Problems with ovality on diameters

[Sketch 0]

Hi.

 

I have issues with ovality when printing diameters, both internal and external and of all sizes.

 

The direction and distribution between X & Y of this ovality is always the same.

 

This is not to do with the short belts or out of alignment of the print head rails.

 

To start with and ignoring any small print seams left, I was getting about 0.4mm ovality on a 50.mm external diameter.

After adjusting the short belts and aligning the print head rails I have reduced this to 0.15mm.

To align the print head slides I used a precision 100.mm engineering calibration slip and used this to align each rail in 4 places (at each end and at both limits of axis travel) to its relevant drive / slide rail.

Both X and Y were done in this manner until not discernible error could be detected.

 

I have repeated this procedure several times with no further change. Still got 0.15mm ovality.

 

So, effectively I need to know if there is a method to 'dial' out this last 0.15mm of ovality.

Is there a further hardware related issue I do not know about? Or, can this axis error be adjusted with a setting or software?

 

I really do need to find a solution, as I use our S5 to make Engineering fixtures that are used directly in a Automotive production environment.

[SandervG 1,521]

Hi @Sketch, thank you for your message and welcome to the community! 

Would it be possible to share a photo of your print?

About your ovality, you say that it is 15mm. Is it always 15mm off in a certain direction, or a percentage? If one direction is 15mm off, is the other direction dimensionally accurate or just ~15mm less inaccurate? 

 

Is your Ultimaker stable on its feet, no corners lifted from the surface it is on? Have you ever had any other issues with the Ultimaker that you resolved, or anything else that may not be as it was out of the box? 

 

Thank you,

 

 

 

 

[Sketch 0]

Hi SandervG:

 

Thanks for getting back so quickly.

 

This issue is 0.15mm (150 microns), so is not detectable to the human eye.

I've added a couple of pics anyway.

Largest diameter visible is 65.0mm.

This is an assembled component consisting of.

Black ABS - Main Body.

Silver ABS - Disk Insert (This will be replaced by Turned and Milled PEEK in the final use part - not printed).

Black Nylon - Centre Locator.

There are other Stainless Steel inserts and fixings present (not printed).

 

The ovality encompasses both X and Y to almost equal degree (i.e. both axis combining to give ovality).

Ovality is consistent in both magnitude and direction (therefore this indicates actual mis-calibration).

The machine is stable and does not rock or move.

Machine is in humidity and temperature controlled environment.

I have the additional Accante top enclosure to aid temperature stability.

I print with pretty much every type of material (PLA, ABS, PP, PET, Nylon, PC, PET CF, PAHT CF, PP GF, PA GF etc. etc.) and always found the print quality to be very good.

Only other issues I have had, has been blocked nozzles and dirty feed wheels which accidentally trigger material out alarm. However, both of these issues are to do with printing composite materials and are simple cleaning jobs.

Aside from this, the machine is kept clean and is looked after. I gentle wipe the slide rails clean and oil them once every few weeks, along with greasing the lead screw and cleaning the feeders.

 

I first noticed the issues several weeks ago and found that the main culprit to be loose short belts. However, this got me looking and digging deeper, trying to improve the ovality (roundness) issue. Can't say for sure how long it has been there before this, as have not had to do such a size and shape sensitive part until now. It may well have been there since the start.

 

This type of problem is usually calibration related on most CNC machines and can usually be adjusted in soft settings to cure the problem. Obviously with the S5 being a far simpler device, I would imagine that the adjustment is only possible by hardware, hence asking. However, if stepper motor calibration is possible, this may be the answer.

 

 

 

Many thanks.

 

Sketch

[SandervG 1,521]

Hello,

 

lol, yes obviously it would not be 15mm. That would be ludicrous. I read it correct but I typed it too quickly without giving it much thought. My apologies if I don't understand you correctly, but by saying 'The ovality encompasses both X and Y to almost equal degree (i.e. both axis combining to give ovality).' the 'equally' part makes me think the offset is on both axis and that you might talking about a 'normal' dimensional inaccuracy, not 1 axis more than the other. 

 

I think when you print a square, it might also be around 0.15mm?

 

As you probably know most machines can't operate at a 100% 0,0mm level of accuracy. The same goes for Ultimaker. Especially since working with thermoplastics which get extruded under pressure and which shrink when they cool down there is likely always a small offset. 

Unfortunately it is not possible to calibrate the motors, but perhaps Cura has a few options. Have you ever looked into the setting of horizontal expansion? When set to negative it could compensate for the 0,15mm.

[Sketch 0]

Hi Sander

 

The problem resembles what is known as backlash and due to the shape that is makes, it looks as if the problem is with the Y axis.

 

The image above gives an indication of the shape (or at least distribution of shape) produced.

 

In a nut shell when the Y axis reverses in direction, it does not react in the desired manner to produce a perfect circle. This is a matter of calibration / setting of the stepper motor, assuming no other inefficiencies in the slide rails, belts etc.

 

I did rather think it may just be a machine limitation.

 

As for Horizontal Expansion, I use this all the time for sizing of models but this will have no effect at all on the issue, as it is for size in both X and Y simultaneously. Basically I get the same problem but on a slightly bigger or smaller diameter.

 

Maybe worth suggesting this to the powers that be, as a possible addition to UM machines, to allow fine tuning of stepper motor behaviour. Maybe an additional calibration feature.

 

If I have any other breakthroughs, I will add to the post, or if anyone else has any other ideas, it would be great to hear them.

 

Many thanks anyway.

[SandervG 1,521]

I discussed this yesterday with one of our product experts and we both felt that it might also be good to note that an offset of 0.15mm is not that bad. We can of course help to see if we can isolate a cause but we don't consider this as a strong underperformance. It is possible that the cause lies in a small difference in length between the belts in X or Y, or the spring inside the sliderblock could be of a slightly different tension. Have you checked/compared those during your initial maintenance? 

 

Did you also check if all your rods are firmly pressed and tight into the sliderblocks with no play on them?

Did you ever cut any pieces of your bowden tubes? 

 

I'll pass your feedback along to our developers, it is good feedback. Thanks!