Notifications
Clear all

Mk3 vs Tarantula -- Ghosting  

  RSS
david.b107
(@david-b107)
New Member
Mk3 vs Tarantula -- Ghosting

Hello All,

I got my Mk3 kit on Friday and have been trying to tune it all weekend. I'm now on my 16th XYZ calibration cube.

After hours of tuning I can't get any better print quality out of the Mk3 than my Tevo Tarantula.

I've included some pictures comparing the prints from the two different printers. (I took two different pictures to see the cubes in different positions.)

These cubes were printed with Simplify3D. I tried to match all of the settings I could, but settings like retraction are quite a bit different between the printers. I did match the speed settings, perimeters, temp, and used same filament. Also, they are sitting side by side on the same table.

The ghosting is noticeably (and disappointingly) worse on the Mk3.

I've seen some discussion about the ghosting issues with the Mk3. But, I haven't really seen a true acknowledgement of it from Prusa.

Do you think the ghosting is mechanical or in the firmware? Any suggestions?

Thanks for the help.

Posted : 30/04/2018 4:01 am
Nikolai
(@nikolai)
Noble Member
Re: Mk3 vs Tarantula -- Ghosting

Hi David,

I can see two things.
1. If your test prints are done with the same filament, then you're printing with different temperature. MK3 cube is glossy, TT is more diffuse.
2. I made a screen shot of my cube because no ghosting is visible (for me) with a bare eye. On the screenshot, I can see it. The reason is the overshot corner. Because LA is deactivated, we don't get a real square cube. All corners are slightly round. So what you think is ghosting, is just a shadow.

Extrusion and LA should work better with 3.2 firmware (coming soon).

Speed up your printer to 200mm/s and you will see how real ghosting looks like 😉

Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram

Posted : 30/04/2018 6:53 am
reid.b
(@reid-b)
Reputable Member
Re: Mk3 vs Tarantula -- Ghosting

I have noticed that Slic3r PE still has trouble with ghosting, even the latest beta. I am also running the MK3 RC2 firmware, which has Linear Advance enabled which will help with your corners (but won't help much with ringing or ghosting.) That said, Simplify 3D did a much better print with nearly identical settings, and it did it 6 minutes faster. Below, Slic3r PE on the left, Simplify 3D on the right.

EDIT- I noticed you did these with Simplify3d- Not sure why you are having that ghosting problem, other than you may have printed it real fast? Also, you might want to turn down your extruder temp 5 degrees or so.

Posted : 30/04/2018 7:53 am
RetireeJay
(@retireejay)
Reputable Member
Re: Mk3 vs Tarantula -- Ghosting

Just a thought here... I tend to doubt that the slicing program "creates" ghosting as part of the G-code. In fact, I would venture to say it definitely does not.

If you look closely at the printed parts provided by Prusa as part of the construction of the machine, you don't see ghosting on them. These parts were printed on Prusa printers (presumably MK3's). So the machine design, when used properly, is capable of doing prints without ghosting.

So that leaves us with two possibilities:
(1) the SETTINGS used in the slicer are not optimum for the machine, resulting in this artifact. As noted above, a prime candidate for examination in the settings is speed.
(2) the MACHINE is not assembled perfectly. Perhaps the belts are not tensioned correctly. Perhaps there is some part of the assembly that is not rigid, or is not aligned correctly that allows the machine to vibrate mechanically. I have been using a much older non-Prusa printer made of plywood for several years, and ghosting has always been a problem that I have attributed to the lack of mechanical rigidity on that machine. My new MK3 is very rigid and produces beautiful prints with no ghosting.

Posted : 30/04/2018 5:13 pm
david.b107
(@david-b107)
New Member
Topic starter answered:
Re: Mk3 vs Tarantula -- Ghosting

Hello All, thanks for the replies. I did try the same print with Sli3er PE and the ghosting was much (*much*) worse. Interesting observation about the glossy vs diffuse. I'll try lowering the temp on the Mk3 and try the same print again tonight. I had the temps matched in the settings -- my Tarantula temp readings must be a bit low.

@RetireeJay: Can you please list the settings you think are the most critical to create your beautiful prints? I'll compare to what I'm using. I printed the ones in the picture at 60 mm/sec with 40% underspeed for perimeters. The same for the Tarantula. What speeds are you printing at?

