Notifications
Clear all

Edges/corners poorly defined  

  RSS
max.i
(@max-i)
New Member
Edges/corners poorly defined

Hi everyone, I am lucky enough to have a Prusa i3 MK3 here in the office for a project we are working on. The problem I am having is the edges/corners are printing out too rounded for our purposes. When printing the same STL on a low end Delta printer, the edges are significantly sharper so I must be doing something wrong with the MK3.

Here is what I've tried:
1) Slow down printing to 50% via the printer LCD
2) Set M900 K value to 30 (this is the default)
2) Set M900 K value to 60
3) Comment out the M900 command in gcode (after power cycle)
4) Different types of filament, including the Prusa filament provided with the printer from Prusa Research
5) Calibrate XYZ
6) Calibrate PID
7) Updated firmware to 3.5.2-1999
8) Sliced with Cura 3.6
9) Sliced with PrusaControl 0.9.4_415
10) Sliced with Slic3r 1.41.3+win64
11) Verified belts are tight
12) Verified the extruder assembly is tight on axis
13) Calibrated extruder

Both the MK3 and the Delta printer have a 0.4mm nozzle, with 0.2mm layer height, same brand of filament. Regarding Slic3r, I have not modified any settings at all. Should be all pristine Prusa settings.

I have been using a calibration cube( https://www.thingiverse.com/thing:1278865 ) as a test since it prints quick. The differences are a little hard to spot in the following image, I tried to take the picture with a magnifying glass. It doesn't really help but it is more noticeable in person.
Top Left - M900 K60, Top Right - From Delta printer, Bottom Left - No M900, Bottom Right - M900 K30

Here is another picture of the actual part I am trying to print(I am not able to share STL at this time). It's a hinge with the center cylinder tapered to prevent the arm from popping off. The print from MK3 is rounded to the point that the arm will pop off where as the print from the Delta printer is more defined and as a result the arm stays on.

Red - MK3, White - Delta

Can anyone spot what I am doing wrong?

Posted : 07/03/2019 8:43 pm
--
 --
(@)
Illustrious Member
Re: Edges/corners poorly defined

X/Y position accuracy is what sets the squareness of corners. Any belt slop will deform corners. So I'd be looking at belt tension.

Position the extruder hard left, and bed far back. Use a 10lb fish scale, and lift at the middle of the lower belts until you see 2 pounds on the scale. The belt should be deflecting about 1/4 inch from it's initial position.

Posted : 08/03/2019 8:49 am
max.i
(@max-i)
New Member
Topic starter answered:
Re: Edges/corners poorly defined

It was difficult to get a completely accurate measurement but the deflection is very close to 1/4th inch at 2lbs. I tried to tighten the belts for good measure and I don't think I can physically get it much tighter. I was pulling about as hard as I could then tightening the tensioners as far as they go. Maybe with another person to help pull the belt, I could get it tighter but that seems excessive, the assembly guide certainly didn't imply that kind of tension. Due to the inherit design differences between the delta and MK3 printers, should I expect the MK3 to meet or exceed the precision of the corners of the delta?

Posted : 08/03/2019 10:55 pm
Mustrum Ridcully
(@mustrum-ridcully-2)
Honorable Member
Re: Edges/corners poorly defined

a simple 2 lb lead weight can be had from folks that sell things like free weights and wrist and ankle weights. or ask someone who loads their own ammo the molds that they poor reclaimed lead into create 2 lb bars. Then an simple hook to hang the weight from the x belt and a scale or digital calipers to measure the drop will give a repeatable number... the Y belt is a bit harder but you can elevate the printer between two platforms (chairs etc) and adjust it the same way this may be the cheapest way to an accurate number; since a good accurate scale such as those designed for adjusting triggers will cost quite a bit (unless you can borrow one from that chap casting lead bullets 😉 )

Posted : 09/03/2019 12:58 am
max.i
(@max-i)
New Member
Topic starter answered:
Re: Edges/corners poorly defined


a simple 2 lb lead weight can be had from folks that sell things like free weights and wrist and ankle weights. or ask someone who loads their own ammo the molds that they poor reclaimed lead into create 2 lb bars. Then an simple hook to hang the weight from the x belt and a scale or digital calipers to measure the drop will give a repeatable number... the Y belt is a bit harder but you can elevate the printer between two platforms (chairs etc) and adjust it the same way this may be the cheapest way to an accurate number; since a good accurate scale such as those designed for adjusting triggers will cost quite a bit (unless you can borrow one from that chap casting lead bullets 😉 )

This is about the approach I took. I took a bag and kept adding screws and nuts into it until it reached 2lbs according to a scale (+/- 0.1oz). I took two rulers. First one I laid across the top of the belt then took another ruler and measured how far down the belt went. Would have been easier with another pair of hands but I think it's close. I really don't think I could have gotten the belts much tighter without someone else bracing the printer while I pull(I think this is excessive). They are very tight now (it's subjective), tighter than what I see on youtube and such. I didn't see much (if any) improvement.

Posted : 09/03/2019 3:46 am
--
 --
(@)
Illustrious Member
Re: Edges/corners poorly defined

The 2 lbs@1/4" is about as tight as you want to go. It's near the rated bearing radial load limits on the motors. You can get by with less, but the belt application manuals suggest more than twice the tension applied here. I probably wouldn't go much less than 1/2" deflection at 2 lbs (belts touching).

Now that your belts are snug, print a test cube in vase mode and look at the corners. This is as tight as you can do with a 0.4 mm nozzle: 0.2 mm radius.
If not sharp enough, you can move to a 0.25 mm nozzle, and get the radius down to 0.125 mm. Any printer with a 0.4mm nozzle is limited to 0.2 mm, and thinking one printer does better is probably seeing other issues, like X or Y overshoot or mechanical ringing. Proper belt tension reduces these, but there are other factors: frame strength, extruder mass, acceleration profiles and stepper accuracy/resolution. Add the slicer interaction and you have a plethora of causes to examine to get optimal quality.

All I can add is that friends and colleagues who have a lot more 3D experience than me recommended the Mk3 over FDM printers costing fthree to our times as much. I had planned to buy a $3,000 printer. They said don't bother. Then again, maybe they don't like me? 😯

Posted : 09/03/2019 8:11 am
Share: