PK
Peter Kriens
Wed, Sep 4, 2024 7:35 AM
I am very new to 3D modeling and printing. For my project I picked OpenSCAD because I love the programmatic approach and I need lots of variations on the same core components. After some grueling time to learn OpenSCAD and the (complex but very good) BOSL2 library I built a number of prototypes and printed them on a Ultimaker S3 and tested them with focus groups.
I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
To my horror I found that STEP is the standard file in the IM world, they genuinely hate STL files it seems. I can convert the STL files decently through Fusion360 but this is almost sufficient to exit OpenSCAD. Conversions are always lossy and it is kind of worrying that OpenSCAD is this not used in the IM world.
3D printing has the issue of support >45 degree overhangs, IM requires that the mold can be easily released freely. This profoundly influences the design. There are also the issues of plastic uses, sink marks, warpage, etc. Curious if anybody has thoughts about this.
As a software developer I'd like to have a single source but the IM people seem to just want to take the STEP file and then they do their magic to it. This not completely insane since the mold making process is very expensive.
In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
My design requires very accurate holes for connectors like USB, HDMI, etc. They should have enough friction (~0.2mm) but not too much, the tolerances in the connectors and the printer are very complicated to handle. I started out tuning this in the OpenSCAD source but this does not work in IM which has a much smaller tolerance. I recently reverted to exact measures in the source and now use CURA to expand the polygons so they have the required size. Are there other approach for this?
Would be very interested to have a discussion that have a similar problem.
Peter Kriens
I am very new to 3D modeling and printing. For my project I picked OpenSCAD because I love the programmatic approach and I need lots of variations on the same core components. After some grueling time to learn OpenSCAD and the (complex but very good) BOSL2 library I built a number of prototypes and printed them on a Ultimaker S3 and tested them with focus groups.
I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
To my horror I found that STEP is the standard file in the IM world, they genuinely hate STL files it seems. I can convert the STL files decently through Fusion360 but this is almost sufficient to exit OpenSCAD. Conversions are always lossy and it is kind of worrying that OpenSCAD is this not used in the IM world.
3D printing has the issue of support >45 degree overhangs, IM requires that the mold can be easily released freely. This profoundly influences the design. There are also the issues of plastic uses, sink marks, warpage, etc. Curious if anybody has thoughts about this.
As a software developer I'd like to have a single source but the IM people seem to just want to take the STEP file and then they do their magic to it. This not completely insane since the mold making process is very expensive.
In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
My design requires very accurate holes for connectors like USB, HDMI, etc. They should have enough friction (~0.2mm) but not too much, the tolerances in the connectors and the printer are very complicated to handle. I started out tuning this in the OpenSCAD source but this does not work in IM which has a much smaller tolerance. I recently reverted to exact measures in the source and now use CURA to expand the polygons so they have the required size. Are there other approach for this?
Would be very interested to have a discussion that have a similar problem.
Peter Kriens
MP
Marcus Poller
Wed, Sep 4, 2024 10:22 AM
I am very new to 3D modeling and printing.
I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to compensate them.
- FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
- Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
- sintering: you need additional holes to let metal powder run out of your model
- Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
- laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
Are there other approach for this?
Create one model per fabrication process. There is no general solution.
Would be very interested to have a discussion that have a similar problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that instrument.
I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
Hi Peter,
> I am very new to 3D modeling and printing.
> I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
> In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to compensate them.
* FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
* Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
* sintering: you need additional holes to let metal powder run out of your model
* Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
* laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
> Are there other approach for this?
Create one model per fabrication process. There is no general solution.
> Would be very interested to have a discussion that have a similar problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that instrument.
I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
MM
Michael Möller
Wed, Sep 4, 2024 11:18 AM
🙂 to the above.
"In theory, there is no difference between theory and practice. In
practice, however ..."
I have a whole book for the many different FDM processes, and each of their
quirks, and even that is more for the (serious) hobbist, and not industrial.
ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <
discuss@lists.openscad.org>:
I am very new to 3D modeling and printing.
I am now working on productizing this and I decided it will need to be
injection molded (IM). Studying the issues it seems there are quite a few
of them.