How tight are your belts? How do you "measure" them -- the "guitar string" method? The "belt status" on my Mk3 shows 271 for both X and Y. If I understand correctly, 280 is the loosest within spec number Prusa recommends. I had to make the belts on the Mk3 much tighter to get the 271 number. I had them about as tight as the belts on my Tarantula at first and "belt status" was 298.

As for your comments about rigidity: I agree generally speaking. But I was able to build my $250 Chinese Tevo Tarantula and get good prints. Is the construction of the Mk3 so much different that the approach I took when building the Tarantula wouldn't apply to the Mk3? Do you have any pointers?

Thanks for the help.

Posted : 30/04/2018 8:29 pm
RetireeJay
(@retireejay)
Reputable Member
Re: Mk3 vs Tarantula -- Ghosting

Here is a copy of the "0.15mm OPTIMAL MK3" settings I used - out of the box - when I printed a strain relief holder for the wires coming out of the X axis motor. The part I printed has the same quality look as the printed parts directly from Prusa. You can get the settings from Prusa's website when you download their "drivers" package. The filament settings were Prusa's default for "Hatchbox PETG" printing at 230 for the first layer and 240 thereafter.

# generated by Slic3r Prusa Edition 1.39.1-prusa3d-win64 on 2018-04-23 at 10:41:24
avoid_crossing_perimeters = 0
bottom_solid_layers = 4
bridge_acceleration = 1000
bridge_angle = 0
bridge_flow_ratio = 0.8
bridge_speed = 30
brim_width = 0
clip_multipart_objects = 1
compatible_printers =
compatible_printers_condition = printer_notes=~/.*PRINTER_VENDOR_PRUSA3D.*/ and printer_notes=~/.*PRINTER_MODEL_MK3.*/
complete_objects = 0
default_acceleration = 1000
dont_support_bridges = 1
elefant_foot_compensation = 0
ensure_vertical_shell_thickness = 1
external_fill_pattern = rectilinear
external_perimeter_extrusion_width = 0.45
external_perimeter_speed = 35
external_perimeters_first = 0
extra_perimeters = 0
extruder_clearance_height = 20
extruder_clearance_radius = 20
extrusion_width = 0.45
fill_angle = 45
fill_density = 20%
fill_pattern = grid
first_layer_acceleration = 1000
first_layer_extrusion_width = 0.42
first_layer_height = 0.2
first_layer_speed = 30
gap_fill_speed = 40
gcode_comments = 0
infill_acceleration = 3500
infill_every_layers = 1
infill_extruder = 1
infill_extrusion_width = 0.45
infill_first = 0
infill_only_where_needed = 0
infill_overlap = 25%
infill_speed = 200
interface_shells = 0
layer_height = 0.15
max_print_speed = 250
max_volumetric_extrusion_rate_slope_negative = 0
max_volumetric_extrusion_rate_slope_positive = 0
max_volumetric_speed = 0
min_skirt_length = 4
notes =
only_retract_when_crossing_perimeters = 0
ooze_prevention = 0
output_filename_format = [input_filename_base].gcode
overhangs = 0
perimeter_acceleration = 800
perimeter_extruder = 1
perimeter_extrusion_width = 0.45
perimeter_speed = 45
perimeters = 2
post_process =
print_settings_id =
raft_layers = 0
resolution = 0
seam_position = nearest
skirt_distance = 2
skirt_height = 3
skirts = 1
small_perimeter_speed = 20
solid_infill_below_area = 0
solid_infill_every_layers = 0
solid_infill_extruder = 1
solid_infill_extrusion_width = 0.45
solid_infill_speed = 200
spiral_vase = 0
standby_temperature_delta = -5
support_material = 0
support_material_angle = 0
support_material_buildplate_only = 0
support_material_contact_distance = 0.15
support_material_enforce_layers = 0
support_material_extruder = 0
support_material_extrusion_width = 0.35
support_material_interface_contact_loops = 0
support_material_interface_extruder = 0
support_material_interface_layers = 2
support_material_interface_spacing = 0.2
support_material_interface_speed = 100%
support_material_pattern = rectilinear
support_material_spacing = 2
support_material_speed = 50
support_material_synchronize_layers = 0
support_material_threshold = 45
support_material_with_sheath = 0
support_material_xy_spacing = 60%
thin_walls = 0
threads = 4
top_infill_extrusion_width = 0.4
top_solid_infill_speed = 50
top_solid_layers = 5
travel_speed = 250
wipe_tower = 1
wipe_tower_per_color_wipe = 15
wipe_tower_width = 60
wipe_tower_x = 180
wipe_tower_y = 140
xy_size_compensation = 0

