Holes too small.

Hi everyone. I'm really stumped by this and am desperate for help. Every hole in every print is .5mm too small. It doesn't matter what axis the holes are on. It doesn't matter what shape the holes are. If I try to print a part with 8mm holes for smooth rods, the holes are 7.5mm. This happens in Repetier host using Slic3R and Cura and also in Simplify3D. I have tried 2 and 3 outside shells. I have also tried adjusting the extrusion multiplier. My exterior dimensions are not perfect, but very close. My printer is a Prusa i3 steel frame. I have the E3D V6 and E3D Titan. Any suggestions would be greatly appreciated.

Holes allways come out smaller than what you specify,due to shrinkage of the plastic and the fact that
holes are made up of many straight lines rather than pure arcs.
There is a forular to calculate the needed diamater as follows

They suggested that where X mm was the required diameter of the hole then Y mm was the dimension needed to achieve it and that:

a) for vertical holes Y = 1.0155X + 0.2795 mm

b) for horizontal holes Y = 0.9927X + 0.3602 mm

These figures were for PLA, and as I understand it, they take into account various factors including the fact that the inside of the hole is unrestrained against shrinkage.

Just buy some drill bits and drill them out.

I like having the holes come out smaller because I can drill them out to the correct size, get a good finish and know they are the correct fit.

Its much easier to fix an undersized hole then an oversize hole

Quote
obewan
Holes allways come out smaller than what you specify,due to shrinkage of the plastic and the fact that
holes are made up of many straight lines rather than pure arcs.
There is a forular to calculate the needed diamater as follows

They suggested that where X mm was the required diameter of the hole then Y mm was the dimension needed to achieve it and that:

a) for vertical holes Y = 1.0155X + 0.2795 mm

b) for horizontal holes Y = 0.9927X + 0.3602 mm

These figures were for PLA, and as I understand it, they take into account various factors including the fact that the inside of the hole is unrestrained against shrinkage.

The shrinkage makes sense, but I have a friend with the same printer and the same brand of filament, and his come out perfect with no compensation for shrinkage. This confuses me.
I just printed a test piece that is a 20mm circle with a 10mm peg sticking off and a corresponding piece that has a 10mm hole for the peg to fit in. The 20mm dimension printed at 20.18mm and the peg printed at 10.16mm. Not perfect, but close. The 10mm hole printed at 9.56. The height of the part is almost perfect. Also, the slight inaccuracies in the outside dimensions are same no matter the size of the part, so I know it's not a steps per mm issue.

Quote
scottybfg
Just buy some drill bits and drill them out.

I like having the holes come out smaller because I can drill them out to the correct size, get a good finish and know they are the correct fit.

Its much easier to fix an undersized hole then an oversize hole

I'm fine with drilling them out, but its not just round holes. It's any shape. Makes for a lot of filing to get parts to fit.

This is a known problem. There's a thing with several holes to measure the size against the planned or calculated size so you can check the difference. Main reason is the approximation of circles done by (more or less) small lines, the extrusion width affecting both sides of an ideal line, and the shrinking of the extruded plastic.
From my experience the difference between the theoretical hole size against the printed one is smaller if the number of segments to approach the circle is higher. In OpenSCAD try to increase the number by setting $fn=20 or 30 maybe, depends on the size. Cura's holes are closer than Slic3r's holes - Slic3r is (with my Mendel90) about 0.4-0.5 smaller, Cura is 0.3-0.4.

Edited 1 time(s). Last edit at 02/13/2017 02:31AM by Inhumierer.

What is your nozzle size and have you set an extrusion width?
Just thinking that if you over extrude on a curved profile, dependant on how the software calculates it, the problem is larger on the inside than on the outside.
You could run a 20mm test cube and see if the dimension is the same as the round object.
If you get an issue with holes that have a X or Y axis then this would still be relevant as you can be causing the straight sides to be over extruded.

My nozzle is .4mm. I printed a hollow square with .4mm walls and my printed walls are .77mm. Pretty sure that's my problem. Does anyone know how to set extrusion width in Simplify3D?

[/quote]
The shrinkage makes sense, but I have a friend with the same printer and the same brand of filament, and his come out perfect with no compensation for shrinkage. This confuses me.
I just printed a test piece that is a 20mm circle with a 10mm peg sticking off and a corresponding piece that has a 10mm hole for the peg to fit in. The 20mm dimension printed at 20.18mm and the peg printed at 10.16mm. Not perfect, but close. The 10mm hole printed at 9.56. The height of the part is almost perfect. Also, the slight inaccuracies in the outside dimensions are same no matter the size of the part, so I know it's not a steps per mm issue.
[/quote]

If you put your figures into the formula in reverse it tells you the size that the hole will end up if you make no adjustment. SO if we let Y=10mm then X mm will be the resulting size.

