Calibrating a 3D Printer With Test Prints

I got my extruder clog cleared up and, after a delay for a summer road trip, I got back to calibrating my MakerFarm Pegasus 3D printer.

Here are some objects that I printed and used to calibrate my printer. They aren't the only ones available but they worked for me. Search "calibration" on Thingiverse to find other tools.

It takes some time to get a printer working well and it takes vigilance to keep it printing well.

Bed Level

By printing this simple test object (Thingiverse 13053), I learned that a lot of my clogging problems came from a build bed that wasn't level enough.

Bed Level Test

The first time I printed this rectangle I saw I was getting places where the nozzle was so close to the glass that nothing was being extruded. In other places it was above the glass and plastic was dropping down.

I talked a little about bed level as a possible cause of clogging in the last post and mentioned that I reinstalled the springs under the bed.

Using what I saw in the test print, I carefully adjusted each corner to get the bed as level as possible. Then I used the bed leveling option in MatterControl to fine tune it. Another test print looked much better with an even layer of PLA in each of the rectangles.

Extruder Calibration

Before you can get accurate dimensional measurements, you need to make sure the extruder motors are outputting very close to the right amount of filament. Checking the extruders is an easy test and doesn't require any part to be printed.

I measured 120mm up from the top of the extruder body and made a mark on the filament there. Then I used MatterControl to send the command to extrude 100mm of filament. This is the length that comes off the roll and NOT what comes out of the nozzle.

Next, I measured from the top of the extruder up to my mark. If 100mm was extruded, the measurement should be 20mm. In my case it was 27mm. I asked for 100mm and got 93mm so I was under extruding but by less than I expected.

Extruder Calibration

The formula to calculate the E (extruder) steps is:

Expected measurement / actual measurement X current steps

So, 100 / 93 X 117.3 = 157.32

I set that step and ran the test again. This time the mark ended up at 20mm where it belongs.

20mm Cube Test Print

A test cube is probably the first thing most new 3D printer owners try.​

This cube is Thingiverse 214260. I took 33 minutes to print on my MakerFarm Pegasus when sliced with MatterSlice using .2mm layer thicknes

I like this one because it has the X, Y and Z marking so I don't have to remember to mark it. It is infilled to make it more rigid so takes longer than the hollow calibration cubes you will also find on Thingiverse.

The first time I printed the cube it came out with a lot of shifts in the Y axis and the X, Y and Z dimensions were also off.​ I tightened the Y belt and used the formula here to make a adjustments to the steps.

The formula for using a calibration cube to set the X, Y and Z steps per millimeter is​:

Expected measurement / actual measurement X current steps

I printed it again and got a whole new set of wrong dimensions. You can see some ghosting in it too. After chasing numbers for a few tries, I decided just changing the steps settings wasn't working.

20mm Test Cube

​A little online research turned up several articles that claim adjusting steps based on measurement is never a good idea. The step setting should be based on mathematical calculations that take into account the actual mechanical parts used in the drive chain for each axis.

The math is given in on Triffid Hunters calibration guide page at http://reprap.org/wiki/Triffid_Hunter's_Calibration_Guide.

I didn't do the calculations.​ MakerFarm already figured that stuff out so I went back to the recommended setting in the firmware they provide.

Then I took a very close look at the hardware mechanics and finally found out why my numbers were off. You can read all about that process in my Circles Have Me Going in Circles post. Now my cubes are pretty good.

My recommendation is to use the cubes as a check but don't use them to adjust the steps except, maybe, for very precise final adjustments. Before doing that, find out what is really throwing your dimensions off.

Stringing and Oozing

To test for and get rid of the oozing and stringing I used this helpful guide from MatterHackers and the test object they offer in the article.

I have been seeing stringing in the prints I did so I was hoping this would help me get rid of it.

I didn't!​

For the first plot I used the settings the article ended up with as the mid point of improvement from their testing. That was 2mm of retraction and 150mm travel speed.

