shaking bearings

[xeno 108]

I'm having trouble making smooth large prints,

and after some investigation I found that the original bearings in the extruder head, don't fit properly.

both bearings wiggle ever so slightly, this results in a larger movement lower down,

they just don't fit well :(

I can imagine that making fast and sudden moves of the head would cause ripples in the print, like on the image

can I replace these with solid sliders, en where could I get them ?

 

[illuminarti 18]

Try reducing the x/y-jerk settings on your printer. That will reduce the speed that the print head carries through corners, and so reduce the shaking forces on the assembly. You can do it with the Ulticontroller (Control -> Motion), or via gcode:

M205 X15 ; set XY jerk to 15

If you go the gcode route, you'll need to do it every time (perhaps as pert of your start gcode) unless you save the setting to EEPROM with an M500 command.

The default jerk value is 25, iirc. Try smaller values and see if that helps.

 

[3dcase 37]

If you don't tell anyone!!!

It's just a greasemonkey solution.... But it works, and helps finish the android...

When the linear bearings of the printhead wiggle a little, you can put a little more tension on them.

Lift the lower 6mm shaft a little. Fold a piece a paper ca. 0.5mm and put it in between shaft and wooden sliders.

Lowering the jerk (accelaration?) seems more sophisticated.

 

[xeno 108]

If you don't tell anyone!!!

It's just a greasemonkey solution.... But it works, and helps finish the android...

When the linear bearings of the printhead wiggle a little, you can put a little more tension on them.

Lift the lower 6mm shaft a little. Fold a piece a paper ca. 0.5mm and put it in between shaft and wooden sliders.

I will try that for a temporary solution,thanks

Edit: But would that cause friction when the printhead moves to the sides ?

while lowering the speed would solve a part of the problem, but I don't want a 3day print turn into a week print

the bearing just don't fit properly, the wiggle at the bottom of the extruder assembly is more then 1mm

stupid that I never noticed it !, I wonder if all the UM1's have it ?

 

[illuminarti 18]

Lowering the jerk settings won't make much difference to your print time; it only affects the speed at the beginning and end of line segments that end in fairly sharp corners. Compared to the acceleration/deceleration settings, the change is tiny. But it will give less kinetic energy to dissipate at in the corners, and so should reduce shaking effects.

 

[xeno 108]

Lowering the jerk settings won't make much difference to your print time; it only affects the speed at the beginning and end of line segments that end in fairly sharp corners. Compared to the acceleration/deceleration settings, the change is tiny. But it will give less kinetic energy to dissipate at in the corners, and so should reduce shaking effects.

Unfortunately, I don't have a Ulticontroller :(

and I don't have the experience working in G-code, I wouldn't know where to put the code

 

[3dcase 37]

I will try that for a temporary solution,thanks

while lowering the speed would solve a part of the problem, but I don't want a 3day print turn into a week print

the bearing just don't fit properly, the wiggle at the bottom of the extruder assembly is more then 1mm

stupid that I never noticed it !, I wonder if all the UM1's have it ?

I don't no about all UM1's, but mine had it, hard to tell how much, it has to be as little as possible of course.

The wiggle disapperared after a while. The head can only wiggle when the shafts are perfectly aligned. that will allmost never happen i Guess.

At the end its just a wooden/lexan or alu. box with some electronics in it. :shock:.

 

[illuminarti 18]

Just add the line to the end of the start gcode in Cura.

You should take the time to learn basic gcodes. It helps get a much better understanding of what the printer is doing. Just load the code into a text editor, or into Repetier Host. The latter even has built-in contextual help that tells you what the codes do.

Here's the list I usually refer to to check up on codes:

http://www.ctheroux.com/2012/11/g-code-commands-supported-by-marlin/

 

[gr5 2,189]

I strongly recommend you read about these tests before you do anything:

http://umforum.ultimaker.com/index.php?/topic/2532-prevent-ringing-wobbly-surface-after-sharp-corners/

I recommend doing what that guy did and leaving your jerk a lone and cutting acceleration time in half. It will probably only increase your print time by 10% since your parts are so large (long straight lines compared to say a UM robot).

M201 X3000 Y3000

Just add that at the end of your "start" gcode. It sets max accel to 3000mm/sec/sec.

 

[xeno 108]

Could I just add both to the list of the startup code ?

M205 X15 ; set XY jerk to 15

M201 X3000 Y3000

or am I being silly :-P

@3dcase:

I added the 0.5mm underneath to add tension, but I don't feel any improvement,

I will probably start looking for replacements that have a snug fit, like the X-Y sliders, and maybe change to 8mm also ?

 

[illuminarti 18]

yes, you can add both

If your acceleration is currently more than 3000, it would make sense to reduce it to 3000, for sure.

 

[xeno 108]

I added both, and started printing, lets see what happens :cool:

I'll let you know tomorrow.

 

[nick-foley 5]

Yeah, this is a problem I've been trying to solve for months. It is hard to eradicate. I've tried several different pairs of linear bearings and none of them have been any less sloppy than the others.

 

[xeno 108]

Printed the head last night, but still had the same problem, even after reducing the speed to 65%.

I sanded the front to see if it was still usable, or I would sand away the detail to get it smooth.

it took a lot of work and still some lines still show very clearly.

 

[gr5 2,189]

First of all: reducing speed does not help "ringing" so that may have been a waste of your time. Reducing acceleration does help ringing. Ringing is one cause of *vertical* lines. Are you talking about vertical lines or horizontal? Because I see horizontal lines much more clearly than I see vertical (I see both).

Horizontal lines are typically caused by a Z screw or Z stage that isn't moving the same amount on every layer. So it might move .2mm most of the time but sometimes it moves .3mm and you get underextrusion and the next layer it might move .1mm and get overextrusion. Both of which cause horizontal "lines".

Also I see a horizontal line through the nose. That looks strongly like underextrusion. It could be caused by a clog or partial clog or more likely your filament might have gotten tangled for a minute (happens to me usually on a full reel of filament). Or the z stage may have dropped too far (e.g. .4mm instead of .2mm).

 

[gr5 2,189]

Xeno - the very first picture you posted showed VERY CLEAR horizontal lines but you mentioned "ripples" so I thought you ment the curved lines parallel with the circle pattern. Which is it you are talking about? Because all of the subsequent posts are talking about "ringing" which is COMPLETELY different than horizontal layering issues. All the advice about "jerk" and "acceleration" were advice to fix the ringing pattern. But now I wonder if you meant the horizontal lines?

 

[xeno 108]

I was indeed talking about the horizontal lines :eek:

and you saying it might be z-stage related makes sense somehow, but the hands I printed are perfect smooth and smaller then the other parts,

as soon as I print a larger part it shows the horizontal lines.

might be time to clean the z-screw and see if that maybe improves it,

I also have already added some weight to the z-stage, might need a bit more.

thanks

 

[illuminarti 18]

I imagine that larger parts mean that you are putting more torque on the bed, because you're printing higher up, and closer to the edges. Which is why I suggested lowering the jerk setting to reduce the shaking of the printer on direction changes.

You might also check belt tension, and make sure that there's not any backlash causing slight positional inaccuracy from one layer to the next.

Also, check the bed leveling screws, and make sure that they are tight, and have plenty of thread in the delrin blocks, so that there is no tendency for the whole bed to slide under the screws, or wobble around on the screws, as the forces on the bed get further up and out from the center.

 

[xeno 108]

I already redone the z-bed, and is very sturdy,

Belt tension is always a problem, what is a good tightness ?, I did re-tighten them recently.

this is what my revised bed looks like

 

[gr5 2,189]

I still think the horizontal lines are somehow related to your Z height control. I understand what Illuminarti is saying - how it could be shaking/swinging side to side and that would indeed create layers out-of-place. But that would be worse I think on the sides and not as bad a problem on the front and back of the part.

I think more likely your Z state isn't always moving .1mm when you ask it to. I would target the linear bearings the most. Unfortunately it is probably extremely difficult to debug this - I really don't know how. I guess I would get a dial gauge like the one pictured below touching the build plate then using pronterface lower the bed .1mm at a time and record from the dial how much the bed actually lowered and repeat 30 times and make sure it really moves .1mm every time. If not - something is sticky - like the linear bearings. You would need a gauge accurate to about .03mm because your error is occasionally quite large - I think the bed moves .15 instead of .1 sometimes and then later .07 instead of .1. I think the error is occasionally on the order of 30% (.03mm) which gives you 30% under or over extrusion for that layer which is enough to be visible when you look closely. On average I'm sure it is more accurate than 1% but some layers the bed moves too little, and some too much.

 

[gr5 2,189]

It's also possible that the bed bounces up and down when the head is jerking around. But this wouldn't cause an entire layer to be over or underextruded. It would show up on very short line segments.

 

[xeno 108]

The evidence points to the z-stage, as bouncing of the head would produce blobs on the outside,

and I see a few of those on the prints too, but the difference in layer thickness all around the object points to the z-screw not working right.

I have added some more weight to the platform and printed a part, and it looks better,

when I have the time I will thoroughly clean the z-screw parts and see what happens then