In 3D printing chamfering with 45 degree has an advantage over rounding,
which needs support. However, rounding looks a lot better so is required
for a commercial product. Do I need to spread if (im) ... all over my code?
Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the production
process driver (e.g. slicer for FDM-Printing) will make sure your model can
be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to
compensate them.
- FDM-prints schrink after printing. Overhangs >45° require support.
Supports leave support marks
- Milling: Removing wood or metall from a block of material creates
tension, your workpiece will bend. You need to compensate for that, e.g. by
running a prototype slightly greater than your model. Some CNC-machines
only have 2.5 axes, so you need to remodel overhangs.
- sintering: you need additional holes to let metal powder run out of
your model
- Injection Molding: you cannot use cubes, the model needs to be
releasable in a machine. You need additional injection tubes.
- laser cutting: the heat generated by your laser will generate tension
within the material. The cut out pieces will generate tension. Your
workpiece will bend on the laser bed. Good drivers may compensate tension
by chopping your cutline into segments. Sometimes the machine operator
needs to treat the worki piece with heat within the cutting process. The
machine operator needs to know how to put your piece onto a raw supply
material to compensate for in-material-tension.
Are there other approach for this?
Create one model per fabrication process. There is no general solution.
Would be very interested to have a discussion that have a similar
problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears.
The operator rejected my 3d model, it could not be smaller that 2mm in each
dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by
surface elevation/texture. I want to replace the texture with a multi color
FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that
instrument.
I have talked to some carpenters and some mill operators about the
process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard.
There is no guarantee that I will get a flat guitar neck, just educated
guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
🙂 to the above.
"In theory, there is no difference between theory and practice. In
practice, however ..."
I have a whole book for the many different FDM processes, and each of their
quirks, and even that is more for the (serious) hobbist, and not industrial.
ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <
discuss@lists.openscad.org>:
> Hi Peter,
>
> > I am very new to 3D modeling and printing.
>
> > I am now working on productizing this and I decided it will need to be
> injection molded (IM). Studying the issues it seems there are quite a few
> of them.
>
> > In 3D printing chamfering with 45 degree has an advantage over rounding,
> which needs support. However, rounding looks a lot better so is required
> for a commercial product. Do I need to spread if (im) ... all over my code?
> Or just take the cost of support?
>
> You need one model per production process.
>
> In theory it is all fun an games: You have one 3d model and the production
> process driver (e.g. slicer for FDM-Printing) will make sure your model can
> be produced on any machine (printed, cutted, milled).
>
> In practice:
> Each fabrication process has its limits and your 3d model needs to
> compensate them.
>
> * FDM-prints schrink after printing. Overhangs >45° require support.
> Supports leave support marks
> * Milling: Removing wood or metall from a block of material creates
> tension, your workpiece will bend. You need to compensate for that, e.g. by
> running a prototype slightly greater than your model. Some CNC-machines
> only have 2.5 axes, so you need to remodel overhangs.
> * sintering: you need additional holes to let metal powder run out of
> your model
> * Injection Molding: you cannot use cubes, the model needs to be
> releasable in a machine. You need additional injection tubes.
> * laser cutting: the heat generated by your laser will generate tension
> within the material. The cut out pieces will generate tension. Your
> workpiece will bend on the laser bed. Good drivers may compensate tension
> by chopping your cutline into segments. Sometimes the machine operator
> needs to treat the worki piece with heat within the cutting process. The
> machine operator needs to know how to put your piece onto a raw supply
> material to compensate for in-material-tension.
>
> > Are there other approach for this?
>
> Create one model per fabrication process. There is no general solution.
>
>
> > Would be very interested to have a discussion that have a similar
> problem.
>
> Parametric models:
> I wanted to try SLA printing. The model were some credit card size gears.
> The operator rejected my 3d model, it could not be smaller that 2mm in each
> dimension. I forked my own model.
>
> Color models:
> I have a FDM pirate treasure chest that emulates the looks of wood by
> surface elevation/texture. I want to replace the texture with a multi color
> FDM print. I remodeled the box.
>
> Production processes:
> I have been thinking about CNC milling a guitar. I intend to play on that
> instrument.
>
> I have talked to some carpenters and some mill operators about the
> process. Sourcing the wood in such small quantities will be hard.
> Arranging the milling process to compensate for bending is very hard.
> There is no guarantee that I will get a flat guitar neck, just educated
> guesses.
> Milling from two sides will be hard.
> I will need to postprocess by hand to remove tool marks.
>
> You cannot operate a CNC mill without being a carpenter first.
>
> I have given up on that idea.
>
> Marcus
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email to discuss-leave@lists.openscad.org
>
NH
nop head
Wed, Sep 4, 2024 11:59 AM
I use $variables to customise my library for different processes. At the
moment I only print 3D objects with FDM from STL files and CNC mill 2D
objects from DXF files. I do have a large SLA printer that I haven't
tried yet. I plan to add a variable to make 3D objects more suitable for
SLA. E.g. change the polyholes and teardrops to normal round holes. I also
plan to build a laser cutter and for that I will set $cnc_bit_r to the tiny
laser kerf.
Unfortunately $variables have been broken, so I have to use the last
release to produce my output files. I use later snapshots in the GUI to get
the faster rendering.
I know production houses hate STLs but I do all my own manufacturing. If I
was doing injection moulding I would just set $fa and $fs very high and
create a higher res STL file. I would then probably make an EDM machine to
cut the mold and make my own injection machine using a design from Buster
Beagle 3D.
STL only has straight lines but my machines only move in straight lines so
not a problem.
On Wed, 4 Sept 2024 at 12:18, Michael Möller via Discuss <
discuss@lists.openscad.org> wrote:
🙂 to the above.
"In theory, there is no difference between theory and practice. In
practice, however ..."
I have a whole book for the many different FDM processes, and each of
their quirks, and even that is more for the (serious) hobbist, and not
industrial.
ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <
discuss@lists.openscad.org>:
I am very new to 3D modeling and printing.
I am now working on productizing this and I decided it will need to be
injection molded (IM). Studying the issues it seems there are quite a few
of them.
In 3D printing chamfering with 45 degree has an advantage over
rounding, which needs support. However, rounding looks a lot better so is
required for a commercial product. Do I need to spread if (im) ... all over
my code? Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the
production process driver (e.g. slicer for FDM-Printing) will make sure
your model can be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to
compensate them.
- FDM-prints schrink after printing. Overhangs >45° require support.
Supports leave support marks
- Milling: Removing wood or metall from a block of material creates
tension, your workpiece will bend. You need to compensate for that, e.g. by
running a prototype slightly greater than your model. Some CNC-machines
only have 2.5 axes, so you need to remodel overhangs.
- sintering: you need additional holes to let metal powder run out of
your model
- Injection Molding: you cannot use cubes, the model needs to be
releasable in a machine. You need additional injection tubes.
- laser cutting: the heat generated by your laser will generate tension
within the material. The cut out pieces will generate tension. Your
workpiece will bend on the laser bed. Good drivers may compensate tension
by chopping your cutline into segments. Sometimes the machine operator
needs to treat the worki piece with heat within the cutting process. The
machine operator needs to know how to put your piece onto a raw supply
material to compensate for in-material-tension.
Are there other approach for this?
Create one model per fabrication process. There is no general solution.
Would be very interested to have a discussion that have a similar
problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears.
The operator rejected my 3d model, it could not be smaller that 2mm in each
dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by
surface elevation/texture. I want to replace the texture with a multi color
FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that
instrument.
I have talked to some carpenters and some mill operators about the
process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard.
There is no guarantee that I will get a flat guitar neck, just educated
guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
I use $variables to customise my library for different processes. At the
moment I only print 3D objects with FDM from STL files and CNC mill 2D
objects from DXF files. I do have a large SLA printer that I haven't
tried yet. I plan to add a variable to make 3D objects more suitable for
SLA. E.g. change the polyholes and teardrops to normal round holes. I also
plan to build a laser cutter and for that I will set $cnc_bit_r to the tiny
laser kerf.
Unfortunately $variables have been broken, so I have to use the last
release to produce my output files. I use later snapshots in the GUI to get
the faster rendering.
I know production houses hate STLs but I do all my own manufacturing. If I
was doing injection moulding I would just set $fa and $fs very high and
create a higher res STL file. I would then probably make an EDM machine to
cut the mold and make my own injection machine using a design from Buster
Beagle 3D.
STL only has straight lines but my machines only move in straight lines so
not a problem.
On Wed, 4 Sept 2024 at 12:18, Michael Möller via Discuss <
discuss@lists.openscad.org> wrote:
> 🙂 to the above.
>
> "In theory, there is no difference between theory and practice. In
> practice, however ..."
>
> I have a whole book for the many different FDM processes, and each of
> their quirks, and even that is more for the (serious) hobbist, and not
> industrial.
>
> ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <
> discuss@lists.openscad.org>:
>
>> Hi Peter,
>>
>> > I am very new to 3D modeling and printing.
>>
>> > I am now working on productizing this and I decided it will need to be
>> injection molded (IM). Studying the issues it seems there are quite a few
>> of them.
>>
>> > In 3D printing chamfering with 45 degree has an advantage over
>> rounding, which needs support. However, rounding looks a lot better so is
>> required for a commercial product. Do I need to spread if (im) ... all over
>> my code? Or just take the cost of support?
>>
>> You need one model per production process.
>>
>> In theory it is all fun an games: You have one 3d model and the
>> production process driver (e.g. slicer for FDM-Printing) will make sure
>> your model can be produced on any machine (printed, cutted, milled).
>>
>> In practice:
>> Each fabrication process has its limits and your 3d model needs to
>> compensate them.
>>
>> * FDM-prints schrink after printing. Overhangs >45° require support.
>> Supports leave support marks
>> * Milling: Removing wood or metall from a block of material creates
>> tension, your workpiece will bend. You need to compensate for that, e.g. by
>> running a prototype slightly greater than your model. Some CNC-machines
>> only have 2.5 axes, so you need to remodel overhangs.
>> * sintering: you need additional holes to let metal powder run out of
>> your model
>> * Injection Molding: you cannot use cubes, the model needs to be
>> releasable in a machine. You need additional injection tubes.
>> * laser cutting: the heat generated by your laser will generate tension
>> within the material. The cut out pieces will generate tension. Your
>> workpiece will bend on the laser bed. Good drivers may compensate tension
>> by chopping your cutline into segments. Sometimes the machine operator
>> needs to treat the worki piece with heat within the cutting process. The
>> machine operator needs to know how to put your piece onto a raw supply
>> material to compensate for in-material-tension.
>>
>> > Are there other approach for this?
>>
>> Create one model per fabrication process. There is no general solution.
>>
>>
>> > Would be very interested to have a discussion that have a similar
>> problem.
>>
>> Parametric models:
>> I wanted to try SLA printing. The model were some credit card size gears.
>> The operator rejected my 3d model, it could not be smaller that 2mm in each
>> dimension. I forked my own model.
>>
>> Color models:
>> I have a FDM pirate treasure chest that emulates the looks of wood by
>> surface elevation/texture. I want to replace the texture with a multi color
>> FDM print. I remodeled the box.
>>
>> Production processes:
>> I have been thinking about CNC milling a guitar. I intend to play on that
>> instrument.
>>
>> I have talked to some carpenters and some mill operators about the
>> process. Sourcing the wood in such small quantities will be hard.
>> Arranging the milling process to compensate for bending is very hard.
>> There is no guarantee that I will get a flat guitar neck, just educated
>> guesses.
>> Milling from two sides will be hard.
>> I will need to postprocess by hand to remove tool marks.
>>
>> You cannot operate a CNC mill without being a carpenter first.
>>
>> I have given up on that idea.
>>
>> Marcus
>> _______________________________________________
>> OpenSCAD mailing list
>> To unsubscribe send an email to discuss-leave@lists.openscad.org
>>
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email to discuss-leave@lists.openscad.org
>
JB
Jon Bondy
Wed, Sep 4, 2024 12:01 PM
Marcus:
I agree with your observations, in general, but I can successfully
compensate for post-print FDM shrinkage, and support on the BambuLabs
X1C, using their slicer, is almost invisible. Typically, the support
breaks off quickly and easily. I've not had similar luck with
PrusaSlicer supports on the Prusa XL.
Jon
On 9/4/2024 6:22 AM, Marcus Poller via Discuss wrote:
I am very new to 3D modeling and printing.
I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to compensate them.
- FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
- Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
- sintering: you need additional holes to let metal powder run out of your model
- Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
- laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
Are there other approach for this?
Create one model per fabrication process. There is no general solution.
Would be very interested to have a discussion that have a similar problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that instrument.
I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
--
This email has been checked for viruses by AVG antivirus software.
www.avg.com
Marcus:
I agree with your observations, in general, but I can successfully
compensate for post-print FDM shrinkage, and support on the BambuLabs
X1C, using their slicer, is almost invisible. Typically, the support
breaks off quickly and easily. I've not had similar luck with
PrusaSlicer supports on the Prusa XL.
Jon
On 9/4/2024 6:22 AM, Marcus Poller via Discuss wrote:
> Hi Peter,
>
>> I am very new to 3D modeling and printing.
>> I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
>> In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
> You need one model per production process.
>
> In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
>
> In practice:
> Each fabrication process has its limits and your 3d model needs to compensate them.
>
> * FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
> * Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
> * sintering: you need additional holes to let metal powder run out of your model
> * Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
> * laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
>
>> Are there other approach for this?
> Create one model per fabrication process. There is no general solution.
>
>
>> Would be very interested to have a discussion that have a similar problem.
> Parametric models:
> I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
>
> Color models:
> I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
>
> Production processes:
> I have been thinking about CNC milling a guitar. I intend to play on that instrument.
>
> I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
> Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
> Milling from two sides will be hard.
> I will need to postprocess by hand to remove tool marks.
>
> You cannot operate a CNC mill without being a carpenter first.
>
> I have given up on that idea.
>
> Marcus
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email to discuss-leave@lists.openscad.org
--
This email has been checked for viruses by AVG antivirus software.
www.avg.com
RW
Raymond West
Wed, Sep 4, 2024 12:33 PM
Throw away the 3d printed design, and start afresh if you are actually
thinking that you need an injection moulded part. It is entirely
different design requirements. Also remember, all the cost is loaded up
front, and a mistake in the design is expensive to correct. Wrt design
software, It is probably worth paying a few thousand dollars a year for
something like solid works.
On 04/09/2024 08:35, Peter Kriens via Discuss wrote:
I am very new to 3D modeling and printing. For my project I picked
OpenSCAD because I love the programmatic approach and I need lots of
variations on the same core components. After some grueling time to
learn OpenSCAD and the (complex but very good) BOSL2 library I built a
number of prototypes and printed them on a Ultimaker S3 and tested
them with focus groups.
I am now working on productizing this and I decided it will need to be
injection molded (IM). Studying the issues it seems there are quite a
few of them.
- To my horror I found that STEP is the standard file in the IM
world, they genuinely hate STL files it seems. I can convert the
STL files decently through Fusion360 but this is almost sufficient
to exit OpenSCAD. Conversions are always lossy and it is kind of
worrying that OpenSCAD is this not used in the IM world.
- 3D printing has the issue of support >45 degree overhangs, IM
requires that the mold can be easily released freely. This
profoundly influences the design. There are also the issues of
plastic uses, sink marks, warpage, etc. Curious if anybody has
thoughts about this.
- As a software developer I'd like to have a single source but the
IM people seem to just want to take the STEP file and then they do
their magic to it. This not completely insane since the mold
making process is very expensive.
- In 3D printing chamfering with 45 degree has an advantage over
rounding, which needs support. However, rounding looks a lot
better so is required for a commercial product. Do I need to
spread if (im) ... all over my code? Or just take the cost of support?
- My design requires very accurate holes for connectors like USB,
HDMI, etc. They should have enough friction (~0.2mm) but not too
much, the tolerances in the connectors and the printer are very
complicated to handle. I started out tuning this in the OpenSCAD
source but this does not work in IM which has a much smaller
tolerance. I recently reverted to exact measures in the source and
now use CURA to expand the polygons so they have the required
size. Are there other approach for this?
