Notifications
Clear all

Oh no we're skewed! Prusa Mini edition.  

Page 2 / 3
  RSS
Andrey Maslyuk
(@andrey-maslyuk)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

Thank you @kranex !

Used the tool and two models to get all three axis corrections.

https://github.com/Kranex/goskew

https://www.prusaprinters.org/prints/112122-skew-detection-and-correction

Posted : 03/01/2022 5:01 pm
Rainer
(@rainer-2)
Reputable Member
RE: Oh no we're skewed! Prusa Mini edition.

 

Posted by: @kranex
Posted by: @turro75

I discovered gskewer, available as octoprint plugin. Not tested  but it looks promising.

Here the link https://github.com/Kranex/octoprint-gskewer.git

If you don't want to use octoprint I can create a prusaslicer post-processing script that does the same thing. I'm currently experimenting with a new method of calculating the error in the xy plane as I've found the one outlined for gskewer fiddly and prone to error. 

I've just tried goskew over the Prusa Slicer and I'm thrilled.Will you also expand it for the Z axis?

I just installed your octoprint plugin.Unfortunately, the .PDF with the explanations of how to calculate it no longer exists.Maybe you can explain it briefly.The Octoprint plugin where you only have to enter the measurement of a test cube would of course be awesome.

Posted : 05/01/2022 8:39 am
Andrey Maslyuk
(@andrey-maslyuk)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

 

Posted by: @rainer-2

 

Posted by: @kranex
Posted by: @turro75

I discovered gskewer, available as octoprint plugin. Not tested  but it looks promising.

Here the link https://github.com/Kranex/octoprint-gskewer.git

If you don't want to use octoprint I can create a prusaslicer post-processing script that does the same thing. I'm currently experimenting with a new method of calculating the error in the xy plane as I've found the one outlined for gskewer fiddly and prone to error. 

I've just tried goskew over the Prusa Slicer and I'm thrilled.Will you also expand it for the Z axis?

I just installed your octoprint plugin.Unfortunately, the .PDF with the explanations of how to calculate it no longer exists.Maybe you can explain it briefly.The Octoprint plugin where you only have to enter the measurement of a test cube would of course be awesome.

 

@rainer-2, I just posted a model here; I was able to correct all three axis using it and goskew.
Best case scenario you need to print one flat model and two larger ones (last one to confirm correctness).

Posted : 05/01/2022 4:07 pm
Rainer liked
Rainer
(@rainer-2)
Reputable Member
RE:

Hi, thank you very much, I had already discovered it and tried it today.But I printed the large version twice.In any case, it works.

My measurements my measurements before / after    XY  139,97/138,95 -----139,55/139,53

                                                                                               XZ   139,40/139,63-----139,40/139,63

                                                                                               YZ   139,91/138,80-----139,30/139,35

i realized it with the octoprint plugin.With the XZ, I had a negative score. I think it works anyway. the measurements before and after are identical but only have a slight difference.

I used this tool to calculate.

 

     

Posted : 05/01/2022 6:20 pm
Kranex
(@kranex)
Active Member
RE:

 

Posted by: @andrey-maslyuk

Thank you @kranex !

Used the tool and two models to get all three axis corrections.

https://github.com/Kranex/goskew

https://www.prusaprinters.org/prints/112122-skew-detection-and-correction

Thank you for this, I've now found a great test print for calculating skew in all axis, and made a spreadsheet calculator for it.

Posted by: @rainer-2

 

Posted by: @kranex
Posted by: @turro75

I discovered gskewer, available as octoprint plugin. Not tested  but it looks promising.

Here the link https://github.com/Kranex/octoprint-gskewer.git

If you don't want to use octoprint I can create a prusaslicer post-processing script that does the same thing. I'm currently experimenting with a new method of calculating the error in the xy plane as I've found the one outlined for gskewer fiddly and prone to error. 

I've just tried goskew over the Prusa Slicer and I'm thrilled.Will you also expand it for the Z axis?

I just installed your octoprint plugin.Unfortunately, the .PDF with the explanations of how to calculate it no longer exists.Maybe you can explain it briefly.The Octoprint plugin where you only have to enter the measurement of a test cube would of course be awesome.

Shame that the PDF has disappeared, I can't find similar instructions so I've found an alternative.

You can use this calibration print: https://cults3d.com/en/3d-model/tool/easy-skew-calibration-tower-with-spreadsheet

With this spreadsheet calculator: https://bit.ly/3sXsLTa

Note that this spreadsheet is not the same as the one linked to by the calibration print.

I will update the Readme's for GoSkew and the octoprint plugin soon. This thing can honestly replace the Cali-triangle thing I was trying to do, so I might update GoSkew and OctoSkew to simply use the values from this calibration tower.

This post was modified 2 years ago 4 times by Kranex
Posted : 05/01/2022 8:20 pm
Rainer liked
Rainer
(@rainer-2)
Reputable Member
RE: Oh no we're skewed! Prusa Mini edition.