My test came out with essentially no stringing at all. So what to do. I see stringing in my prints but the test says there is none. I just got a bunch of nice 10mm cubes.

No Strings

I only thing I could think of was to try a more strenuous test.​

Make magazine uses a pretty tough one for testing 3D printers in their annual shootout. I printed that next.

This Got Strings!

Plenty of stringing to see in this print. Although, I am pretty impressed with how well the Pegasus handled the points on these pyramids. As the points get smaller, the printer spends less time going around them and the plastic has less time to cool. That makes pointy things hard for 3D printers.

There are so many factors that can effect stringing it was hard for me to know where to start. I tried to break it down and tackle one potential cause at a time.

Temperature

Extruder temperature is definitely a factor in stringing. The higher the temperature, the more stringy the plastic will be.

Filament manufacturers put temperature ranges on the packages and I started with that. For this red PLA from eSun, the recommended range is 190-220 degrees C.

After some experiments printing single walled cylinders and checking them for oozing and stength. Lower temps mean less ooze but also can cause weak prints. I settled on 190 degrees C as an OK balance of strength and oozing. This can vary by roll and even by color so I will have to watch each one I try.

I will also try reducing the temperature below what eSun recommends to see if it helps even more.​ See the torture test below. I am still trying to improve that one.

Filament Flow

I adjusted the E steps above but I wanted to see if it needed any fine tuning. I used the same cylinders I printed for the temperature test to check extrusion.

First, I saved the g-code for the cylinders to a file and opened it with a text editor. I was looking for the extrusion width that was expected. In this case 0.67mm which is pretty typical for a .4 mm nozzle.

Extrusion Width

I used my digital calipers to measure the wall thickness of the cylinders I printed at several points around the perimeter. Since they were printed as a single perimeter, that number should be 0.67mm.

It is important to measure at the top since the bottom of the part will be a little oversize because of the heated bed and pressure from layers above. I also tried to get just one layer between the caliper jaws but that is hard to do.

The measurements I got ranged from 0.67 to 0.71. Not bad but I did tweak it just a little.

To make the adjustment I divided the expected measurement of .67 by the average of my actual measurements, .69, to get the percentage I would need to modify the extrusion rate by. .67 / .69 = .97 or 97%. I then plug that number into the slicer software.

I know some people change the filament diameter to adjust the flow but why? The ​multiplier is right there and that is what it is meant for.

Extrusion Multiplier

That did reduce the stringing slightly.

Retraction Length

This is supposed to be the key to string control. The MakerFarm default setting is 1mm. I upped that to 2mm of retraction and then went all the way up to 6mm but didn't really see any difference in stringing or oozing. You might see different results. I settled on 2mm as my setting.

Retraction Speed

I set mine at 150mm/s which is what the MatterHackers guide suggested. They claim anything over 150 wont make a significant difference.

Minimum Travel Requiring Retraction

MakerFarm says 2mm and I left it there. I might try a smaller number if I am printing parts with tight spacing and getting strings.

Stringing Still a Work in Progress

I have improved the stringing from my printer but I still feel it can be a lot better. I will continue to work on it by changing one setting at a time and seeing if it reduces the strings.

3D Benchy Test

3D Benchy is a cute little boat that will test the accuracy and capabilities of your printer.There are details and features in this object that will test almost every aspect of your 3D printer.

Benchy is so popular he has his own website, Facebook Page, Thingiverse Group and other social media presence.

When  you download the file from Thingiverse you get a couple variations for single and multi extruder printers. There is also a .pdf file with the full set of measurements so you can compare the ideal to what you actually get when you print it.

3DBenchy Printer Test

If you have children or grandchildren they will be happy to take your test models off your hands. No need to scrape these!

No Waste Here!

Torture Test Your 3D Printer

Thingiverse number 1363023 is a very tough test of printer adjustment and capabilities. It will really show you what needs to be tweaked to get better prints.

Torture Test

Getting perfect prints of this one is something to strive for but I know, in reality, I will never get there. That is the whole idea of this test - to see just where the print quality begins to drop off.

