Help, internal size of part(s) too small...

Posted by Qsilver
 Re: Help, internal size of part(s) too small... December 18, 2013 08:34PM Registered: 12 years ago Posts: 2,947
Quote
tmorris9
OK, I know just enough about gcode to be dangerous. True my assessment of the tool patch is based on assumption as no setting I have tried (hundreds and all I have seen across several boards) produces the correct size. I admit that it was a logical conclusion rather than fact.

So if you look at the enclosed image, I have highlighted an inner square hole (I wanted to eliminate issues with radius problems for this example) can you tell me based on the highlighted info what the size of the hole is in this object. And if possible explain to me how you calculate the size using the gcode?

Thanks!

That's a perfect way to do it. If we look at the X min and max coordinates and/or the Y we can then do some maths.

With the numbers you have highlighted we get.

X = 98.175 - 87.825 = 10.35
Y = 118.175 - 107.825 = 10.35

Which tells us you are printing a 10mm hole with a 0.35mm path. If that is not correct then you have proven the problem exists.

The strange thing is the hole looks like a rectangle not a square but the maths says otherwise.

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 18, 2013 08:37PM Registered: 11 years ago Posts: 1,171
Very interesting, here is the same basic info once sliced in Kisslicer. A few things I noticed right away are that Kisslicer lays down the very inner perimeter first while Slic3r lays down the inside inner perimeter first. Next is that slic3r used 4 lines to create the inner perimeter while Kiss was like 21 or 22 lines to create the same 4 lines of perimeter. I wonder if either of these could be an issue as to why slic3r won't make the right sized holes?
 Re: Help, internal size of part(s) too small... December 18, 2013 08:39PM Registered: 11 years ago Posts: 1,171
The hole looks odd as it was taken at an angle. But yes it's a 10mm hole and the nozzle is .35mm
 Re: Help, internal size of part(s) too small... December 18, 2013 08:49PM Registered: 11 years ago Posts: 1,171
If you do any woodworking or manufacturing that uses cutting tools you must account for the kerf of the of the cutter be it blade or router... In additive machines I would think it would be the reverse of this. So, based on a .35 path do those numbers account for half the patch width? (.175mm) Still I would be very happy if Slic3r were that close but none of the tests I have printed come close to that accuracy.
 Re: Help, internal size of part(s) too small... December 18, 2013 09:02PM Admin Registered: 17 years ago Posts: 7,879
I could work it out but I would need to know the layer height to work out the filament width from the E numbers. It isn't anything to do with the nozzle width.

Edited 1 time(s). Last edit at 12/18/2013 09:04PM by nophead.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 18, 2013 09:24PM Registered: 12 years ago Posts: 2,947
Quote
tmorris9
Very interesting, here is the same basic info once sliced in Kisslicer. A few things I noticed right away are that Kisslicer lays down the very inner perimeter first while Slic3r lays down the inside inner perimeter first. Next is that slic3r used 4 lines to create the inner perimeter while Kiss was like 21 or 22 lines to create the same 4 lines of perimeter. I wonder if either of these could be an issue as to why slic3r won't make the right sized holes?

Yes Kisslicer rounds off the corners to prevent blobs on outside corners. The fact it does it on inside corners is actually a Bug that is known but does not seem to cause to many issues. It was supposed to be fixed in the next release but that does not look like it will ever happen (kisslicer forum currently 15,000 new spams in one week and a missing developer for 3 months).

But if we look at the numbers from Kisslicer we can see something else.

X 98.19 − 87.81 = 10.38 which says it is inset an additional 0.03mm. (this may be a result of the seam hiding and should be turned off to confirm)

Edited 3 time(s). Last edit at 12/18/2013 09:53PM by Sublime.

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 18, 2013 09:45PM Registered: 11 years ago Posts: 1,171
Quote
I could work it out but I would need to know the layer height to work out the filament width from the E numbers. It isn't anything to do with the nozzle width.

My layer height was .2mm
 Re: Help, internal size of part(s) too small... December 19, 2013 12:19AM Registered: 12 years ago Posts: 2,947
Quote
tmorris9
The hole looks odd as it was taken at an angle. But yes it's a 10mm hole and the nozzle is .35mm

Do you have the extrusion width manually set to 0.35mm or do you have the nozzle size set to 0.35 and auto extrusion width?

Also did you see my last post about Kisslicers results?

