Filament testing routine

[Nicolinux 288]

Hi,

I am sure most of you print with different filament colors or types. Do you have some routines to share about how to test filament for its optimal printing speed and temperature? Also, since the bowden tube is quite sensitive about filament diameter, how do you measure it accurately?

I wanted to create a generic gcode file with a simple object (something like Illuminarti's infamous extrusion test object - http://umforum.ultimaker.com/index.php?/topic/3976-almost-always-missing-layers-underextruding/?p=33427).

So how are you doing it? Can you share some tips?

 

[IRobertI 520]

Frankly I just print whatever it was I was planning on printing and see how it goes. I'm real sophisticated like that

For diameter I use a digital calliper and measure with the thin/sharp part of it. Using the flat parts can introduce error due to curvature in the filament. I wiggle the filament around until I get the lowest value, then I rotate the filament 90 degrees and repeat and then I use a value somewhere in-between those two measurements. It's good practice to unspool a bit of filament and repeat that procedure just to make sure there aren't huge variations in the diameter.

Mostly I'm lazy though and just go with 2.85 as a good middle ground. I only print PLA though and most of my filament seems to be around that size anyway so I might not be representative of the masses.

 

[takei-naodar 0]

Frankly I just print whatever it was I was planning on printing and see how it goes. I'm real sophisticated like that

For diameter I use a digital calliper and measure with the thin/sharp part of it. Using the flat parts can introduce error due to curvature in the filament. I wiggle the filament around until I get the lowest value, then I rotate the filament 90 degrees and repeat and then I use a value somewhere in-between those two measurements. It's good practice to unspool a bit of filament and repeat that procedure just to make sure there aren't huge variations in the diameter.

Mostly I'm lazy though and just go with 2.85 as a good middle ground. I only print PLA though and most of my filament seems to be around that size anyway so I might not be representative of the masses.

 

Same here.

 

[printedsolid 3]

haha. Same here, and I have a decent amount of experience validating manufacturing processes, so I really should know better : ) Going to follow this thread to see if someone comes along with some good suggestions...

 

[Dim3nsioneer 557]

I use a micrometer screw and take 10 measurements at different spots. Then I take the average as starting point and produce a first print. Then comes one or two corrections. And then I enjoy the great print quality of my Ultimaker... :-P

 

[drayson 75]

I'm also mostly too lazy to make accurate measurements. Do it once, write it on the lable of the spool and come back to that value when changing...

[3dcase 37]

Hi..

Until now did not have any trouble with diameters ( or I did not notice them )

Difference from one roll to the other is easy corrected with changing the flow on the fly.

I am more concerned in the material itself. Would like to see some better coding of the material.

Even when buying PLA from the same supplier, you never now what you get until you try it, perhaps buying 10 rolls at once might help to be sure it's the same batch, but I only need 1 a month :(.

Would be nice if it was more standardized, like steel, or industrial plastics, so you know what you buy.

For example I use a 2.3Kg roll of grey PLA, for almost a year now, and still flexible. from the same supplier some white pla rolls, brake overnight while in the printer?

 

[yellowshark 153]

I have three zones – prototype at 60/80 m/s and 300 microns; good quality at 40/50 m/s and 200 microns; best quality at 20/30 m/s and 100 microns. Which speed I use for any zone will depend on geometry and filament.

Each of the above six tests will use three geometries.

Standard cube

Doing a lot of architectural work, pristine square corners are important to me and this test fits the bill well. It is also a good test for identifying elephant’s feet.

Vertical cylinder

This is a test for surface finish and cylindrical shapes appear frequently in engineering work. I am always trying to push the higher speed in the zone for this one.

Nasty geometry.

This is a shape I have always struggled with. It is a sphere with a small flat bottom. Imagine a cylindrical extrusion through the middle of the sphere and then another extrusion at 90 degrees to the first extrusion. You end up with four vertical circles and four “corners” firstly moving out from the centre and then curving back to the centre. On top of this is a vertical cylinder. It is actually a part from a drive shaft.

 

Each zone has a starting extruder temperature (from past testing/experience)– 220 for prototype; 210 for good quality and 200 for best quality.

For each test if the result is bad I will adjust the temperature by +/- 10 degrees depending on what the problem is. If the result is good I will then do two more tests, one + 5 degrees and - 5 degrees to establish the optimum.

 

I have just added a fourth geometry, a hollow cube with square extrusions cut through the surfaces; this will be good for testing small overhangs.

 

The one aspect I do not test at the moment is retraction. I must say that if retraction is working well then changes in the above parameters do not seem to affect retraction performance for me. IE it seems to be a printer setting rather than a quality/filament setting.

 

 

 

 

 

 

[Nicolinux 288]

Hey yellowshark. Thanks for the tipps. Did you add images to your post? Because if you did, they are not showing up.