A Better Way To Calibrate Your 3D Printer. Importance of Dimensionally Accurate 3D Printed Products
HTML-код
- Опубликовано: 12 окт 2021
- Hey everyone. I want to talk about the importance of dimensional accuracy when running a 3D Print farm and selling products. I hope the information is useful to you and helps you achieve the highest level of accuracy for your products. Thanks for watching.
link to calibration print: www.thingiverse.com/thing:502...
Caliper I use: amzn.to/3T28Sac
Dial Gauge Used in other videos: amzn.to/3I6ySLi
My Favorite Filament (Hatchbox PLA): amzn.to/3SILXiU
Ender 3 Silicone Spacers (A must have over springs): amzn.to/3UH6fvJ
PEI Flex Plate: amzn.to/3UOncUW 
Наука
tolerance must be measured in decimal not in percentage. 20.06 and 100.06 has same tolerance. but if you convert in percentage there will be Major difference between two.
You are correct. I used percentage as a way to make a comparison.
It is clear that the deviation is not caused by the movement of the motors, gears and belts, but certainly in extrusion, so you should not try to make modifications on the number of steps/mm or scaling the STL at the slicer. You must then try to correct using extrusion factor (in this case reducing from 100%) and to obtain the percentage it must be done with measuring wall thickness (printing the cube without infil, only 2 walls stopping the print at about 5 mm of height), but be sure that your nozzles are in perfect condition.
Nozzle holes are too tight and if made of brass can easily be damaged if it hit the table or some parts and it modify the result of extrusion.
Vatiation between the layers (Z axis) are linked to belt adjustment and extrusion (and, of course, speed) too.
Don't forget the extruder: these bowden type extruder mecanism are subject to have some variation because of the movement and restrictions inside the tubes. I think it could be more precise if it is direct drive extrusion.
haha exactly
I think he meant margin of error, or error relative to the part
makes me feel better that I up-scaled the cube up to 100mm. but I love how your design is because it also helps show holes/bridging even over a short span
This is because of material shrinkage which gets worse with more infill and more walls. So it´s really hard to calibrate to but we can make some approximations.
Like 0.5% shrinkage with PETG, you can now simply scale your part by 100.5% in your slicer and voila.
This is why your cube will be fine since it´s tiny and a 200mm piece will be off by about 1mm.
Also what YOU measure with the cube is the "corner bulge" never measure corners, always round up your calibration pieces since the corners will have acess material. Measure in the middle of the cube, this alone will be more precise by 0.2mm.
my calibration cubes don't have elephant foot problems and the layers are way more even. you have something else going on like maybe loose belts or x,y,z rollers loose. your make sure you z rod isn't bent.
Thanks so much for the video! I've been pulling my hair out trying to get repeatable accuracy but those dang cubes always come out so inconsistent and have inconsistent layers themselves. You made me realize we've been calibrating wrong this whole time.
You definitely earned a subscriber in me!!!!
I like your calibration sheet too. It's great for tracking and to make sure you keep it on point in a schedule
Great video! It’s logical, if you cal small and then print bigger you multiply the margin of error whereas if you cal big and print smaller then you divide the margin !! I print and sell prototypes from the engineer at my job, will be calibrating this way from now on !
Amazing findings, thanks for the calibration design, going to print one now.
Awesome Tip! Never thought about it this way before. Thanks a lot! 😁
I think you are super right about that, even though i already did not use the cube because of my very same experience.
Solid advice. Thanks for taking the time to do this.
Agreed! The bigger the dataset, the more accurate the reading. It averages any minor measuring variables over a much longer distance which drops % of error significantly.
if he can't use vernia gauge, then he will never get accurate data, no matter how big his spreadsheet is
I've only been printing for fun and dimensional accuracy hasn't been critical. But if it's as simple as calibrating with something other than the cube then I'm more happy to get better results from my printer! Thanks!
Wow I have to say this made so much difference. Tried twice with the 20mm cubes and I was still getting results that were out, but with your 100mm test print my printer is now bang on. First print (after trying with cubes) was 99.3, 98.5, 99.6, second test after adjustment was 100.01, 100.03, 100.00.
The 20mm cubes are just too small, so absolutely any printing variation translates into how accurate you can measure them. Turns out that using the cubes made my printer go from over sized at the stock settings to under sized. At worst its now 0.03% out, which to me is bang on.
How do you actually calibrate it though? Let's say I found out I'm off by 0.3mm how which settings do I adjust to compensate for that?
@@1SmokedTurkey1 there are other videos that explain how. It's slightly different for each printer.
@@chrism2964 been specifically trying to find it but I can’t seem to find the right search term “calibration adjustment” doesn’t bring up anything useful. Can you guide me somewhere? I have a Bambu P1S
Go on then, have another subscriber!
I've had an E3V2 for a few months now. I've installed an SKR Mini E3, BLTouch and have spent a good amount of time learning how to calibrate etc. Currently in the process of building a Prusa i3 mk3s for a more reliable print.
It's a never ending learning curve though, videos like this help.
Many ways to calibrate a 3d printer to what you are doing. Most would measure distance traveled on the axis and use Flow Rate to control but you are using a method which works for a specific print. Your elephants foot is either banding or you have the nozzle to close at the start of the print, or both. The banding on the cube you showed looked to be banding. Sounds like you care about your product which is a great thing! Good idea, nice design and quality control! Nice!
How do you calculate the steps pet mm then for X Y and Z with his designe? I don’t understand
I am going to use this on my CR-10. I always thought the same way about the calibration cube. It is so small that the percentage your printer is off is shown by such a small measurement difference in something so small. Your calibration print looks really good.
Excelent Information! Go ahead with this videos!
Excellent video, thanks!
This is amazing! Thank you. I hate using cubes but for some reason I just haven’t stopped using them. Until now. I also really appreciate you mentioning the over extruded corners. So frustrating! I’m going to use your calibration model from now on.
thank you for the video, this will solve so many of my probelms.
If your measurements vary from print to print, what parameters do you adjust to "tighten" things up and get what you're printing to be more accurate?
hi it's realy intreasting what you propose i will try it but in the mean time i have this probleme the outer dimension of my 3d printes parts are accurat but the inner dimension like if there is a hall or square in the middle of the part the dimension of that hall or sqaure are smaller by 0.4 mm in X and Y axis do you have maybe an asnwer for that thank's
I'm just getting started as a maker and this advice should help a lot. Any screw holes I set in my prints always came out slightly oval, which threw everything out, I'll try this and see if that fixes the issue.
TYVM this is the type of info ppl need :D
Hi, you are absolutely right. Thanks, subscribed
How do you work out the correct data to put into cura
Thanks for your advices, are very helpful.
great topic, thanks 👍
thank you friend!
Do we need specific settings for this?
You point out very good discrepancies. No matter the filament, you should get closer tolerances with the same roll and settings.
My Z axis is spot on. but my X/Y axis are off 0.15mm. The external perimeter is 0.15mm too large and internal features is 0.15mm too small...any ideas?
Thanks for your video! I tried your calibration file. But with the updates steps I find that screw holes are now printed elongated. X and Y are no longer equal as in the factory Settings of my Ender 3 pro. Is there more to it?
Hey Stephan, there shouldn't be any difference, as long as your x and y measurements both read 100mm the holes should not be an issue.
Tus vídeos son inspiradores, gracias.
Gracias a ti Adolfo :)
How could I adjust it in the Orcaslicer for X1C Bambu Lab? I have 99,80mm on X, 99,87mm on Y and 99,84mm on Z.
You can't adjust steps on bambu firmware. You have to compensate in the slicer, I don't use those products so I can't give instructions how.
Thanks for the great tip, and for the drawing.
hi. i just started 3d printing but my prints rarely fit lol how do i use your 100mm thing to calibrate??
thanks for this video. it is still unclear how you would calibrate your printer. What are the actual steps you use to dial in your printer
From a Metrology fan who just realized you can calibrate your Esteps, you just earned yourself a sub!
i was just wondering why on earth did anyone thought of using a 20mm calibration cube for something with a 200mm+ bed XD
Nice video! Thanks!
I’m a kind of calibration fanatic!
In my channel I’m always talking about the importance of calibration.
An I almost never print the cube. Just for extrusion multiplier.
Very good! I’d like to make a live with you in my channel. And dont worry about language barrier… It’S funny to translate during the live!
Solid video man thanks! Have a great day
Great stuff. Would love to get your guidance on z offset adjustment and belt tensioning too. Maybe u already covered those two things.
Great video. Like the calibration print and how it avoids the over extrusion corners. Which 3d printer do you consider more repeatable than the Ender 3 and that we can afford?
A thought for you. Why not have your calibration print in some smaller sizes like 20mm, 40mm so we can get the calib close then use your 100mm. Your print shows 3hrs in my Cura setup.
Thanks
Printing your XYZ calibration as I type this. Just before this I did the nozzel extrusion check and after 5 extrudes of 100mm I am getting an average of 100.03mm. Crazy. This is on an Ender 3 S1 Pro. Thanks for all your videos. It’s so refreshing watching someone that’s not a professional RUclipsr reviewing printers that were given to them like “it’s a great printer!” Yeah no kidding they payed you to say that. So far I’m following your advice with leveling and extrusion and stuff and the proof in the pudding is in the eating or whatever that saying is.
While looking online I see people, including me, are printing holes that are about .4mm undersized. Have you come across this? Have any fix for this?
Oh yeah, and I just subscribed. Keep doing what you do. Cheers from Sweden.
They are not payed. You are overextruding because you need to calibrate steps per mm. I have the same printer, after calibrating the steps I get the exact 100mm. A manufacturer can't make the steps exact because it depends on ambient temperature, the filament you are using, among other things, so you need to calibrate it yourself.
I was printing xyz cubes on my ender 5 pro and noticed I have a problem with the z-axis of the cube being “smushed down”. Do you know what it could be?
You could have binding of the z axis. Something may be causing the z axis to not move up smoothly. Check your lead screws and v rollers and make sure everything's is moving smoothly. Try spinning the v rollers by hand to make sure they are not too tight.
Very good information, i have some problems with tolerances, i was thinking in this, good work
Came across this looking for help with my calibration cube but...do you have an STL for your PSA card holder? I can only find the regular stands. Thank you!
Hey Mink sorry I don't have the files available or for sale at the moment.
but how do you use that tool to calibrate?
So what are the ways you adjust the printer then, do you update slicer settings, or do you adjust number of steps per mm, or do you adjust printer's mechanics somehow ... ?
I do not change slicer settings because I have to use the same gcode file for 17 other ender 3v2s that I have. The adjustments I make are gantry leveling and bed leveling. I also check for excessive backlash in my lead screw which I made a video about that.
I watched the video, searched the comments, and looked on thingiverse. But I could not find a mention of the print settings you use for the calibration test? Like how many walls, top/bottom layers, % infill, flow %...etc. So I was wondering, what are your typical slicer settings when testing your calibration part?
The slicer settings should be what you decide to use for your prints. There is no wrong or right answer to your profile. In my printing farm every printer and every file I create use the exact same profile. So that profile is what I use for my calibration print. I do use 10% Infill and walls. The layer height and other settings are your preference!
yup, i always scale my cube to 150mm and only print a few mm high to at least get an idea of my x and y. then i print a 150mm tall cube and adjust accordingly. The smaller the cube or part, the more thermal expansion, and tolerance loss
thanks for this, going to use and calibrate my printer tonight , + 1 sub :)
How do you account for material shrinkage into the Calibration print? If I measure this using PETG should I scale the print this up by 0.8% to account for material shrinkage?
What I have done is I have waited until the part has cooled down to take measurements. Then calibrate. That way the shrinkage is factored into the calibration.
Great video, I will try calibrating my Ender 3v2 with your tower, My prints are a hit or miss at the moment and I'm doing a lot of research lately as to why...
I started maybe a month ago and my prints were flawless, and half way through the new PLA spools I ordered (some cheap Chinese make) my hot-end fan died I started having all sorts of problem with skipping layers, heat creeping ghosting and generally really bad looking prints.
I did get a replacement fan but my prints were never the same.
I tried drying the material in the oven @50 C for a few hours, lubricated and re-aligned the X-axis, cleaned the hot end several times with the needle and cold pulls, got a CR touch for auto bed leveling and upgraded the firmware to Jeyrs. Another thing I played a lot with is slicer settings, copied and tested a bunch of different settings from content creators who use Ender 3 v2 to try and correct those issues.
I am now waiting on new material eSUN PLA+ and Overtune PLA spools, an anti backlash spring (which might help with the elephant foot you're having as well) and a couple of 4020 fans to use with Briss Fang cooler, but one thing I haven't tried is to calibrate my e-steps which I didn't think of because of because I didn't change any parts so far besides replacing a fan and the prints were great when I just started.
Perfecting 3D printing is a long process but I'm so interested in the process I almost don't mind when my prints go wrong because its another lesson lol
Trial and error is definitely the key to learning! Once you dial everything in, its an easy process to rinse and repeat for more printers. The more printers you have the more you start to learn because the issues are replicated across many printer, and when you find a solution, you know it was the right one because all of your remaining printers have the issue corrected as well. I learned the most about the ender 3 v2 when I started managing 6 printers. I now have 17 and still learning everyday.
The single biggest change I made to my v2 regarding print quality is swapping to dual blower fans and the hero me fan ducts. The one dinky fan it ships with was never up to the task and even if the fan was amazing it still only blows one direction. I would also recommend a bowden tube mod using a short piece of capricorn from the back of the nozzle to just before the fitting on top of the heatsink. Then that short run is sacrificial if need be and I use the stock white tube the rest of the way so I can see problems if the arise. I previously had changed the whole tube to the capricorn but ended up having to keep replacing it every couple of months. Here's a video explaining it better than I can.
ruclips.net/video/7tCxO17XZtw/видео.html
Awesome, because I'm doing the same project, except I'm making a case for the case that you make for the manufacturer's case. Seriously tho, this has saved me from printing more 20mm cubes with varied results.
This is really good content, I really appreciate you sharing your experience, and it inspires me to do the same if ever get where you are
Thank you Michael! I hope to continue to inspire and help you make your ideas come to life!
Whats the barrier to entry in doing injection molding? I'd love to know more info about it if I can use my printers to make the molds to then to injections for car parts etc.
I know people are using SLA printers to make molds, haven't seen any made from FDM.
what do you mean by "calibrate to 100mm"? so, what is it exactly I am supposed to do? Adjust E steps? or Flow? So what am I supposed to do after a calibration cube or any other model? thanks
Isn't elephant foot caused by print head being too close to bed in first layer?
In top layer maybe the ironing setting is wrong?
Maybe wrong I don't know just an idea. I am trying to find a solution to get better prints myself.
So far no luck with my Geeetech A20 printer
In this example, the 2nd layer is oozing over the first layer because the z is not moving up high enough as it should. So this is causing the elephants foot.
Hey thx for the information. Im still new to this but does this applies to TPU filament? Thx!
I have never used TPU so I cannot confirm.
temp, flow, e-steps, retractions, etc will make the size vary each and every time. it will never be exactly the same each print. You want really better prints. Get Kipper and do the vibration calibration on it and then use SuperSlicer to make it the best possible slice of a print. when i look at your cube i see the corners not sharp. They are bulging out. I also see that the Z layers. some are bigger then others. i don't know. why but i don't use the narrow part of the calipers. I move the part up to the flat part of the calipers. also when you do that type of measurement. the calipers can be skewed just a little but and not entirely square. point A-B. I wouldn't use that cube for 0.0X calibration. Within. 0.5 is decent.
Klippers is definitely something I want to try out.
How are you imputing your findings into calibrating the extruder. Flow ?
Hi Sean, I’m going to make a video soon on how to calibrate the printers e steps. In the mean time you can view this video that I followed when I Originally calibrated my printer. The only thing I did differently was use my 100mm print, and change his excel formula to reflect 100mm instead of the 20mm in his calculator. It’s an excellent video to get you started
ruclips.net/video/SN9NCn-Dhf8/видео.html
im still getting realy bad elephan foot on the bottom. any advice?
Check out my other video on elephants foot. It has helped a lot of people solve their problem.
Where do you calibrate? Just in scaling? Or do you really dial in the steps/mm first and then just the shrinkrate by scaling?
I dont scale at all. I calibrate the steps per MM
may i ask what should be changed when the calibration is not what we expected? modify design and troubleshoot what's wrong with printer?
Definitely troubleshoot the printer. Your printer should output the exact dimensions your design has at least to .05mm tolerance.
Good advice! TY!
Grate video
Thanks for the calibration print, it really helped calibrate my Ender 5 pro which wasn't possible with the cube. My question is what is the purpose of the holes in the part? Do you think it's necessary? because it only slows down the printing speed.
I ended up removing the holes. So essentially they are pointless lol. It was a newbie mistake
Good video and a way to get better prints. However the real reason for the inaccuracy in the Z is because of the use of lead screws with backlash so the first few layers are compressed (elephants foot). The only way to eliminate this fully is to go belted Z drive.
Can you post the Fusion file? I wanna make one that's 150mm which is pretty much the max of my calipers. at some point id like to try and calibrate for the entire bed. then anything smaller than that should be just fine. also, what are your infill/wall settings? currently, with my settings, it says it will take 3-4 hours.
I use 10% Infill with 2 walls
@@collect3d OK so check this out.. no matter what size I print calibration I am always exactly 0.1 short on the Y-axis vs the X-axis. what would you do to troubleshoot such a thing?
@@rayge if the deviation is the same for every size.. for example over 20mm it is 20.1 over 40mm it is 40.1 over 60m it is 60.1.. then not much is going to help. If you adjust the steps/mm you will introduce a change in dimension as the model size increases so you maybe get 20mm correct but then 100mm will me 0.5mm short.
Also as stated in the video, be aware of the rounded corners bulging and affecting readings! only measure from the top down in the middle of two sides.. or at two points on the sides and average even better.
Great tip!!
Thank you!