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 19, 2013 06:12AM Admin Registered: 17 years ago Posts: 7,879
Quote
tmorris9
Quote
I could work it out but I would need to know the layer height to work out the filament width from the E numbers. It isn't anything to do with the nozzle width.

My layer height was .2mm

I also need to know the filament diameter you have set in the slicer of course.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 19, 2013 04:53PM Registered: 11 years ago Posts: 1,171
Sublime : .35mm nozzle -0- (default) extrusion width though I have tried others.

Nophead: it's 3mm but actually 2.87mm and that is how it's set in Slic3r. Tried 3mm to see but holes still the same.
 Re: Help, internal size of part(s) too small... December 19, 2013 05:03PM Registered: 11 years ago Posts: 1,171
So I ran 6 prints f that object last night trying various settings (did not write them all down). Like larger and smaller nozzle diameters, super slow speed (5mm/s). Consistently my 10mm hole was 9.40mm. The only setting that seem to do anything was changing perimeters from 2 to 1 and then I got a 9.45mm hole, a small change but in the right direction and it may be a clue to I did one print with no top/bottom layers and no infill (just a single inner perimeter not attached to anything else) and it was even closer at 9.5mm.

Unless anybody has any ideas, I am done testing for awhile and will use Kisslicer on anything that needs to fit together properly. I really like the ease of use of Slic3r to so that's too bad.
 Re: Help, internal size of part(s) too small... December 19, 2013 06:31PM Registered: 12 years ago Posts: 2,947
Quote
tmorris9
Sublime : .35mm nozzle -0- (default) extrusion width though I have tried others.

Well I think you may have proven the problem is with Slic3r . If you look at the beginning of the Slic3r gcode file ( the one you posted the picture of with the coordinates) you will see a comment on the extrusion width slic3r decided on. In auto this would something greater than the nozzle diameter. That extrusion width is what the inset needs to be half of not the nozzle diameter. I would love to know what the extrusion width is in the file.

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 19, 2013 07:10PM Admin Registered: 17 years ago Posts: 7,879
Calculating the width from the E numbers I get 0.4mm, but that would only make it 0.05mm too small, not 0.5mm. Multiple outlines would cause a bigger error if they are done from the inside of the object first.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 19, 2013 07:58PM Registered: 11 years ago Posts: 1,171
Here is the top of the gcode

; generated by Slic3r 1.0.0RC1 on 2013-12-18 at 17:06:30

; layer_height = 0.2
; perimeters = 2
; top_solid_layers = 3
; bottom_solid_layers = 3
; fill_density = 1
; perimeter_speed = 80
; infill_speed = 80
; travel_speed = 300
; nozzle_diameter = 0.35
; filament_diameter = 2.84
; extrusion_multiplier = 1
; perimeters extrusion width = 0.35mm
; infill extrusion width = 0.52mm
; solid infill extrusion width = 0.52mm
; top infill extrusion width = 0.52mm

G21 ; set units to millimeters
M107
M104 S235 ; set temperature
G28 ; home all axes
M109 S235 ; wait for temperature to be reached
G90 ; use absolute coordinates
G92 E0
M82 ; use absolute distances for extrusion
G1 F1800.000 E-2.00000
G92 E0
G1 Z0.200 F18000.000
 Re: Help, internal size of part(s) too small... December 19, 2013 08:27PM Admin Registered: 17 years ago Posts: 7,879
>; filament_diameter = 2.84

You said 2.87 earlier!

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 19, 2013 08:41PM Registered: 14 years ago Posts: 1,092
tmorris9: Are you completely sure that your machine moves 10mm when you tell it to move 10mm in X and Y?

Marlin for example comes with the default value for X and Y steps of 78.7402 (which is for the Ultimaker). Using this on my old machine, 10mm was actually about 9.8mm. Turned out that the value I needed to use was exactly 80. The difference in my case isn't huge, but perhaps your machine is different.
 Re: Help, internal size of part(s) too small... December 19, 2013 09:10PM Registered: 11 years ago Posts: 1,171
Quote
>; filament_diameter = 2.84

You said 2.87 earlier!

This comes from a mixture of faulty memory and dyslexia ;-) 2.84 is the right number.
 Re: Help, internal size of part(s) too small... December 19, 2013 09:18PM Registered: 11 years ago Posts: 1,171
Quote
Cefiar
tmorris9: Are you completely sure that your machine moves 10mm when you tell it to move 10mm in X and Y?

