Ultimate clearance calibration test

by energywave Apr 27, 2016
Download All Files

Thing Apps Enabled

Please Login to Comment

I love this, I have used this many of times. Helps troubleshoot and dial in my machine. Thanks again.

I'm very happy to know that it was so useful! Thank you for telling me :)

thanks for this calibration test, a great idea !

You're welcome! :)

I would highly suggest adding a 0.1mm clearance test as well. Why start at 0.2mm and limit it to average printers and above? A very well calibrated printer should be able to perform a clearance of 0.1mm I think.

That's a good observation. I made this test thinking about my printer that is not very precise and 0,2 clearance without sticking together was a miracle, in my vision. I understood, after have designed this thing, that good calibrated printers where more precise. I think, however, that if you have 0,2 clearance without sticking that's a good result :) I can made a rev. 2, however, starting at 0,1...

100% infill!?! How big is this thing? Thats a lot of filament.

This model was born to economize filament as all clearance tests I've found here in thingiverse where pretty big and slow to print, with big unneeded surfaces. This object is small, it makes nonsense to use an infill <100%.

Thanks for this test!!

Printed this on my Prusa i3 MK2S and ALL nubs separated with no problems. Woot!!
Printed using PLA filament using the stock 0.4 mm nozzle and close-to-default slic3r settings.

You're welcome! ...really! Your nubs separated all without problems?! That's why I would like to buy a Prusa MK2S too... :)

Thank you! I am printing this right now to try but in the mean time I wanted to say thanks up front for sharing this. I have wasted so much material trying to print parts that fit together. I will report back my results after trying this.

You're welcome! It's a pleasure to share what can be helpful for me to other kind people :)
I've used this test to understand the correct tolerance parameter for printing this: http://www.thingiverse.com/thing:53451
I've detached till third pin in the top row (0,4mm), with some force but with bare hand. So I printed that object using same slicer/parameters/filament and using tolerance 0,4 and the result was perfect! (you can see it here: http://www.thingiverse.com/make:216642 )
So I can confirm that this test is very useful :)

Gear Bearing
by emmett
Gear Bearing

Yes, this calibration is an very important. Also I have wasted a lot of material at downloaded projects. At first you have to know how much it is at your machine. Because it is a characteristic of the machine, it should be adjusted at the machine ( this means, settings of the slicer). A normal designer for milling machined parts writes a tolerance of +-0.05mm to the drawing, independent of the machine used. We should learn. At slic3r there is a parameter " XY-size-compensation ", insert the radius you need ( negative is smaller) and slic3r will adjust this. Z is not affected by this problem. Also repsnapper will provide this feature. I mean this is an important feature and I will name all slicers without this feature 'poor software'. (sorry for that). There is no need to accept because slic3r and repsnapper are open source and free of charge. In one of my designs I have described this problem.
A suggestion to all designers here is to describe if there is a gap included in slt files or not, so I can use this parameter or not. Suggestion to all "I made one" is to write if they have used this parameter. There is precision available, we should take it!

Interesting. I mostly used Cura as my 3d printer (closed source) works best with it. I like very much Slic3r but didn't used it much for that problem. I've now seen that parameter "XY-size-compensation" (that I didn't knew existed) and I'm now experimenting, using this object, the effect on the printed model.
Having learned that, I totally agree with you, if every creator write how much clearance he inserted in his model we know how much compensation to add/remove when we print with every specific printer. That seems a right and scientifical approach to the problem.
So the correct steps for calibrate (using Slic3r or other slicers that can compensate XY) are the following:
1) print this object using xy-size-compensation of 0
2) take note of the lowest clearance pin you can detach (for future reference too). This can vary with different print settings but even with different filament.
3) load in slic3r the model you want to print and, knowing the model clearance and your printer compensation, write a correction value in XY-size-compensation.
I'll add that proedure in the model description later (after my tests). Thank you very much for your contribution!!! :)

My only concern here is that I had tried SlicR using a config I found for the mod-t a few months ago and the print head drove itself down into the build plate and put a hole in the build plate and messed up my hot end. So be careful. Being new to 3D printing, I do not know what caused that

Wow, another owner of a MOD-t! I'm not alone :) Ok, I also had some problems at beginning using Slic3r so I was stick with Cura, even because MOD-t team focused on Cura gcode generation to make their firmware, they use cura engine for printing on the cloud with the store.
But at some point I realized that I had to better test Slic3r and I've found it much better in some aspect (not for all objects, though). I can generally produce better quality objects with slic3r, due to the flexibility you can calibrate speeds of various parts. and I prefer the hex filling. Not to mention the just discovered XY-size-compensation...
To overcome my worries, like yours, I analyzed the resulting start fragment of gcode files to understand if there is something different and... there isn't! (if you correctly set start / end gcode in Cura and Slic3r!)
The initial calibration (head going down, then up, then wait to warm up, then down again to do those 1.000 times movement on the X axis) is TOTALLY in the firmware, there isn't something in gcode to tell the printer to do something different. When that is done your gcode is being executed, where the first visible action you can see is the priming on the right.
That said I also had some firmware bugs that dove my head into the bed or into the model (melting it and doing a mess!) and other problems. But those are firmware problems. (I've been in contact with the lead engineer about that).
If you want I can share to you my slic3r configuration (that is working for sure). Hm, even better, I can share it here on thingiverse. Are you interested?

Sure. I would gladly try your Slicr config. My main interest in getting a 3D printer was to learn how to make interlocking parts but so far anything I find on the web to try to print to see how that would work results in the parts all fused together or if I print them separately they don't fit. They are apparently aware of this issue of the Mod-T printing 'heavy', but there is no word on if and when a solution will be available.

I've been in touch with New Matter and they have a new firmware that will greatly improve printing experience, as it should handle acceleration to extrude a constant quantity of material (now the movements are linear, causing differences in extrusion width).
The new firmware should be released this week or the next (they said). Let's hope this will improve the situation.
In the meanwhile the only solution is to print MUCH slower. I print the perimeter at 15 mm/s and the small gaps and small perimeter pieces at 7 mm/s. It takes a lot of time but the result is much more precise. And with Slic3r I've managed to print interlocked objects not fused together.
When I'll have some spare time I'll finish my slic3r profile (I had to do it for me too) and I'll send it to you.
In the meanwhile I can tell you that in all my attempts I've found that 0.3 mm is the clearance to put between two pieces, printed in separate times, if they are not too complex/large (in that case should be slightly increased). But for interlocked pieces printed in one time (like the bearing I've printed) the value is 0,4 mm.
The problem is that this value is much sensible to slicers/settings/filament... You have to find your correct value (using this test object :)