With this spreadsheet calculator: https://bit.ly/3sXsLTa

at least it doesn't work for me I can't enter any values ​​or make a copy of them.

I also did the other test and got the same results.But it is easier to measure and because of the label 'foolproof'With just one layer, 2 top / bottom and 10% infill in 3 hours to print with 200%

Posted : 06/01/2022 9:15 am
Andrey Maslyuk
(@andrey-maslyuk)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

 

Posted by: @kranex

 

Posted by: @andrey-maslyuk

Thank you @kranex !

Used the tool and two models to get all three axis corrections.

https://github.com/Kranex/goskew

https://www.prusaprinters.org/prints/112122-skew-detection-and-correction

Thank you for this, I've now found a great test print for calculating skew in all axis, and made a spreadsheet calculator for it.

Posted by: @rainer-2

 

Posted by: @kranex
Posted by: @turro75

I discovered gskewer, available as octoprint plugin. Not tested  but it looks promising.

Here the link https://github.com/Kranex/octoprint-gskewer.git

If you don't want to use octoprint I can create a prusaslicer post-processing script that does the same thing. I'm currently experimenting with a new method of calculating the error in the xy plane as I've found the one outlined for gskewer fiddly and prone to error. 

I've just tried goskew over the Prusa Slicer and I'm thrilled.Will you also expand it for the Z axis?

I just installed your octoprint plugin.Unfortunately, the .PDF with the explanations of how to calculate it no longer exists.Maybe you can explain it briefly.The Octoprint plugin where you only have to enter the measurement of a test cube would of course be awesome.

Shame that the PDF has disappeared, I can't find similar instructions so I've found an alternative.

You can use this calibration print: https://cults3d.com/en/3d-model/tool/easy-skew-calibration-tower-with-spreadsheet

With this spreadsheet calculator: https://bit.ly/3sXsLTa

Note that this spreadsheet is not the same as the one linked to by the calibration print.

I will update the Readme's for GoSkew and the octoprint plugin soon. This thing can honestly replace the Cali-triangle thing I was trying to do, so I might update GoSkew and OctoSkew to simply use the values from this calibration tower.

It looks nice, but I'm not confident in this model:

- the hanging vertical diagonals - would these be printed straight enough?

- the large flat ends of each bar - it is easier to make a mistake taking exact measure of these.

Posted : 07/01/2022 2:25 am
Kranex
(@kranex)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

@andrey-maslyuk good points. I'm having a little think to see if there's something better than a cube, but aside from the material cost I don't think there's much better. I think I can make a calculator that uses the diagonals of a cube to generate the tangents, but this time I want to print it first. If it works out I'll host a calculator on a github page rather than using a spreadsheet.

biggest issue that I can think of right now, is that the shape needs to be made in a way that doesn't interfere with measurements. the reason I was using a triangle to calibrate the XY plane is that I can print them at the maximum size I can measure and the other corner doesn't get in the way of the measurement. Measuring the diagonals of a cube is trickier as there's no nice flat edge to measure against, you can measure over the face, but I found that to be a little tricky.

Posted : 23/01/2022 7:25 pm
Chris Laudermilk
(@chris-laudermilk)
Estimable Member
RE: Oh no we're skewed! Prusa Mini edition.

Ah ha! Your post made a design feature of the Voron calibration cube suddenly become obvious!

https://www.prusaprinters.org/prints/101683-voron-design-calibration-cube

Now I realize why the corners of the cube are chamfered--diagonal measurement. So that offers face-to-face measurement, diagonals measurement, and a bearing press-fit test.

Mini+ (kit) - Revo Micro | Antler Cooling | WiFi (4.4.0RC1)

Posted : 24/01/2022 2:56 pm
Kranex
(@kranex)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

@chris-laudermilk That is also an interesting print for calibration, I'm pretty sure that I could make a calculator for that one too, but the diagonals cutting through all planes will probably make the math kinda fun.

I'm thinking rather than simply thinking about which I think will be best, I'm just going to make calculators for multiple models and see which I find easiest and most accurate. Maybe make a poll somewhere and see what other people think too.

Posted : 25/01/2022 8:00 am
LaserGuru
(@laserguru)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

 

With this spreadsheet calculator: https://bit.ly/3sXsLTa

This isn't public; no one except you can use it.

 

Posted : 05/03/2022 5:39 pm
cjnaz
(@cjnaz)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

@nizmox, thanks much for this thread.  I didn't see any activity on the github page, so I'll post here. 

In your original post you have to sections titled "XY Skew".  The second one should probably be titled "YZ Skew".  (I see that its different on the github posting.)

It's odd that Prusa doesn't talk about XY skew in their Squaring your MINI guide.  Apparently it's not a thing - until you try to print things that have to fit together, like a box top.  Hum...  I had hours on the phone with tech support a year ago where I came to understand the problem.  Their fix was to send me a new Buddy board box and new T-nuts.  The crux of the problem is that simply tightening down the screws will lead to a skewed XY.