Posted : 30/04/2018 10:08 pm
david.b107
(@david-b107)
New Member
Topic starter answered:
Re: Mk3 vs Tarantula -- Ghosting

Hello All, I've taken a look at the settings RetireeJay provided. I tried them with no improvement. In fact, using Slic3r produced worse prints.

I need to walk away from this things for a couple days. But, I wanted to provide a summary of what I tried in case it helps others.

Here is what I've tried with no improvement to ghosting/ringing:
* Slowed the print down to 40 mm/sec with 40% under-speed for perimeters
* Tightened the belts
* Loosened the belts
* Checked bearings to ensure they are smooth, and they are
* Tried multiple print temperatures
* Tried Slic3er default Mk3 settings, made the problem worse
* Removed the rubber feet and placed printer directly on two boards
* Tried 3.1.3 with LA enabled and put the M900 g-code in the script, no changes to ringing

Good luck if you're working through a similar problem.

Posted : 01/05/2018 6:50 pm
RetireeJay
(@retireejay)
Reputable Member
Re: Mk3 vs Tarantula -- Ghosting

Since you have run experiments at more than one printing speed, can you tell if the "wavelength" of the ghosting changes with speed?
I'm thinking that if there's a mechanical vibration, like a loose part of the extruder assembly, that the vibration would be at a constant mechanical frequency. So when you "record" this constant tone onto a surface at different speeds, the distance between peaks would change. Faster speeds would result in longer distances between peaks, in direct proportion to the ratio of the speeds tested.

Posted : 01/05/2018 11:06 pm
david.b107
(@david-b107)
New Member
Topic starter answered:
Re: Mk3 vs Tarantula -- Ghosting

@RetireeJay, thanks for the continued help.

I see exactly what you describe. When I look at a cube printed at a perimeter speed of 34 mm/sec, the waves are spaced further apart than one printed at 15 mm/sec. This is on the X and Y axis both. Am I correct in thinking that if it was the extruder, I'd only see the effect on the "Y" side of the cube?

I did check the extruder. All bolts that were accessible from the outside were tight. I grabbed the hotend and it seemed secure. Is there something in particular I should look at?

On a different note: I did more research into the linear bearings. I mentioned before that I thought mine were smooth. I need to qualify that a bit. Both X and Y moved smoothly on the rods, meaning there was no binding. Once they were moving they moved easily. But the movement was "gritty" at slow speeds.

I took the belts off the motors so both axes would move freely. If I tried to manually move the extruder by say 1mm by pushing it with my finger, the "gritty" action i described would cause it to "jump". In contrast, my Tarantula has rollers. I did this same experiment with it and the extruder didn't jump. Could this be causing the ringing? Maybe the bearings jumping and causing the motors to have to "home back in"?

Posted : 02/05/2018 2:14 am
RetireeJay
(@retireejay)
Reputable Member
Re: Mk3 vs Tarantula -- Ghosting


If I tried to manually move the extruder by say 1mm by pushing it with my finger, the "gritty" action i described would cause it to "jump". In contrast, my Tarantula has rollers. I did this same experiment with it and the extruder didn't jump. Could this be causing the ringing? Maybe the bearings jumping and causing the motors to have to "home back in"?

That sounds quite plausible. Although your printer is new, it may still benefit from lubrication. Prusa's own recommendation is here: https://www.prusaprinters.org/top-6-tips-original-prusa-i3-3d-printer-maintenance/

Also, the linear bearings and the rods are manufactured to very tight tolerances. If the space between the rods doesn't match the space between the bearings, then there is a constant sideways force on the assembly, increasing friction in the bearings. This is taken into account (to some extent) by the assembly process for the Y axis, but there's not much opportunity to tweak spacings on the X axis.

Posted : 02/05/2018 3:35 am
Share: