diff --git a/calibration.html b/calibration.html index 06316ca..eeaf777 100644 --- a/calibration.html +++ b/calibration.html @@ -212,6 +212,7 @@

Initial setup of the printer, regular maintainence, if first layer quality diminishes, any time the frame or mechanical components have been disassembled or replaced, any change of bed surface or nozzle, a change in filament that has significantly difference bed/hot end temperatures. There is a lot that can throw the bed level off, but careful use of your printer without any hardware changes should see it remain consistent for an extended period of time.

Tools:

The gcode generator on this page. A standard sheet of office paper.

+

General Principles

Getting a good first layer is an essential part of 3D printing successfully and is probably the number one cause of failed prints for new users.

@@ -245,7 +246,7 @@

If BABYSTEP_ZPROBE_OFFSET is enabled in Marlin, setting the Z offset can easily be done as the first layer goes down. Don't forget to save to EEPROM afterwards. Newer versions of Marlin also have a Z offset wizard that can be included when you compile. I have a dedicated video for this:

Another advantage of some ABL systems is that once the Z offset is set, you can interchange build surfaces of various thicknesses, with no changes needed for a successful first layer. Assuming the probe is triggered the same way on the bed surface, the Z offset is applied to this trigger point and the first layer height should be the same. On a manually levelled bed, the four corner knobs would need to be turned in unison to raise or lower the bed in accounting for thickness of the new build surface.

-

First layer gcode generator

+
First layer gcode generator

The following form will create a series of five squares that you can use to live level your bed or set the Z offset. It is quick to print and features one square in the middle of the bed, with four others in the corners. You can use these to turn the levelling knobs in each corner until they are consistent, or ensure your ABL system is working if you have one in place.

@@ -271,7 +272,9 @@

Before general calibration or before a significant modification is to be fitted.

Tools:

Gcode generator on this page.

+ +
Baseline test print generator

The form below will create a customised version of the XYZ 20mm calibration cube by iDig3Dprinting. It is fast to print and gives a good indication if there is any fundamental problem with the printer.

@@ -296,6 +299,7 @@
Tools:

Ruler, permanent marker, terminal software such as Pronterface or Octoprint.

Instructions on how to setup terminal software can be found here.

+

For the X, Y, and Z axes, the steps per mm is usually consistent between printers and rarely changes with modifications. As long as belts are tight and true, it rarely needs to be tuned.

For the extruder however, variations in extruder hardware and filament means it is worth properly calibrating the extruder steps per mm, or E-steps.

@@ -322,6 +326,7 @@

The filament will then very slowly go through the extruder (and hot end). Once the extrusion finishes, we measure the distance between the mark and the entry to the extruder.

Ideally, 20mm remains, which means exactly 100mm was extruded. If your distance is anything other than this, complete the form below to calculate the correct E-steps:

+
E-steps calculator

@@ -365,6 +370,7 @@

Base calibration, as well as any time there has been a change to the extruder/hot end. You may wish to revisit this after tuning linear advance.

Tools:

Your favourite slicer. Accurate digital/vernier calipers (two decimal places is much more preferable to a set with only one).

+

Our E-steps are now correct in the firmware, so we will move on to calibrating the slicer. Each slicer has a setting to control the overall amount of filament extruded by the printer. If the flow rate is increased, more filament will be extruded. If the flow rate is decreased, less filament will be extruded.

In Simplify3D and PrusaSlicer, this is called Extrusion Multiplier. Cura calls it Flow.

@@ -446,6 +452,7 @@

Use digital/vernier callipers to measure the outer wall thickness of the hollow cube. Take measurements in multiple places/sides and average them. You may wish to cut/tear off the lower and upper layers of the cube. This is to remove portions with elephant's foot and/or other abnormalities.

If your measurement is significantly off, the following calculator can then be used to calculate the new flow rate:

+
Flow rate calculator
@@ -616,6 +623,7 @@

Initial calibration, any time the hot end or extruder is changed, when trying a new type/brand of filament.

Tools:

Gcode generator on this page.

+

FDM works by melting plastic filament and extruding it accurately one layer at a time to build up 3D geometry. By its nature, the plastic will continue to ooze and drip out of the nozzle even when not pushed by the extruder. To combat this, our slicers use retraction, where the filament is withdrawn from the hot end, alleviating pressure and minimising ooze. When properly tuned, this has the effect of removing stringing, the unwanted oozing of plastic between two points of the model.

An example of fine stringing can be seen in the following image. It appears like cobwebs:

@@ -667,21 +675,7 @@
-
-
I can't translate my test results to my own slicer! Other factors beyond the scope of this test - Important!
- -
+
Retraction tuning tower generator

The following form will create a retraction tower to conveniently test back to back parameters in the same print. Of the three available parameters, it is best to change only one per test print. For example, keep the retraction speed and extra restart distance the same, but vary the retraction distance over each segment. Changing more than one parameter makes is hard to tell what made the difference. The print is quick, so repeat the test varying other parameters until you are happy with them all.

Here is the STL if you would like to slice a similar test yourself: retractiontestv2.stl. This file has been updated to V2, which changes the external shape from circular to pentagonal. It is also slightly shorter to print faster. The original file is still available here: retractiontest.stl

@@ -697,6 +691,23 @@

I would then repeat the test, setting the same retraction distance for each segment and instead altering the retraction speed to dial that in. A third test could then take place to test extra restart distance.

If you would like to be able to customise additional parameters for a retraction test, Prahjister has made a great tool: Retraction Calibration Tool. It has a higher degree of difficulty due to needing more parameters but is ultimately more powerful. Warning! This is an external website and beyond my control. Some users have reported success and others have had issues with the gcode generated. As with the gcode made by this website, monitor your printer during printing with a view to cutting the power if needed.

+
+
I can't translate my test results to my own slicer! Other factors beyond the scope of this test - Important!
+

After you have found a combination of parameters which works well on your machine, the idea is to then translate them to your own slicing profile. If you can't replicate the results, please work through the following:

+ +
@@ -708,6 +719,7 @@

Initial calibration, any time the hot end is changed, when trying a new type/brand of filament.

Tools:

Gcode generator on this page.

+

For this calibration, we are only concerned with the temperature of the hot end, not the bed. The bed temperature will need to be matched to any given filament, and once a good value is found, you will generally stick with it.

Instead here we are tuning the temperature at which the filament is extruded. There is no universal temperature for a given filament. Variations in heater blocks and thermistor placement dictate this.

@@ -721,6 +733,7 @@

Your 3D printer firmware will have a minimum hot end temperature extrusion is allowed and a maximum hot end temperature for safety. Make sure to keep within these boundaries to avoid errors.

Here is the STL if you would like to slice a similar test yourself: temperaturetowerv2.stl. This is an updated model that prints in less time, has more variation in overhangs, and has a narrow pyramid in each band to try and snap off to test layer adhesion.

The original design can be found here: temperaturetower.stl

+
Temperature tuning tower generator