Marlin for example comes with the default value for X and Y steps of 78.7402 (which is for the Ultimaker). Using this on my old machine, 10mm was actually about 9.8mm. Turned out that the value I needed to use was exactly 80. The difference in my case isn't huge, but perhaps your machine is different.

Not 100% sure. I bought my machine "pre calibrated" from a reputable source, he use to be an automotive engineer and worked with calibration of milling machines and such. And the fact that the outsides come out near perfect (+/- .02mm) I would think it's right. My machine came with Repetier firmware. I noticed the firmware is older 0.62 VS. the latest 0.83. Not sure if that would make a difference. I am not very versed on what all is involved in updating it and I do use my printer for important stuff so I can't afford to mess it up. Don't know if I should try to upgrade it.

Besides, lets assume that my printer is at fault, why would so many other people be suffering the same problem?
 Re: Help, internal size of part(s) too small... December 19, 2013 09:36PM Registered: 12 years ago Posts: 2,947
@ tmorris9

Are you sure you have all the extrusion widths set to 0 (auto) as it seems unusual for it to be ( ; perimeters extrusion width = 0.35mm ) in auto mode unless something has changed recently. But either way it is trying to produce a 0.35mm path which is correct for the inset we see in the previous code so that conforms there is no bug in path placement.

These seem more like the numbers I would expect from auto:
; infill extrusion width = 0.52mm
; solid infill extrusion width = 0.52mm
; top infill extrusion width = 0.52mm

@ Nophead are you simply doing a square cross section or a rounded one for your numbers? I would like to know both so we can at least say whether it is doing one of those or something else completely wrong. Also is it possible the extra 0.03mm that Kisslicer inset the path is to allow for the rounded edge?

Edited 1 time(s). Last edit at 12/19/2013 10:14PM by Sublime.

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 19, 2013 09:54PM Registered: 11 years ago Posts: 1,171
Here is a pic.

What are the proper methods for calibration (I would like to go through and check it myself)?

And I assume the calibration is done in the firmware? Any tips on this for Repetier?
 Re: Help, internal size of part(s) too small... December 19, 2013 10:09PM Registered: 11 years ago Posts: 1,171
By the way. I know this thread was started by Qsilver (sorry for hijacking it).

And I really do appreciate Sublime & Nophead for taking the time and effort to help.

Just wanted to give thanks where it's due.
 Re: Help, internal size of part(s) too small... December 19, 2013 11:58PM Registered: 14 years ago Posts: 1,092
Quote
tmorris9
Quote
Cefiar
tmorris9: Are you completely sure that your machine moves 10mm when you tell it to move 10mm in X and Y?

Marlin for example comes with the default value for X and Y steps of 78.7402 (which is for the Ultimaker). Using this on my old machine, 10mm was actually about 9.8mm. Turned out that the value I needed to use was exactly 80. The difference in my case isn't huge, but perhaps your machine is different.

Not 100% sure. I bought my machine "pre calibrated" from a reputable source, he use to be an automotive engineer and worked with calibration of milling machines and such. And the fact that the outsides come out near perfect (+/- .02mm) I would think it's right. My machine came with Repetier firmware. I noticed the firmware is older 0.62 VS. the latest 0.83. Not sure if that would make a difference. I am not very versed on what all is involved in updating it and I do use my printer for important stuff so I can't afford to mess it up. Don't know if I should try to upgrade it.

Besides, lets assume that my printer is at fault, why would so many other people be suffering the same problem?

I asked because I was trying to rule that out 100%. I was also wondering if perhaps we had an accumulation of errors that produced the right widths for the outside of objects, but that screws up the internals.
 Re: Help, internal size of part(s) too small... December 20, 2013 01:21AM Registered: 11 years ago Posts: 1,171
Like I mentioned I am not a software guy so it's possible something in firmware is set wrong but I don't have a clue what it might be and how to fix it.
 Re: Help, internal size of part(s) too small... December 20, 2013 02:00AM Registered: 11 years ago Posts: 1,171
So I decided to test accuracy a way I could figure out. So I placed solid metal clamps exactly 101mm away from a known flat point on my X axis and used Repetier Host to move it 100mm (I did this so I would have a measurable gap and remove the possibility of it hitting and moving the marker. Using my trusty caliper it left a 1mm gap. I did this several times on both the X and Y axis (did not bother with the Z as it's not in question).

