Calibration Cubes: More Harm Than Good?

Digital calipers? Disgusting. Vernier is all you need.

Major props for letting the author of the cauliflower update his design, it's not one I've seen before but I've been meaning to my Voron even more dialed in and it looks like it'd be a big help. I'm fully in agreement with your conclusions. Steps-per-mm are defined by the gearing of your printer. Shrinkage adjustments, line width, etc will dial in a printer far better. And don't forget to print a cooling tower to dial in your fan speeds as well!

I noticed the cube was correct, but my hole distance wasn't correct anymore. That is when I started testing my calibrations with different methods. Great video btw 👍

Strange…but I have never printed a cube yet. The printer prints accurately enough..

A Mitutoyo gadget for 150 buck to do some fiddling with FDM toys?

Ah, komm jetzt ...

Yes.
Been telling people all of this since 2013. You are correct in every detail in this video.
I like Adam's design, I already made something similar but your joint update looks great.

I thought that myth of calibration cubes died 5 years ago already? It's obivous that the steps/mm are derived from the mechanical system, nothing else.

Thanks!

Firstly, I'm sold, but secondly, extremely casual mention Adam's cancer diagnosis! Funny delivery aside, love to see makers supporting makers

Califlower is not enough. It doesn't tell you anything about Z skew or Z height calibration. I still think a calibration cube is better as you can check for right angle in Z direction. It is just a matter of printing it larger if you want to be more accurate. You can then calculate the skew just as good. This matters alot when you print a complex object which needs to be accurate in X,Y and Z. Especially if you print a box with lid in a 45 degree position to the bed. Checking for skew and dimensional size accuracy of a printer are two entirely different things anyway. The smaller your print object is the less X/Y skew matters. Dimensional size accuracy always matters.

i like it

that was prety balsy saying leave a s***h coment

very confused right now - it says a core xy cannot use skew fix but your rat rig is a core xy? So how come you can set your skew? when in the docs it says you cannot change skew if it's core xy

Everything I printed in PLA is slightly too small and everything in PETG is slightly too big, so it was obvious to me the error was elsewhere (the error doesn't scale with the object either, it's the same 'offset').

I never calibrated like you describe. I began designing printers when they were first a thing and have always used toothed belts and pulleys. They have a fixed number of teeth and with the belt pitch you have a fixed scale and steps per mm. There is no need to change this. No belt slippage will occur. Nominal diameter of pulley is irrelevant as it will mesh with the teeth.
Over and under-extrusion is the only issue and it has to do with pressure, driven by different speeds of extrusion again coming from machine velocitys.

1+ to leaving steps per mm alone. Steps per mm should be mathematically solved for and maybe a very small allowance for belt stretch (esp if cheap belts were used).

It seems to me that if you have calculated steps per mm, printed a calibration cube or flower, and your X Y dimensions don't match - the source of error would likely be the belts. Maybe they are cheap and too elastic, maybe one is permanently stretched, maybe they are of different quality form each other, maybe one is far over tensioned and or the other far under.

I primarily have been using the "cube" as a first print just to make sure the printer is actually working (just ask that half printed Voron cube with a snake coming out of it). Then I worry about tuning and calibration later.

I've bought a few things from Adam like the Wire-Stripping adapter in Feb 2022, his workbench schematics and his Califlower.

Im going to have to get that califlower

Please don't swear

I want to party with you! SLAMS BEER!!

do you have the files that you used for filament profile calibration? I did some heavy modifications to my old Lulzbot and really want to start from ground zero to tune all aspects of this printer to get it working better.

Great video. I appreciate the idea of not using steps per mm to adjust scale, I configure my steps per mm by reading the actual travel of the machine with a dial indicator, so I know that any dimensional error I get in my prints are results of print settings not hardware configuration.

The only way to do calibration is with machinists gauge blocks

I always suspected these calibration cubes to be flawed concept! Exactly for the reasons you listed :)

Really very good content (always!). Thanks a lot!

I always had the same thoughts about the calibration cubes, I had come up with a simpler solution then califlower, but not as robust. Thanks for this video now I have to test my solutions against the califlower :)

I like the idea of the califlower. That said if you are going this far why not just use a Dial indicator on each axis to get your X and Y and Z steps correct? I realize that this still doesn't account for skew so you should do a calibration print after the fact.

Good point about the shrinkage of materials skewing the measurements. I gues printing larger thing would help. And not havign cheap made in china callipers too.

Location? does it matter if you print it in the center or in one of the corners?

Thanks for all that testing!!
Sorry to propose some more tests: How does print-speed affect accuracy?

Stephan Thank You 🙏. Your videos are very informative

The link to the Cliflower download returns "Service Unavailable" is there another source I can obtain this from

Machinist here, 4 thou is pretty good, especially for a machine that sits on ur desk and costs less than a couple grand

Never even occurred to me that anyone would actually use a calibration cube to actually calibrate their printer... That is crazy. Only ever used them as a kinda quicker to print Benchy to see printing quality.

"Unsubscribe or leave a shitty comment, I don't care"


*Clicks the like button immediately after hearing this 🤣

Does this also affect delta printers?

With standard ANSI uncertainty calculation and with a sample population smaller than 31 units per printer, you are probably using the t-value for find the sample standard deviation no? Regardless, this must have been a very tedious and lengthy test, so thank you for all the effort.
I am curious how you isolated what the bias errors were and if you use a gaussian distribution to minimize the precision error before coming to a conclusion.

There are way too many variables to isolate when 3D printing for size calibration in my opinion. I have 2 MK3S+ prusa's that are both in their own insulated enclosures. For size calibration like you are doing, and especially when trying to engineer/design a new calibration print, it would make isolating the bias error easier if you had 3 or more of the same printer. Only after song that would I expand to different printer styles and brands.
It was a good idea to use the exact same filament brand, type, & color, so that was nice to see.

Again, so many things to introduce variation here, so this couldn't have been easy, thank you for the video 🎉

I calibrated my e steps (travel and extruder) with digital calipers as indicator , then calculated shrankage for my particular filament.
For the holes i give .4mm clearance (i print with 0.6mm nozzle 0.8 line width and 0.28 layer height), and otherwise model true to size. Then i scale model .3% in slicer. i print mostly in petg and usually flat\wide parts. I didnt propertly test z shrink, but so far uniform scaling didnt cause any problems.

Mmmmmmm I love the confidence

calipers are only accurate to .001 of and inch that can be about .05 of a mill you can calibrate them but still hard to do much better then that. dial calipers a bit better don't sweat .05 it's nothing.

I feel sorry for people who believe in the Abrahamic Religion of Allopathy - they are told they will die unless they inject extremely toxic poisons into their body which are the REAL cause of death from 'cancer'. The victims have no-one else to blame than themselves because they choose to inject the poison into their veins and essentially kill themselves.

We can repurpose this califlower as a cup coaster!


as opposed to XYZ cube. NICE!

bad calibration cube! bad! go put your nose in the corner and think about what youve done. nope nope, dont want no talk -back.....nose-corner, now!

I greatly appreciate you doing research across your brands, and especially testing multiples of the same brand where possible. You are a gift to the community!

"Have you ever measured yours". Is that some juvenile German humor? ;)

Brilliant!!!!!! TY :-)

You should have used white material on the Cauli- flower.... would have been more realistic.

Can you make a video of the other calibrations? Flow calibration, temperature etc?

Should've been sponsored by Cubespace.