10 = 1.0155X + 0.2795 or 1.0155x = 10 - 0.2795. This gives a resulting hole size of 9.57mm diameter which is pretty damn close to the 9.56mm that you measured.

There are various factors involved
1. It depends on the slicer and firmware as to whether the printer can produce a perfect circle, or whether there is an approximation by a series of short straight lines. Most free slicers and most low cost printers will use an approximation. For minimum error part of the straight line will fall outside the true curve and part inside. Thus the inside diameter will be slightly smaller.

2. All prints contract on cooling. Once the "glass" temperature of the filament is reached it becomes a rigid solid and cooling stresses are locked within the print. It is these stresses that will pull the print from the bed and warp it if they are excessive. Where the print is solid, equal cooling stresses act in all directions. Where there is a hole there is no material within the hole to resist them and the local effect is shrinkage of the hole.

3. Most people assume that the cross section of a laid down run of filament is a perfect rectangle. It isn't, the edges of the filament are unsupported and while the filament is molten, surface tension will cause them to round. The resulting cross section is similar to that obtained by cutting lengthways through a sausage. Flat top and bottom with rounded ends. For a 0.4mm nozzle and 0.2mm layer thickness the width over these rounded ends is 0.48mm. The difference varies according to nozzle size and layer thickness (no I don't have further information). This is why you can both see the layers in a print and feel them with a finger nail.

For a hole all the effects will act so as to reduce its size. When considering a pin or plug the first effect will add to the diameter as will the third effect while contraction will tend to reduce it. On top of all the above there is the physical accuracy of the printer itself.

Calibration of the X and Y axes for a given dimension of cube will NOT guarantee that a cube of larger or smaller size will print accurately, because while 1 and 2 are size dependent 3 is not.

The answer to achieving exact precision dimensions or fit, if that is what you must have, is to know the material and process that you are dealing with, and make the necessary adjustments to the design of the model itself to achieve them. After all, if you needed part of something to be exactly 12.5mm high and another part 10mm you wouldn't expect to achieve both with a 0.2mm layer thickness.

Edited 1 time(s). Last edit at 02/13/2017 07:10PM by Supermec.

Quote
Supermec
1. It depends on the slicer and firmware as to whether the printer can produce a perfect circle, or whether there is an approximation by a series of short straight lines. Most free slicers and most low cost printers will use an approximation. For minimum error part of the straight line will fall outside the true curve and part inside. Thus the inside diameter will be slightly smaller.

I agree with all of your points, but this one is even worser. There are no circles in .STL files, so no matter which tool you use to design something, after exporting it into .STL, the circles are history.
Next point: All 3D printers (of which I know) use stepper motors to move parts in some directions. So even if you'd feed the printer with the G codes to make circles, the firmware of the printer needs to approximate the circle by small straight lines, because a stepper motor makes steps, no smooth movements. Yay, that's why they're called "steppers" ;-)

Be careful with what you are adjusting. A 0.4nozle extruding at 0.77 shows way over extrusion. You don't directly fix that with extrusion width.
Those familiar with the software need to step in here because I am still confused by the extrusion width.
Check the filament size, extruder calibration and any extrusion multiplier settings first.

Quote
Inhumierer

Quote
Supermec
1. It depends on the slicer and firmware as to whether the printer can produce a perfect circle, or whether there is an approximation by a series of short straight lines. Most free slicers and most low cost printers will use an approximation. For minimum error part of the straight line will fall outside the true curve and part inside. Thus the inside diameter will be slightly smaller.

I agree with all of your points, but this one is even worser. There are no circles in .STL files, so no matter which tool you use to design something, after exporting it into .STL, the circles are history.
Next point: All 3D printers (of which I know) use stepper motors to move parts in some directions. So even if you'd feed the printer with the G codes to make circles, the firmware of the printer needs to approximate the circle by small straight lines, because a stepper motor makes steps, no smooth movements. Yay, that's why they're called "steppers" ;-)

Yes perfectly correct. No STL file contains a smooth curve. I just hadn't picked that out separately, and as you say the minimum movement distance for a stepper is one microstep (on a really good one !).

Just trying to convince our friend that 3D printing needs thinking and learning about, rather than blindly following a set of instructions.

Edited 1 time(s). Last edit at 02/14/2017 01:06PM by Supermec.

Quote
MCcarman
Be careful with what you are adjusting. A 0.4nozle extruding at 0.77 shows way over extrusion. You don't directly fix that with extrusion width.
Those familiar with the software need to step in here because I am still confused by the extrusion width.
Check the filament size, extruder calibration and any extrusion multiplier settings first.

I'll check into over extrusion, but aside from dimensional problems, I'm pretty happy with my print quality. What other symptoms would over extrusion cause?

