How to do X/Y dimension calibration?

[Tigerbeard 4]

I noted on a test model of 20mm in X that the print was actually 20.6mm. Thats about 3% error.

I can get similar values in Y.

 

Edit: I changed the topic, because I realized, that the XY Calibration in the maintenance menu of the printer does not adjust this setting at all (I mean this procedure).

So when my printer does not print 10mm as 10mm, I need to change a factor, not an offset.

 

Is there a procedure for that?

In the official list of UM3 maintenance procedures here nothing like that is listed

 

 

Edited July 18, 2020 by Tigerbeard

[Smithy 1,135]

What you mean is the calibration of the steps, but this is not needed on the UM printers, they are calibrated very well and there is no way to change the steps without hacking into the firmware.

 

To get better results and more dimensional accuracy you want to print slower (<30mm/sec) and set all speeds to the same value. 
 

Search the forum for dimensional accuracy, there are a lot of threads with advices which settings are relevant to get a better results.

[Tigerbeard 4]

This seem to be exactly my problem. I found it with a google search outside the forum 🙂

 

Ultimaker Forum Post 18879-how-to-calibrate-xy-and-z-steps

 

Is here someone from Ultimaker who can put up this information on Ulitmaker Support / Printers/ Ultimaker 3 / Troubleshooting.  Currently there is only one very lonely entry anyway..

[Smithy 1,135]

Not really, you found how to hack the steps which is not needed. If you now follow these instructions if will just fit for your example model and the next one will be different.

 

Search for dimensional accuracy and follow the steps there, print slower, print cooler helps a lot without tinkering into the steps calibration.

[Smithy 1,135]

I wrote sometime ago this article and it covers the important settings:

 

In order to print dimensionally accurate parts, we need to adjust some parameters. Each one of them helps us to go one step further, and in combination we have a perfect engineering profile which we can save in Cura. In principle, the following rules apply to any 3D printer, but some values may need to be adjusted for other printers.

 

Some golden rules

1. Print slow
   The slower the printing, the more positive the effect on dimensional accuracy. Ideally, you should print between 25-45mm/sec.
2. Print cool
   The best results are obtained by printing as cold as possible. One way to find the right printing temperature would be to reduce the temperature until you see under extrusion and then raise it 5-10°C again.
3. Unify speeds
   One of the most important rules is to set all speeds in Cura to the same value. Cura usually calculates the print speeds according to a specific formula. Infill, for example, prints much faster than an outer wall. Due to the permanent acceleration and deceleration, a certain under or over extrusion always occurs, which in turn negatively influences our dimensional accuracy. 
4. Acceleration & jerk control
   Acceleration and Jerk Control decelerate the print head before and accelerate it after a change of direction, so that less vibration occurs when printing e.g. a 90° corner. However, as with the different speeds, over- or under-extrusion can also occur. This can be seen very clearly at a 90° corner, which is then slightly bulged outwards. For dimensionally accurate prints, the two options should therefore be switched off.
   But there is also a disadvantage, because if the print head then goes around the corner unbraked, vibrations arise which then show up through so-called ringing on the surface. But this ringing is only optically visible and not really measurable. So for technical parts, where the surface is not important, but the dimensional accuracy, Acceleration and Jerk Control should still be switched off.
   You have to find a balance between ringing and dimensional accuracy. If the surface quality is important, the values for Acceleration & Jerk Control can be adjusted to obtain a good balance.
5. Elephant foot
   The so-called elephant foot is created on the first layer, where the material is pressed into the print-bed to achieve optimum adhesion. This also displaces the material and creates a lateral bulge. The object is then slightly thicker on the underside than in other places.
   The "Initial Layer Horizontal Expansion" option can be used to adjust this behavior. Common values would be -0.2 or -0-3 mm depending on nozzle diameter.

 

Engineering profile settings for Ultimaker Cura

Line Width: 0.4
Wall Thickness: 1.2
Wall Line Count: 3
Top/Bottom Thickness: 1.2
Print Speed: 40
Infill Speed: 40
Wall Speed: 40
Outer Wall Speed: 40
Inner Wall Speed: 40
Top/Bottom Speed: 40
Print Jerk: 20
Infill Jerk: 20
Wall Jerk: 20
Outer Wall Jerk: 20
Inner Wall Jerk: 20
Top/Bottom Jerk: 20
Horizontal Expansion: -0.03
Initial Layer Horizontal Expansion = -0.2

Additional settings you can try
Inital Layer Height = 0.1
Slicing Tolerance = Exclude
Enable Jerk Control = not checked
Enable Acceleration Control = not checked
Outer Before Inner Walls = checked
Combing Mode = off

 

Finally, an important note

There is no universal profile that fits everywhere and delivers good results. Too many factors play a role here, the filament to be printed, but also the ambient temperature and humidity. You will get different results on hot humid summer days than on cold dry winter days. Therefore one should always make a few test prints before a larger print job and if necessary adapt the parameters to the conditions.