I recently designed an XY alignment adjuster, with procedure, that allows you to reliably adjust out XY skew. The orange first layer cal is after (nearly final) XY adjustment. 

Original Prusa MINI with Silicone bed leveling mod, YZ alignment bracket, Heatbed cable guide, Octoprint, Wyze IP camera, Capricorn Boden tube, Kasa smart socket

Posted : 27/04/2022 4:12 am
Kranex
(@kranex)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.
Posted by: @laserguru

 

With this spreadsheet calculator: https://bit.ly/3sXsLTa

This isn't public; no one except you can use it.

 

Damn, sorry about that. I thought I had published it publicly, but apparently not. Anyhow, I've finally gotten around to making a simple website for calculating error tangents from a test print. It's a little rough around the edges, but it's up and running here: https://kranex.github.io/skew-it/

I made a new stl for this, I think that keeping things cubic is the easiest way of keeping both the math simple and the prints precise. I'm more than happy to add different print calculators in the future if we can find something better.

Posted : 01/05/2022 9:47 pm
Padawan
(@padawan)
New Member
RE: Oh no we're skewed! Prusa Mini edition.

There is so much room for error, we should be able to compensate the skewness in the firmware like with the Z-distance. Print squares and adjust until it's fine.

Posted : 11/05/2022 8:59 pm
Andrey Maslyuk
(@andrey-maslyuk)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

@padawan, try my solution based on @kranex skew correction tool:

https://www.printables.com/model/112122-skew-detection-and-correction

I can now print both 2D and 3D items which fit perfectly from first attempt.

Posted : 12/05/2022 2:08 pm
Padawan
(@padawan)
New Member
RE: Oh no we're skewed! Prusa Mini edition.

This is great thanks !

Posted : 12/05/2022 10:42 pm
Jan One
(@jan-one)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

 

Posted by: @kranex
Posted by: @laserguru

 

With this spreadsheet calculator: https://bit.ly/3sXsLTa

This isn't public; no one except you can use it.

 

Damn, sorry about that. I thought I had published it publicly, but apparently not. Anyhow, I've finally gotten around to making a simple website for calculating error tangents from a test print. It's a little rough around the edges, but it's up and running here: https://kranex.github.io/skew-it/

I made a new stl for this, I think that keeping things cubic is the easiest way of keeping both the math simple and the prints precise. I'm more than happy to add different print calculators in the future if we can find something better.

@kranex The E and F diagonal labels in your skube.stl at https://kranex.github.io/skew-it/ are wrong, they should be switched. I used the cube for reference and had a bad feeling about those labels, so I did another test print and it made my YZ skew worse like I suspected, had to change the correction value sign.

Posted : 15/08/2022 2:00 pm
Jan One
(@jan-one)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

 

Posted by: @jan-one

 

Posted by: @kranex
Posted by: @laserguru

 

With this spreadsheet calculator: https://bit.ly/3sXsLTa

This isn't public; no one except you can use it.

 

Damn, sorry about that. I thought I had published it publicly, but apparently not. Anyhow, I've finally gotten around to making a simple website for calculating error tangents from a test print. It's a little rough around the edges, but it's up and running here: https://kranex.github.io/skew-it/

I made a new stl for this, I think that keeping things cubic is the easiest way of keeping both the math simple and the prints precise. I'm more than happy to add different print calculators in the future if we can find something better.

@kranex The E and F diagonal labels in your skube.stl at https://kranex.github.io/skew-it/ are wrong, they should be switched. I used the cube for reference and had a bad feeling about those labels, so I did another test print and it made my YZ skew worse like I suspected, had to change the correction value sign.

Scratch that, looks like it works with your calculator (the E, F values in your YZ equation are also switched i guess). I used a different calculator and it came out wrong.

Posted : 15/08/2022 4:44 pm
Jimmy
(@jimmy-6)
Active Member
RE: Oh no we're skewed! Prusa Mini edition.

Honestly I'm thinking of sending the Mini+ back. I have a XY skew which is a pain to fix and I fear will be back with time.

Really this should be fixed through software/firmware.

Posted : 28/08/2022 6:24 pm
rlm
 rlm
(@rlm)
New Member
RE: Oh no we're skewed! Prusa Mini edition.

hello, i used goskew with prusasclicer for a mini+ and i notice in the gcode file anomalies on thumbnails after processing in goskew. i guess you have to remove the thumbnails in prusasclicer?

 

for example :

; generated by PrusaSlicer 2.5.0+win64 on 2023-02-03 at 12:40:34 UTC

; thumbnail begin 220x124 11772

; h786ZPQNCVG1lMaIqJq5zqTEX7.000uqFAoc78mbVY0.000DaAUcq8GzDB3nQiUnFmCvlWboX7.000uNk+paElUEX7.000Y0.000

; nspOi3VXUiQ2GToNFkuo/MtM4DDWOEyn7eQ+OaoKFqegp5Y+1jyWBj22VTboDj5rrHvYBzXVIvtdpJ

Posted : 03/02/2023 12:52 pm
Page 2 / 3
Share: