How to 3D Print more Accurately so you can make Lego Bricks

3D Printing doesn't always have to be accurate. But when it does, you need to know the steps to make your 3D printer behave.

Be sure to like, subscribe, share and enjoy.

3D Printing Professor is made possible by YOU!

I'm social!




Licensed under Creative Commons: By Attribution 3.0 License

[jirij]

Few more points from me (to anyone interested, please don't take them as a criticism), going down the rabbit hole:

1) A more common (alternate) way to calibrate extuder steps-per-mm (or flowrate %) is to measure the difference before the extruder, ie. measure 120mm of filament away from the extruder body, make a sharpie mark, tell the printer to extrude 100mm, measure the distance to mark, do the math. Amongst other things, this is easier to measure to a higher precision given the same tools.

2) The flow changes between filaments come mainly from a different effective radius of your drive gear as it bites deeper/shallower into the filament. This is also why PETG tends to need more steps (or flow %) as it is softer, the drive gear teeth "sink in" more, reducing the effective radius. Similar for ABS. Filament width differences still matter, though not as much in my experience (with cheap PLA).

3) As mentioned in (2), effective radius matters. If your printer has an adjustable tension knob on the extruder, set it once, calibrate and don't touch it! Tightening/losening will throw off your calibration.

4) Setting your steps-per-mm for X/Y/Z to a measured value rather than a calculated one is, in some circles, considered a bad idea as it hides real problems with the mechanical system (like skipping microsteps), which will manifest unreliably on shapes more complex than a 20x20 cube. Try printing slower or increase current to your motors. Also check for loose pulleys, loose belts, loose motor mounts, belts running slightly diagonally to X/Y axes, etc.

5) A 20x20 cube is a horrible accuracy test, print a 100x100 flat hollow square instead - any inaccuracies in size will be much more magnified over larger distances.

6) Square your printer in X/Y, X/Z and Y/Z, ie. thing:2563185. If you cannot physically move the rails, turn on SKEW_CORRECTION and input your measurements.

7) Turn up your cooling fan / focus it better. FFF printing, like welding or casting, suffers greatly from accuracy losses due to thermal contraction while (slowly) cooling. A powerful fan will cool the outside of a printed track fast enough to lock-in the as-printed shape. For the same reason, consider printing on a raft as a hot bed will cause the first layers to cool differently (when the print finishes), deforming it.

8) Perform step (1) with different extrusion speeds - many people report different values for different speeds (even though it makes no sense). This may again indicate mechanical issues (microstep loss) or a clogged drive gear. If you can't get repeatable results using the same speed, the drive gear or tension bearing may be bad (non-circular) or badly mounted (different effective radius during one rotation), etc.

9) Enable and tune linear (pressure) advance, BACKLASH_COMPENSATION, ..., tune your acceleration values to eliminate ringing (Teaching Tech made a great video about it recently!).

10) Use an SLA printer. :)

[hydrophobicgoldfish]

Most informative, comprehensive and helpful video on this topic that I have come across. Kudos

[OldCurmudgeonDP]

As others have said, X and Y movements are mechanically defined by the pulley diameter and the deg/step of the motor. Z movements are by deg/step of the motor and the "pitch" of the lead screw. The extruder will have some variation due to the design of the teeth in the job gear, but is mainly based on, again, deg/step of the motor and the diameter of the gear. The linear distance of filament moved will not deviate for a given step count unless, as someone stated, the gear bites more or less into it. Only the extruder should be compensated for in the slicer. If X/Y/Z are not accurate then the steps/mm value was changed from factory defaults or you should look for other problems that need correcting.

[bubbadoolittle]

A very nice video! I have been wondering how to gain better dimensional accuracy and now I know! Many thanks!

[dlh]

This was super helpful, I was able to figure out my flow rate was very low (.35mm for a .4mm nozzle!) and was able to successfully print a lego for the first time!

[mfeldheim]

Great explanation. Feels like common sense often doesn’t help if you don’t know every component and how they work exactly. Learned a lot here

[SamFerlauto]

Thank you. Been printing for a few years now and kept going round in circles .... calibrating, adjusting steps, calibrating, re and all that time was forgetting to calibrate the flow . Just done my CR10 v2 and Anet A8 and miraculously both now print heaps better. Thank you love your videos

[magomat]

This are the tubes we need more great topic prof.

[flymario]

Loving it! Still have a ton of Dino's to print from your last Kickstarter! :)

[NoraFulcanelli]

Just pledged on your kickstarter, you rock Joe. Keep it up.

[berniewasserman]

Excellent video but you missed an important item. Different material have different expansion and shrinkage characteristics like PLA vs PETG. This will also have an effect on accuracy.

[MichealWeinfurtner]

Thanks, lots of useful info in this episode.

[welbot]

One of these days when I get a filament printer, this will be one of those vids I will come back to refer to! Thanks for the tips :)

[Outofthedust]

Just started printing at .06 layer on my Crappy Geeetech A30. the Original Lego Bricks I printed did okay but the top didnt click into lego legos very well. I still need to try the model I adjusted to see how they fit on the top. I have the Bottom of the brick dialed in

[sierraecho]

Also calculation your movement is wrong, you don´t calculate a movement of a system by it´s end product since many more variables go into that like flow rate, material shrinkage, humidity etc.
You can simply tell your machine to move the X axis by 100mm and measure a before and after just like you would do with the filament.
This is really not best practise, it´s somewhat in the ball park but it will way longer to calibrate your printer properly this way.

[colinfielder]

I love your chanel, I really want to meet you some time to discuss the whole 3D printing world and what it has to offer. If you ever come to Australia look me up

[kbshah]

Where can we find the calibration cube for the flow rate? Thanks 🙏

[setver]

I guess this is a good general view of how to calibrate a machine. Question. If you like marlin so much why not change your firmware?

[pwtprojects]

How to recognize if you can't print properly: you have a bed slinger, has open frame, print with PLA instead of good technical materials (PETG), cant control chamber temperature properly.

[zk_]

I disagree that an oval shaped hole means you have crap printer. Most likely the printer's belts are too loose and need tightening. Belt tensioners would be a good investment. Other hardware problems can be fixed as well or the printer can be modded to work. Yes it will take some time and skill and money, but any poorly working printer can be made to work better with both hardware and software tweeks. Saying a printer is not good enough to print certain parts just means you are not willing to put the time and effort to make it work better. I have been working with my printer for a couple of years now and have just now gotten it to the point where it prints decently and somewhat accurate. Yes there was a lot of frustration, but I did not give up and I enjoyed the time I put in. I still have some work to do though to make it better. Replacing the 8-bit main board is high on my list to make to my printer next.