[Tigerbeard 4]

3 minutes ago, Smithy said:

What you mean is the calibration of the steps, but this is not needed on the UM printers, they are calibrated very well and there is no way to change the steps without hacking into the firmware.

 

To get better results and more dimensional accuracy you want to print slower (<30mm/sec) and set all speeds to the same value. 
 

Search the forum for dimensional accuracy, there are a lot of threads with advices which settings are relevant to get a better results.

Thanks for your post. I was typing when you posted it. In the link there was an explanation for what you say - the geometry makes the step sizes accurate. However, it seems my 0.3mm error is not a relative but absolute error. That means the smaller the parts are the more severe it gets.

 

I think it is really a must-read for anyone designing with tolerances of what the printer could do (via spec) to understand what needs to be done to actually be able to print that. Especially if the "instustrial" version S5 is having the same issues.

 

I have the printer for a few weeks now and I think the forum is quite nice, but realistically its really very inefficient. Hundreds of posts are about the same stuff all over again. The search ist little help if I need to post to find out which keywords I have to use. 3D FDM is imho such a learning curve, that a collection of "seleced-reading" for different audiences would really speed up things. Why don't you open a new forum section where moderators can link/copy/move good question/answer posts of common encountered issues? I rather spend 5 hours to read through 100 selected posts by the admins on 10 pages than to search 22.000 posts god knows how many pages, many of them not answered.
Sorry if it sounds ranting, thats not intendend, it just that I can not believe to see that the moderators are bearing this for years now.

[Tigerbeard 4]

1 minute ago, Smithy said:

I wrote sometime ago this article and it covers the important settings:

...

 

There is no universal profile that fits everywhere and delivers good results. Too many factors play a role here, the filament to be printed, but also the ambient temperature and humidity. You will get different results on hot humid summer days than on cold dry winter days. Therefore one should always make a few test prints before a larger print job and if necessary adapt the parameters to the conditions.

@Smithy thanks for your post.

Actually thats exactly what I mean, when i thought some posts shoud be tacked and collected somewhere.

 

I see I have a lot of reading to do, but please allow me one more question to your post. I have a stack off all sort of Filaments to use, but got me a fresh sealed UM roll because I wanted to have a tuned material profile for my "leaning curve". I would take from your reply, that this perfect match profile of UM material to the UM3 would not really help me for dimension accurary on a millimeter level. Of in other words, would you confirm that the tweaks are required regardless of the material used?

 

Thanks

[Smithy 1,135]

Some of us uses personal "templates" to help with the correct answer, a knowledge base would maybe helpful, but 3D printing is so individual in each case that it is hard to give general settings. And 3D printing is a steep learning curve but you will get the needed adjustments very quickly when you print more often and try some of the settings.
 

I posted above an article about dimensional accuracy. These settings are the key to get a better result. The S5 has these settings in an engineering profile, only available for the S5, but you can set it on your own in Cura to get the same result.

 

I know that reading hundreds of posts seems annoying, but I did it too when I started printing and it helped me a lot. You get a much better understanding how all these settings are working and you will get a pro soon. A knowledge base article helps for a very specific case, but you will not learn so much from it. 

[Smithy 1,135]

8 minutes ago, Tigerbeard said:

I would take from your reply, that this perfect match profile of UM material to the UM3 would not really help me for dimension accurary on a millimeter level. Of in other words, would you confirm that the tweaks are required regardless of the material used?

The material profiles are important to get some "basics" like temperature, speed, retractions and so on. But it will not help much with your problem. Based on your needs you have to configure Cura to do the things you want. Some want to print perfect looking parts which needs other settings than when you want to have something with good dimensional accuracy. It sounds complicated now, but it is not. Some key settings like print slow is important for all cases. Also to equalize the speeds is helpful in most or even in all cases. Just when you want to print some drafts you can go faster and use the defaults.