This probably shouldn't be your first print but, if your Benchys are looking good then give it a try. You will learn a lot at the very least.

Here are the tests included in this one print (from Thingiverse):

  • 01 Nut, Size M4 Nut should fit perfectly
  • 02 Wave, rounded print
  • 03 Star, Sharp Edges
  • 04 Name, Complex Shapes
  • 05 Holes, Size 3, 4, 5 mm
  • 06 minimal Distance: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7 mm
  • 07 Z height: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1 mm
  • 08 Wall Thickness: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7 mm
  • 09 Bridge Print: 2, 4, 8, 16 mm
  • 10 Sphere, Rounded Print 4.8mm height
  • 11 Sphere Mix, 7 mm height
  • 12 Pyramide, 7 mm height
  • 13 Overhang: 25, 30, 35, 40, 45, 50, 55, 60, 65, 70°
  • 14 Warp, does it bend?
  • 15 3D Print Font, optimized for 3D printing
  • 16 Surface, Flatness
  • 17 Size, 100 x 100mm x 23.83 (10mm width)
  • 18 Spike, minimum Layer Time, 21 mm height from Bottom (include Baseplate)
  • 19 Hole in Wall, 4 mm diameter, check for proper print
  • 20 Raft Test (if you print it with rafts), raft should be just under the model
  • 21 Retract Travel, check retract settings for longer travel

There are a couple smaller version of this object but they don't include as many tests. They can be handy when you what to try some changes but not wait for the big one to print.

I am still working with this model and probably will be for as long as I own a 3D printer. It really shows were improvements can be made.

Conclusion

Testing and calibration is important with any new 3D printer. You need to get it dialed in for the filaments you use and the projects you print. I know this is something that I will continue to do.

From the last few posts, it must seem that testing and calibration is all I have done with my Pegasus. While I have done a lot of testing, I have been printing other things too. I am getting good use out of my machine. I will be talking about some of those projects in future posts but a few include:

  • A knob for my LCD display
  • Filament filters ti keep my filament clean
  • Toys for my grandson
  • Sewing accessories for my wife.
  • Belt tensioners
  • A Cable clip to help with my latest printer malfunction. Once I get this one fixed, I will explain it in an upcoming post.
Parts

Your comments and questions are always welcome. Do you have calibration tests you prefer? Please let us know what works for you.

Tom

  • Longjohn says:

    If I remember right I have my retraction set for 1.8 mm and generally on a direct drive you don’t want to go much more than 2 mm. You’ll know it’s too much if you get a jamming problem on long prints. I was having more of a problem with a little blob at the start/end point of each layer so I also have set the “Extra length on restart” to -0.2 … don’t forget the negative sign

    For stringing the most important thing is temperature and for that I made a temperature tower . What you want to do is just use the Gcode which has the temperature changes included
    http://www.thingiverse.com/thing:1716435

    You can not only check for stringing but also layer adhesion and use the temperature that gives you the best adhesion with the minimum amount of stringing. I have mine down to what I call ‘angel hair’ which just brushes right off.

    Another setting that can help with stringing is your travel speed AKA Speed for non-print moves. The faster the move the less stringing and the stringing is thinner like … angel hair. You want this as fast as possible without it getting ‘twitchy’ or skipping steps If I remember right 150mm/s was a little too much and I settled for either 130 or 140 mm/s

    For a nice little test print I like to use 3D Benchy which is a little boat that you can print and check your calibration. My PLA ones look as good as any on the website and better than most. The ABS ones are a little trickier because of the need for cooling upper parts and bridges without warping on the bed. You’d probably need something like Simplify 3D and some tweaking of the cooling to get ABS to print as well with PLA.

    • Tom Clough says:

      Thank you very much for the input John.

      I know that I have played with all the settings you mentioned but I don’t think I have just the right combination yet. I will try your settings as a starting point then print the temperature tower.

      I have been neglecting the Pegasus since I got the Klic-N-Print 3D but I do have to get back to using that one too.

      Thanks again,
      Tom

  • >