Would be very interested to have a discussion that have a similar
problem.
Peter Kriens
OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org
Throw away the 3d printed design, and start afresh if you are actually
thinking that you need an injection moulded part. It is entirely
different design requirements. Also remember, all the cost is loaded up
front, and a mistake in the design is expensive to correct. Wrt design
software, It is probably worth paying a few thousand dollars a year for
something like solid works.
On 04/09/2024 08:35, Peter Kriens via Discuss wrote:
> I am very new to 3D modeling and printing. For my project I picked
> OpenSCAD because I love the programmatic approach and I need lots of
> variations on the same core components. After some grueling time to
> learn OpenSCAD and the (complex but very good) BOSL2 library I built a
> number of prototypes and printed them on a Ultimaker S3 and tested
> them with focus groups.
>
> I am now working on productizing this and I decided it will need to be
> injection molded (IM). Studying the issues it seems there are quite a
> few of them.
>
> 1. To my horror I found that STEP is the standard file in the IM
> world, they genuinely hate STL files it seems. I can convert the
> STL files decently through Fusion360 but this is almost sufficient
> to exit OpenSCAD. Conversions are always lossy and it is kind of
> worrying that OpenSCAD is this not used in the IM world.
> 2. 3D printing has the issue of support >45 degree overhangs, IM
> requires that the mold can be easily released freely. This
> profoundly influences the design. There are also the issues of
> plastic uses, sink marks, warpage, etc. Curious if anybody has
> thoughts about this.
> 3. As a software developer I'd like to have a single source but the
> IM people seem to just want to take the STEP file and then they do
> their magic to it. This not completely insane since the mold
> making process is very expensive.
> 4. In 3D printing chamfering with 45 degree has an advantage over
> rounding, which needs support. However, rounding looks a lot
> better so is required for a commercial product. Do I need to
> spread if (im) ... all over my code? Or just take the cost of support?
> 5. My design requires very accurate holes for connectors like USB,
> HDMI, etc. They should have enough friction (~0.2mm) but not too
> much, the tolerances in the connectors and the printer are very
> complicated to handle. I started out tuning this in the OpenSCAD
> source but this does not work in IM which has a _much_ smaller
> tolerance. I recently reverted to exact measures in the source and
> now use CURA to expand the polygons so they have the required
> size. Are there other approach for this?
>
>
> Would be very interested to have a discussion that have a similar
> problem.
>
> Peter Kriens
>
>
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email todiscuss-leave@lists.openscad.org
JB
Jordan Brown
Wed, Sep 4, 2024 3:31 PM
On 9/4/2024 12:35 AM, Peter Kriens via Discuss wrote:
My design requires very accurate holes for connectors like USB, HDMI,
etc. They should have enough friction (~0.2mm) but not too much, the
tolerances in the connectors and the printer are very complicated to
handle.
Really? I don't know anything about the design process there, but as a
user it seems like the critical tolerances are all in the metal-on-metal
interfaces supplied by the connectors themselves. I don't think I've
ever seen (or at least have never noticed) a product where the plastic
is part of the mechanical connection.
On 9/4/2024 12:35 AM, Peter Kriens via Discuss wrote:
> My design requires very accurate holes for connectors like USB, HDMI,
> etc. They should have enough friction (~0.2mm) but not too much, the
> tolerances in the connectors and the printer are very complicated to
> handle.
Really? I don't know anything about the design process there, but as a
user it seems like the critical tolerances are all in the metal-on-metal
interfaces supplied by the connectors themselves. I don't think I've
ever seen (or at least have never noticed) a product where the plastic
is part of the mechanical connection.
BC
Bob Carlson
Wed, Sep 4, 2024 10:27 PM
"Unfortunately $variables have been broken, “
Can you elucidate?
-Bob
On Sep 4, 2024, at 04:59, nop head via Discuss discuss@lists.openscad.org wrote:
I use $variables to customise my library for different processes. At the moment I only print 3D objects with FDM from STL files and CNC mill 2D objects from DXF files. I do have a large SLA printer that I haven't tried yet. I plan to add a variable to make 3D objects more suitable for SLA. E.g. change the polyholes and teardrops to normal round holes. I also plan to build a laser cutter and for that I will set $cnc_bit_r to the tiny laser kerf.
