Arcol blog Just another RepRap weblog


IKEA drawer

Last December I was approached by a friend to do a joint build on a better printer.

I took them up on the deal, because I had an IKEA drawer, where the plastic rail guide had broken, and I was unable to print a replacement part with either of my 3D printers.

I think it is fairly reasonable deciding to build a whole new 3D
printer to repair a piece of furniture which only cost 18EUR originally.

Here is the furniture in question:

We set many goals for our ultimate finished machine,
but the first milestone was to be able to print out this replacement

This model is an excellent stress test, because:
1. it is 200mm long, so with the pads it only fits diagonally on the heated print platform
2. has thin walls
3. well calibrated bridging is necessary
4. retraction is really critical (the small cylinders at the top)
5. relatively fast to print
6. it is big only in one direction, so wants to curl up enormously
7. perfectly levelled bed is necessary. PCB alone does not work.
8. we wanted to print it with ABS plastic (the more difficult one)

Here are some pics to back up the above claims.
It fits only diagonally onto the heated print bed (point 1.)

Bridging is necessary, and has thin walls, so only two filament width.
(point 2. and 3.)

Retraction is really critical (point 4.)

And also another cylinder at the top of the model:

No curling occured (point 6.)

And as a bonus I compiled a calibration object, which is even faster to print:

This model was successfully printed on 5 of March, and here it is installed:

And it has not broken since then, so holding up fairly well
(5 months already!), and we use it daily.

One last photo to show the drawer in action:)

I also published both models on thingiverse, as thing:31115.
Maybe others will find them equally useful.

Technical infos:
1. The model was printed on March 5 using slic3r 0.7.0.
2. We used a 0.5mm v4.0.1 hotend at the time. (currently 4.1.1 is on the sale)
3. Red ABS from reprapsource.

The slic3r config.ini

Layer H 0.4mm
E multi 1
bottom_layer_speed_ratio = 0.3
bridge_flow_ratio = 1.4
bridge_speed = 30
duplicate_distance = 6
duplicate_x = 1
duplicate_y = 1
end_gcode = M104 S0 ; turn off temperature\nG28 X0 ; home X axis\nM84 ; disable motors\nM107 ; Fan off
extrusion_axis = E
extrusion_multiplier = 1
extrusion_width_ratio = 0
filament_diameter = 2.95
fill_angle = 0
fill_density = 0.3
fill_pattern = rectilinear
first_layer_height_ratio = 1
g0 = 0
gcode_arcs = 0
gcode_comments = 0
infill_every_layers = 1
infill_speed = 50
layer_height = 0.4
nozzle_diameter = 0.5
output_filename_format = [input_filename_base].gcode
perimeter_speed = 40
perimeters = 3
print_center = 85,65
retract_before_travel = 2
retract_length = 1
retract_lift = 0
retract_restart_extra = 0
retract_speed = 6
rotate = 45
scale = 1
skirt_distance = 6
skirt_height = 1
skirts = 1
small_perimeter_speed = 25
solid_fill_pattern = rectilinear
solid_infill_speed = 50
solid_layers = 3
start_gcode = G28 ; home all axes\nM106 S150 ; Fan on
temperature = 248
travel_speed = 120
use_relative_e_distances = 0
z_offset = -0.2


I closed down the commenting on this webpage due to the enormous number of penis enlargement suggestions.

If you have any thoughts to share about this article, please discuss it on's mailing list, in this thread.


Machine Calibrating

In this blog post we try to answer the question, what should be the layer height if the head speed is 16mm/s and the motor speed is 40rpm. So calibrating the extruder theoretically, rather than trial and error.

Since I had my printing machine I always had some sort of printing deficiency. I always suspected,
that I made my settings wrongly or didn't adjusted it right.

I asked on the forum how am I supposed to set my machine/skeinforge settings right, and I was told that I should follow this walkthrough blog post.
I tried it, but I lacked the proper tool to do it right. I only have an "analogue" caliper, where the accuracy in theory 0.05mm, but Im only able to read it reliably 0.1mm:

Following the above tutorial requires a digital caliper with 1 ┬Ám (0.001mm) measuring accuracy.

The idea, that a measuring is required, when we know every details of the machine, is rather strange. So I asked nophead, what is his method to adjust the settings right. He said, he calculates the layer height and does not need to do any "trial and error", like the above mentioned blog post. We exchanged some emails, and here it goes the explication (at least what I grok from it):

Filed under: theory Continue reading