Printer Calibration? - for dimensional accuracy.  

I find it better to keep calibration in CAD - machine and slicer settings are there for occasions when you have no control over the original file.

Fine adjustment will vary with different filaments, even different batches and colours of the 'same'.  Thermal contraction will vary a little with ambient temperature and more along than across extrusions; you may find that even the orientation on the bed makes a difference.

For the highest accuracy, print slightly oversized blanks and machine to fit.

Cheerio,

RE: Printer Calibration? - for dimensional accuracy.

What about when you're printing one file on multiple machines? The machine should be able to be calibrated to print what you send it accurately.

I've updated to the latest firmware, I have a feeling it started to ignore M92 at some point. Which means I may have sent out some out of spec product.

I'm running through my whole calibration process from scratch, if it doesn't work I'm going to have to re-slice quite a few models... ugh.

I'm not going to add post processing, I've had product perfect off the bed for years.

Easier to tune them to print identically and then calibrate in CAD for the common correction(s).

Cheerio,

RE:

Posted by: @kenour

I measure and calibrate the machine by putting the M92 command in the start gcode section, for me it's;

M92 X100.29 Y100.53 Z401.06 ; calibration

This brings my models dead on in all dimensions, which is 100% needed for their function.

You’re setting the scaling, but ignoring all other dimensional factors.  Consider what happens, for instance, if the extrusion width is increased by, let’s say, 0.1mm.  The lateral outside dimensions increase by the same amount.  But an internal hole gets smaller.  So, even if you “correct” the outside dimensions with scaling, you’re making the internal hole even smaller!  This is just an example, but it should illustrate the point.

Lots of outside factors can influence extrusion width.  Maybe your nozzle is just oversized due to wear.

Think in terms of tolerance between mating surfaces.  Not overall scaling.

RE:

Lots of outside factors can influence extrusion width.  Maybe your nozzle is just oversized due to wear.

Think in terms of tolerance between mating surfaces.  Not overall scaling.

This isn't how it works, M92 linearly scales up the entire model. I used it to give me exactly what I was expecting, I was able to design, and print, and get within +/- a blonde one with this calibration method. I say WAS, because I tested last night, and it seems the latest firmware ignores M92. I'm not sure when this happened, but as this calibration is in my start gcode, and is being ignored, all my models are printing undersized. Which is super cool.

Also no nozzle wear, using diamondback.

RE:

I've been down this road and I agree with others.  The Prusa printers themselves are kinematicslly identical enough to one another for interchange of parts between printers.  This requires your design to be within the realm of accuracy for a 3d FDM printed part.

Here's what I've found with the countless hours over the years going down your path..... I started the same as you being a mechanical engineer and coming from a background of designing and machining precision die parts with an accuracy of +/- .0001", .00254mm for the metric folks. Yes that's a ten thousanth of an inch.  So you can imagine the adjustment period I went thru 

- The adjusted steps per mm are only good for the size you calibrate it at.  For example, if you use a 50mm cube for calibration then yes you can make it dead nuts at 50mm, but if you print a 70mm cube it'll be off. 

- There are many other variables that affect part size way more than the kinematics of your printer. 

- Filament diameter - manufacturers claim to hold +/- .3mm, but if you actually measure the filament you'll find it all over the place.  And this is with any filament I've tested.  It varies roll to roll. This causes under and over extrusion, and unless your going to adjust this for every roll of filament, your going to be chasing your tail.

- Part geometry.  This includes # of perimeters, infill, and the shape and density of your part. All of these will affect the final size.  Aka 2 parts ~50mm cubes, but with different parameters/ shapes will finish at different sizes after cooling.

- Cooling time between layers.  You can dial in part size printing a single part.  Then if you turn around and print 2 or more at the same time the finished part size changes.  This is because the plastic shrinks at it cools, so the longer the time is between the current layer and the next layer the more it shrinks.  I know this for fact from experience.  Now your can mitigate this some by using the print individual objects feature, however this has a bunch of its own challenges. 

-Plastic shrinkage.  I touched on this, but essentially unless you can accurately control the chamber temperature, your final part size changes based on ambient. 

So in a nutshell the kinematics of your printer are plenty accurate as is, and your best bet for accurate sizing and fits is to design your parts with this in mind.  Things like allowing geometry that has tight fits to be flexible in some manor to accommodate variation in your parts. 

If this is absolutely impossible in your design then printing, measuring and altering your part design is the only accurate way your going to get accurate parts, and even then your at the mercy of all of the other variables I've mentioned. 

Hope this information helps.

RE: Printer Calibration? - for dimensional accuracy.

It sounds like your printer's firmware update may have reset some settings, overriding your M92 adjustments. Double-check if M92 values are being correctly implemented by the printer. Also, ensure that EEPROM settings aren’t altering your calibrations upon reboot. Alternatively, consider using slicer-enabled calibration or even writing a startup script in your slicer to apply M92 commands every time. Some users prefer dedicated calibration cubes/scripts rather than cross/pyramid shapes for more consistent axis calibration. If the firmware continues to cause issues, reaching out to the manufacturer or exploring community firmware options might offer a solution. Good luck!

RE: Printer Calibration? - for dimensional accuracy.

Has a lovely answer for you guys, but this fucking forum timed out AGAIN and I lost my response. Not typing it again, fix your shitty forum.

RE: Printer Calibration? - for dimensional accuracy.

Posted by: @kenour

Had a lovely answer for you guys, but this fucking forum timed out AGAIN and I lost my response. Not typing it again, fix your shitty forum.

