mhackney ) you aren't there yet. In other words if you can feel the weight of the arms swinging through the bearing your holding and feel any tendency to pull the bearing between your fingers one way or the other, it needs more clearance on the arm. I found the worst part was the molded holes needed the chamfer on the face where the bearing pins will insert.

[mhackney]

Thanks forte feedback and it sounds like you are getting things sorted. So my predictave abilities for the Delta Arm Blues are gettting refined!

Very good point on the holes and chamfering them. This is an area that is probably overlooked by many builds.

[Broose]

Broose- if your arms can fall straight down from gravity you may need to calibrate your extruder.

Check that when you send a feed command of 50mm it takes 50mm of filament. Measure this prior to the extruder. Then do polygonhell's extruder calibration. After you have done that try again and let us know what you get.

I definitely had an extruder problem with slipping and filament starvation. It appears that at one point in the rotation of the roller it slipped quite a bit easier. It was especially apparent during the back and forth of the infill, I marked the filiment going into the extruder with a sharpie and could see it stalling periodically. I swapped out the drive roller bolt with one that was straighter, and added a couple more o-rings to lower the spring constant, so the force on the idler would remain more consistent. I am still having an issue to a much lesser degree here (click on the photo to get a closeup):

At the full feedrate of 30mm/s you can see some issues with the infill. Lowering the feedrate to 50% it mostly goes away.

In any case, most of the problem i described at the beginning of this thread was related to the extruder.

-B