 

But it is good that you use a good filament for your tests. It is important that the filament diameter is consistent over the whole spool. Other brands are also good, but some cheaper filament can make problems and you look forever for the right settings until you change the filament.

 

And to answer finally your question, yes the tweaks above are helpful for all materials. I always use these settings with nearly every print.

[Tigerbeard 4]

Just as a a note to myself seom related stuff fomr  forum

 

• 23-a-visual-ultimaker-troubleshooting-guide, an external link, but very good with visual examples of problems and extensive solution text. The don't have something on dimensional accuracy, though
• 33239-profile-for-4mm-strong-walls-using-pla, Very interesting post and inho a very related issue. Unfortunately unresolved.
• 29632-getting-better-cylinder-walls Slightly different topic, but high impact from a source one would not suspect as newbie. Good to know before printing very fine/thin structures

[Tigerbeard 4]

19 minutes ago, Smithy said:

the tweaks above are helpful for all materials. I always use these settings with nearly every print.

@Smithy, thanks for those very helpful posts.

 

31 minutes ago, Smithy said:

A knowledge base article helps for a very specific case, but you will not learn so much from it. 

Thanks for the kind words. Interestingly, I had some steep learning curves in the past, the last being two CAD systems. So I know what you mean, but still I think only a fraction of posts bring you some insights 🙂

[Tigerbeard 4]

19 hours ago, Smithy said:

I wrote sometime ago this article and it covers the important settings:

Engineering profile settings for Ultimaker Cura

Line Width: 0.4
Wall Thickness: 1.2
Wall Line Count: 3
Top/Bottom Thickness: 1.2
Speeds: 40
Jerks: OFF
Horizontal Expansion: -0.03
Initial Layer Horizontal Expansion = -0.2

Slicing Tolerance = Exclude
Enable Jerk Control = OFF
Enable Acceleration Control = OFF
Outer Before Inner Walls = checked
Combing Mode = OFF

 

 

After some initial issues I printed a 20mm test cube with these settings and the dimension measurements were

 

 X= 19,97 +/- 0,02

 y= 20,00 +/- 0,02

 z= 20,02 +/- 0,02

 

Also the walls looked fine and the imprinted letters were good. There were no bulges at the corners. The whole surface had a few tiny bits sticking out, but a very short sanding over smoothed those out.

These are excellent results. Great thing, this profile Thanks again!

 

Currently I am doing a "real" part and I am noting that with PVA support the setting is not perfect. I guess its the "Combing Mode OFF" parameter. After each line the printer lowers the table and moves to the next line. For PLA this is fine and with those settings, I have very few short segments. But the support structure is full of micrometer segments and the table is moving up and down all the time. The problem is, however, that each time this leaves a "spike" of PVA material. So the support surface in Z-direction looks like the lunar surface. I was really surpurised that the nozzle did not rip of the lower layers all the time. I see how the parts are, but I expect the supported surfaces to come out really bad. I am not sure if I can set the combing mode different between both extruders (I guess there is not way to have Cura optimizing the support structure for maximum segment length during slicing).

If the support in a layer is always printed second, I think the suport settings could be more relaxed and closer to the standard without affecting part accuracy but speeding up print and lowering printer mechanical load and wear.

 

 

[AurelienC 0]

Hi

 

Thanks for all these advices. I actually looked at the Ultimaker s5 ingenering profile, and in there, the "Enable Jerk Control" is activated. Also, the speeds are not unified . I printed a clearance test using this profile on the UM3, and the result seems not better than the settings given in this post.

[my3DBr 4]

Hello guys,

Are you still around?

I have a UM2E and I’m getting consistently bigger dimensions in Y than in X, no matter what  I’m printing.

Circles that needs precision (like bearings) como slightly oval, I can see it in that little gap.

Squares, Cubes, everything. A calibration 20mm cube is consistently giving me 20.1 on X and 20.5 on Y.

A rectangle 60x40mm on X gives 60.15 while on Y I get 40.5.

A 8mm cylinder gives me 8.1 diameter on X and 8.4 on Y.

All pulleys ok, belts tensioned. Draft o.2 on Cura, everything basic. PLA 200C on nozzle.

Working with horizontal expansion don’t affect results.

I installed a needle on the print head, put a good technical ruler on the bed and stepping seems pretty good through all the axis.

any clues???

[Torgeir 212]

I've just answered this issue in another tread 15 minute ago.

 

Thanks

Torgeir

[my3DBr 4]

Thank you.