If you calibrate using the cube, and get over shoots ONLY, obviously it's going to be wrong. You want to have the avg on 20mm. So between 19.96 or 20.04 would be avg on 20. So even your 100.08 or 100.01 is not on par either, right?
Good stuff man
My first print was a calibration cube and have no problems. The Cube it is exactly 20.00x20.00mm. But well, I don't print with a Creality but with a Longer LK5 Pro.
Can you elaborate on which things you changed on the printer during the calibration? Just extrusion multiplier?
The e-steps for the extruder, x y and z is all I've changed.
I love that spreadsheet completely filled! lol
Thanks lol I try to be thorough
I subscribed because of this video, thank you for the video
Thank you!
Ty i will try it.
If you are selling a product, I think getting pressure advance /linear advance and input shaping set would really raise quality, speed and repeatability.
That is something I definitely want to test out!
What i don't understand: you said your Ender has a tolerance of 0.06mm. You can calibrate to be in the middle of that. But with a longer calibration length you can't change the absolut tolerance of the printer. So being in the middle of the tolerance is enough?
What I have done is print my 100mm calibration multiple times. Get the average range, and then use the middle size to calibrate. Or if your design cannot be larger than a certain size, calibrate to your lowest value plus .06mm to get your maximum desired size. That is how I do my calibration. I don't print much larger than 100mm for my products, so I know there wouldn't be much issue at all with scaling. So you really only have to compensate for the tolerance the machine has.
You are exactly correct in using a longer sample, the longer the sample is the more accurate your measurement will be. I’m a long time DIY CNC guy, and we set out steps per inch in the system utility if we set up our machines using a one inch distance to calculate the steps it will not be very accurate as your 20 MM blocks show, we always to longest distance that we can on each axis this yields a much more accurate steps per inch figure, thus making our cuts all that more accurate.
To speed measurement and accuracy you could set up a little gage stand consisting of a fixed stop and opposite that a dial indicator, The dimension could be mastered with your calipers, this would make measurement faster an more accurate.
This process being an additive process there are more things that effect to final dimension, filament size, pressure, friction and on.
Good luck
Hager
Thank you for the excellent idea on creating a test fixture for consistent measurement! I will look into creating one.
@@collect3d when you are making a living doing this time and quality are so important. And a test fixture will help you so much. Much more accurate than trying to insure you have the narrow calipers making an accurate reading.
@@IH1940HAY what I have been doing is laying my caliper flat on the table, and taking a measurement. The table surface has been my jig to make sure I keep the calipers parallel to the axis I am measuring. The test fixture is a much better alternative! 👍
That's brilliant idea, I was going to get a dial gauge, and this just added to the reasons for getting one
If you’re using a digital caliper, whats its accuracy? The cheaper the caliper, the worse its accuracy.
Z is heavily reliant on temp. Without an enclosure and a way to regulate it, the temperature of your room is going to change the z axis slightly.
Your Z offset is way too low. How get accuracy with such elephant foot?
So how do we use your 100mm print to calibrate?
I have another video uploaded on how to calibrate.
I'm confused as to which orientation I'm supposed to print this. Is the x supposed to be pointing towards me or to the side?
If you are looking at the printer, the X will be pointing to the left side. Hope this helps!
@@collect3d thank you. That helps a lot
good advice thanks
thanks for sharing!
how would you get rid of those layer lines?
Can be done many ways. Tightening the v rollers. The belts tension properly, and a straight lead screw can help reduce layer lines.
Out of curiosity, do you have a video of your calibration process?
I do. Please look at my channel you will see a video on how to calibrate
Great content. You earned a new sub. How about a follow-up video on how you use your model to calibrate your printer?
Thanks Ben! I have a video out that shows how to calibrate. Hope you enjoy!
I noticed that the 20 mm test cubes were off by around 0.05 mm and that the 100 mm test shape was also off by approx 0.05 mm which means the printer is consistent in it's actual dimensional deviation. Creating test shapes that roughly match the largest dimension of a print is a great solution. It should produce more accurate calibration with less effort. Thanks.
The woman who yelled "hey" towards the end just about gave me a heart attack laying in bed watching this at 11:30 at night with everyone else sleeping in the house hahaha
lol we have some bad puppies in the house that always need correcting!
Did you publish the stl
Yes link is in the description.
Good work on your video, question this is only for X and Y only I see when I printed it there was NO Z on it? Or assume that's Z.
That's correct. The z is vertical.
@@collect3d Why I ask It was the corners where not rounded off like on the X and Y. Thank you for the help.
this is why got more from a calibration x and faster the needed changes
from 62 subs at time of upload to 2.7k in a yr. way to go.
when you sit and think about you quite correct, hot plastic will not do what you want all the time, thanks