@kenour

Sorry, I feel for ya, this happens to many of us here, they have no intention of fixing it. In fact its worse then its been since they moved it pre 2018 but we understand how you feel, it takes the wind out your sails. 😕 

I started years ago typing my longer responses in Google Doc's, then copying and pasting so that when my posts get vaporized I can just paste it again, and yes, it's a pain in the ass, I'd prefer to just type it here and not have to worry about it. The slicer software seems to be treated the same way, they cant seem to add or fix 1 thing without breaking or screwing up 2 others, and then don't do reasonable checks to see what the ramifications of said changes might be.

I'm wondering how long it will be before they start running ads in the software, based on the data they're collecting about us. 🤔 

RE: Printer Calibration? - for dimensional accuracy.

Posted by: @dennis-jay

 Alternatively, consider using slicer-enabled calibration or even writing a startup script in your slicer to apply M92 commands every time. Some users prefer dedicated calibration cubes/scripts rather than cross/pyramid shapes for more consistent axis calibration.

This is how I had it implemented when I was experimenting with it.  I created printer profiles for each of my machines, printed a calibration cube and put the modified steps per mm in the custom startup g code.  I went as far as having a printer profile for each material at which the calibration was done, because of varying material shrinkage.  Aka the same calibration part printed in PETG required different values vs. bring printed in ASA or Nylon.  You can see how this becomes a mess to keep up to date especially if you print lots of different materials and want to keep you profiles updated as Prusa releases updates. 

Next time the forum times out on you try hitting the back button to come back to the forum page where you typed your response and do a refresh.  This always brings what in typed back.  This is using edge. 

Good luck.

RE: Printer Calibration? - for dimensional accuracy.

@kenour, I see you've opened an issue in the firmware repository, and they've already opened an internal ticket on it. Your issue statement is clear enough to easily reproduce the bug.

Have you tried M92 alone on the serial port to see what you get back? Are your values getting stored?

RE: Printer Calibration? - for dimensional accuracy.

Rolled back to 6.13+7898, all issues solved. Including the shitty bed leveling issue that was introduced in one of the subsequent updates, as well as the seam issue I was having.

I normally copy pasta before making a forum post, but just forgot to this time... as is tradition.

Regarding other people's experiences, I don't find that calibrating for one size fucks all others. The shape I made has marks at 100/150/200 in all directions. I measure all, put in a table, that takes the average and calculates the adjustment. After that, everything I print is on size. Maybe I'm doing it differently to you, but this works perfectly for me and my products.

I use the same PETG from the same supplier, and print the same handful of models that need to be absolutely dimensionally accurate to be fit for purpose. The printer is enclosed, and all other variables accounted for.

So, will stay on this version until I see the bug fixed I guess. Hope the fix the bed levelling issue, but that's a story for another time... I guess I'll have to report that too.

RE: Printer Calibration? - for dimensional accuracy.

Posted by: @kenour

Rolled back to 6.13+7898, all issues solved. Including the shitty bed leveling issue that was introduced in one of the subsequent updates, as well as the seam issue I was having.

I normally copy pasta before making a forum post, but just forgot to this time... as is tradition.

Regarding other people's experiences, I don't find that calibrating for one size fucks all others. The shape I made has marks at 100/150/200 in all directions. I measure all, put in a table, that takes the average and calculates the adjustment. After that, everything I print is on size. Maybe I'm doing it differently to you, but this works perfectly for me and my products.

I use the same PETG from the same supplier, and print the same handful of models that need to be absolutely dimensionally accurate to be fit for purpose. The printer is enclosed, and all other variables accounted for.

So, will stay on this version until I see the bug fixed I guess. Hope the fix the bed levelling issue, but that's a story for another time... I guess I'll have to report that too.

With what tool are your accurately measuring a mark?  I'm just curious.  I suspect our interpretation of accurate is different.

RE: Printer Calibration? - for dimensional accuracy.

With what tool are your accurately measuring a mark?

Feeler gauge, micrometer and laser interferometer, you know, has to be spot on... 😅

I make tool guides, the internal measurement needs to be 99.7mm (+0.4mm clearance), -0.1mm/+0.2mm.

So yeah, I have about a 10 thou tolerance, but it has to be in the range or it won't be fit for purpose.

The way I had this machine tuned before the firmware fucked me was completely acceptable.

Our interpretation of accuracy isn't different, but I no doubt have looser tolerances due to the application.

The slicer software seems to be treated the same way, they cant seem to add or fix 1 thing without breaking or screwing up 2 others, and then don't do reasonable checks to see what the ramifications of said changes might be.

Ugh, ain't that the truth... and when I notice and report an issue people are like "NaH iT's YoU". Sure thing knackers, I have no idea what I'm doing! 😅 Especially when I can roll back the firmware and confirm my suspicions. I better go post in the first layer issue I noticed too, that miraculously went away when I rolled back.

RE: Printer Calibration? - for dimensional accuracy.

THIS from a sister thread -

https://www.hubbardhobbies.com/3dprinting/3D-printingnotes.html

I agree with most of his reasoning, even the point about printing SHOULD produce parts to spec (within a certain tolerance) except that technology just isn’t quite there, yet.  Machine tooling is far ahead of 3D printing.

RE: Printer Calibration? - for dimensional accuracy.

printing SHOULD produce parts to spec (within a certain tolerance) except that technology just isn’t quite there, yet.

I mean, I had this, as above, for years after initial calibration. But now I know it's a firmware issue, not my fault, I'll wait until the issue is solved before updating again. 

RE: Printer Calibration? - for dimensional accuracy.

they've already opened an internal ticket on it.

Is that good? Does it mean they'll fix it promptly or no?

RE: Printer Calibration? - for dimensional accuracy.

I type and edit stuff I want to post here in Notepad++ and when I'm happy copy and paste it here.