This may not be the normal method for checking calibration but I think it was good enough to tell me it's accurate. Or am I missing something?
 Re: Help, internal size of part(s) too small... December 20, 2013 05:11AM Admin Registered: 17 years ago Posts: 7,879
@Sublime,
With the corrected filament diameter of 2.84mm the flow works out exact for a 0.35mm x 0.2mm rectangle but for the actual rounded shape it comes out at 0.393mm. For multiple outlines it will have a cumulative error but I haven't done the maths for that yet. Note this would also make the outside of the object bigger as well.

I use a scale factor of about 1.005 for shrinkage in Skienforge, which might explain why outlines don't seem too big..

Edited 1 time(s). Last edit at 12/20/2013 05:11AM by nophead.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 20, 2013 11:54AM Admin Registered: 17 years ago Posts: 7,879
Once the first outline is laid down, each subsequent one is bounded by the rounded edge of the previous one. This means the volume to fill the rectangular area actually contributes an additional width equal to h/2. So for example it it does three outlines expecting them to be 0.35mm each = 1.05mm they will actually be 0.393 + 0.45 + 0.45 = 1.293 which is 0.243mm too wide. That would reduce a hole by about 0.5mm.

The mystery is why the outside is not affected, unless it does the outlines from the outside in. Then the error is only 0.043mm and probably gets lost in shrinkage.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 20, 2013 03:10PM Registered: 12 years ago Posts: 2,947
Thanks for doing the maths. At least we now know it places the paths in the same place as other slicers do which is half the width of a square extrusion profile. This leaves us with even more of a mystery, why can people get the correct size holes with other slicers like Kisslicer? Is it because of the order it does hole perimeters? Is this an adjustable setting in Slic3r? Does the problem exist if there is only 1 perimeter?

@tmorris9
Yes that is basically how you check calibration. You can also clamp one end of your calipers to the bed/carriage and the other end to the frame or even easier use a dial gauge.

 FFF Settings Calculator Gcode post processors Geometric Object Deposition Tool Blog Tantillus.org Mini Printable Lathe How NOT to install a Pololu driver
 Re: Help, internal size of part(s) too small... December 20, 2013 03:26PM Admin Registered: 17 years ago Posts: 7,879
I am writing a blog article at the moment. The error is a quadratic function of layer height, so people who print thin layers are less likely to notice and can probably eliminate it by reducing the flowrate without the infill getting too sparse. The error is there for a single outline but gets a lot worse if you have several outlines and do the outer skin last. Not only is the error cumulative but the second and subsequent outlines contribute bigger errors because they are constrained at one side.

By default kisslicer does the outer skin first and recommends that for accuracy.

[www.hydraraptor.blogspot.com]
 Re: Help, internal size of part(s) too small... December 20, 2013 03:56PM Registered: 11 years ago Posts: 1,171
I just noticed there is a setting in slic3r called "External perimeters first".
 Re: Help, internal size of part(s) too small... December 20, 2013 06:54PM Registered: 14 years ago Posts: 1,092
Quote
Sublime
Thanks for doing the maths. At least we now know it places the paths in the same place as other slicers do which is half the width of a square extrusion profile. This leaves us with even more of a mystery, why can people get the correct size holes with other slicers like Kisslicer? Is it because of the order it does hole perimeters? Is this an adjustable setting in Slic3r? Does the problem exist if there is only 1 perimeter?

I've previously seen the same problems with only 1 perimeter, but I haven't used the last 4 releases of slic3r (been a bit busy with electronics stuff).

Quote
Sublime
@tmorris9
Yes that is basically how you check calibration. You can also clamp one end of your calipers to the bed/carriage and the other end to the frame or even easier use a dial gauge.

I think tmorris9 measured it perfectly. It also avoids any issues caused by motor oscillation at the end of a move throwing out the callipers by a small amount (mainly the callipers being bumped slightly further at the end of a move if they were not directly attached to the moving part of the axis being measured).

A lot of people put stock in the fact that their finished objects are the "correct" size as being their calibration, when that is affected by shrinkage, etc. If your movement is wrong but the finished outside dimensions of the object are correct, then it's quite possible that the inside dimensions would not be correct if your extruded width is also wrong or you haven't allowed for shrinkage.

tmorris9: Glad we've eliminated the printer. Sorry to have doubted you.
Sorry, only registered users may post in this forum.