At the beginning I had a slight over extrusion, resulted in small plastic warts, especially with high infill and at thin walls.

Quote
MCcarman
Be careful with what you are adjusting. A 0.4nozle extruding at 0.77 shows way over extrusion. You don't directly fix that with extrusion width.
Those familiar with the software need to step in here because I am still confused by the extrusion width.
Check the filament size, extruder calibration and any extrusion multiplier settings first.

It depends how you chose to model and print the square.

If you chose to model it solid and tell the slicer a wall thickness and 0% infill, the resulting wall thickness is generally pretty good. If you actually model it as a hollow box of the required wall thickness, and set that wall thickness in the slicer this sometimes seems to interact with the slicer and produce over thickness walls. So much so that I have had 1.2mm thickness walls print out at 1.6mm thick with a 0.4mm nozzle. I still haven't managed to get to the bottom of it.

Quote
bkohn1
My nozzle is .4mm. I printed a hollow square with .4mm walls and my printed walls are .77mm. Pretty sure that's my problem. Does anyone know how to set extrusion width in Simplify3D?

It seems to depend on how you chose to model and print the square.

If you chose to model it solid and tell the slicer a wall thickness and 0% infill, the resulting wall thickness is generally pretty good. If you actually model it as a hollow box of the required wall thickness, and set that wall thickness in the slicer, this sometimes seems to interact with the slicer and produce over thickness walls. So much so that I have had 1.2mm thickness walls print out at 1.6mm thick with a 0.4mm nozzle. I still haven't managed to get to the bottom of it.

I always print the holes in parts 0.3mm to small, afterward I use a reamer(not a drillbit) to make these holes perfectly to size.
This procedure has a few reasons.

All burrs and starting point(from the nozzle) at the inner perimeter of vertical holes are removed, you get a clean hole.
The holes get perfectly round again so the holes will fit perfectly around the rods or bearings.
Using a reamer you actually can decide how tense the rods or bearings will fit into the holes because all reamers are a tiny bit thinner at the tip. So you can decide whether you push the reamer through completely for a normal to loose fit or push the reamer just a little bit in to have tight to press-fit.
Widening horizontal holes you get rid of the flat upper side of the holes. That upper flat side will gently be removed(scraped away) by the reamer. A drillbit will force itself in the obstruction, often widening the hole in the wrong direction.

And yes you can safely buy cheap reamers to do the job in printed parts. I would not use these cheap reamers to use in metal parts, just preserve them for plastic.
I'm the lucky owner of a mill and made a few reamers myself. I often use 625(8x22x7) ball bearings, so for a (maybe) 1 time use(for this project) I decided to make my own 22mm diameter reamer.

Hope it helps...

Chris

So is there like a magic number? I dont want to calculate the y=1.0something X = something else.
Can i just have a number like: x=_, now multiply x by 1.____ in the modeling software to get the actual size in real life, and also get the y multiplication. also, wondering if the z axis gets any problems? just curious..

I am using slicer, printing on migbot. yeah, ive tried printing my own printer parts, ive ended up using allen keys and running them through the holes and twisting to get them big enough. frustrating, i understand

You have been given the magic numbers. These formulae were obtained from a professional 3d printing firm, and are what they use to provide the correct results. If there was a simpler way to arrive at the answer, then I would have given it to you. This is about the simplest formula that you can get, it's not advanced mathematics.

Edited 1 time(s). Last edit at 02/18/2017 10:18AM by Supermec.

Quote
Supermec
You have been given the magic numbers. These formulae were obtained from a professional 3d printing firm, and are what they use to provide the correct results. If there was a simpler way to arrive at the answer, then I would have given it to you. This is about the simplest formula that you can get, it's not advanced mathematics.

You are right. i was confused with the x and y, though they were axis values... lol.
do these values work with squares and rectangles???

In reply to the question, no they don't work for squares and rectangles because one of the factors involved is not present in these shapes. The best thing to do is to design yourself some simple test prints that don't use a lot of filament and have an experiment. The shrinkage factors for PLA and ABS are available on the net and they are different. Have a play, look at the results and then have a think about them and what is involved. That way you will learn far more about the 3D printing process.

By the way, you got it slightly wrong on your post under one of the other subjects. Vertical holes refers to those like a coin lying on its face, the same as you would get using a drill press. Horizontal holes refers to those like a coin standing on its edge, or a tunnel.

Edited 1 time(s). Last edit at 02/20/2017 06:48PM by Supermec.

Oh ok. i will test. but my printer is still broken...
I'm not too keen on making fitting parts, I'm really more interested in printing models, so the scale usually doesnt apply to me.

Edited 1 time(s). Last edit at 03/05/2017 12:17PM by Hedgehog.