Unfortunately $variables have been broken, so I have to use the last release to produce my output files. I use later snapshots in the GUI to get the faster rendering.
I know production houses hate STLs but I do all my own manufacturing. If I was doing injection moulding I would just set $fa and $fs very high and create a higher res STL file. I would then probably make an EDM machine to cut the mold and make my own injection machine using a design from Buster Beagle 3D.
STL only has straight lines but my machines only move in straight lines so not a problem.
On Wed, 4 Sept 2024 at 12:18, Michael Möller via Discuss <discuss@lists.openscad.org mailto:discuss@lists.openscad.org> wrote:
🙂 to the above.
"In theory, there is no difference between theory and practice. In practice, however ..."
I have a whole book for the many different FDM processes, and each of their quirks, and even that is more for the (serious) hobbist, and not industrial.
ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <discuss@lists.openscad.org mailto:discuss@lists.openscad.org>:
I am very new to 3D modeling and printing.
I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
You need one model per production process.
In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
In practice:
Each fabrication process has its limits and your 3d model needs to compensate them.
- FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
- Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
- sintering: you need additional holes to let metal powder run out of your model
- Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
- laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
Are there other approach for this?
Create one model per fabrication process. There is no general solution.
Would be very interested to have a discussion that have a similar problem.
Parametric models:
I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
Color models:
I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
Production processes:
I have been thinking about CNC milling a guitar. I intend to play on that instrument.
I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
Milling from two sides will be hard.
I will need to postprocess by hand to remove tool marks.
You cannot operate a CNC mill without being a carpenter first.
I have given up on that idea.
Marcus
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org mailto:discuss-leave@lists.openscad.org
"Unfortunately $variables have been broken, “
Can you elucidate?
-Bob
> On Sep 4, 2024, at 04:59, nop head via Discuss <discuss@lists.openscad.org> wrote:
>
> I use $variables to customise my library for different processes. At the moment I only print 3D objects with FDM from STL files and CNC mill 2D objects from DXF files. I do have a large SLA printer that I haven't tried yet. I plan to add a variable to make 3D objects more suitable for SLA. E.g. change the polyholes and teardrops to normal round holes. I also plan to build a laser cutter and for that I will set $cnc_bit_r to the tiny laser kerf.
>
> Unfortunately $variables have been broken, so I have to use the last release to produce my output files. I use later snapshots in the GUI to get the faster rendering.
>
> I know production houses hate STLs but I do all my own manufacturing. If I was doing injection moulding I would just set $fa and $fs very high and create a higher res STL file. I would then probably make an EDM machine to cut the mold and make my own injection machine using a design from Buster Beagle 3D.
>
> STL only has straight lines but my machines only move in straight lines so not a problem.
>
> On Wed, 4 Sept 2024 at 12:18, Michael Möller via Discuss <discuss@lists.openscad.org <mailto:discuss@lists.openscad.org>> wrote:
>> 🙂 to the above.
>>
>> "In theory, there is no difference between theory and practice. In practice, however ..."
>>
>> I have a whole book for the many different FDM processes, and each of their quirks, and even that is more for the (serious) hobbist, and not industrial.
>>
>>
>> ons. 4. sep. 2024 12.22 skrev Marcus Poller via Discuss <discuss@lists.openscad.org <mailto:discuss@lists.openscad.org>>:
>>> Hi Peter,
>>>
>>> > I am very new to 3D modeling and printing.
>>>
>>> > I am now working on productizing this and I decided it will need to be injection molded (IM). Studying the issues it seems there are quite a few of them.
>>>
>>> > In 3D printing chamfering with 45 degree has an advantage over rounding, which needs support. However, rounding looks a lot better so is required for a commercial product. Do I need to spread if (im) ... all over my code? Or just take the cost of support?
>>>
>>> You need one model per production process.
>>>
>>> In theory it is all fun an games: You have one 3d model and the production process driver (e.g. slicer for FDM-Printing) will make sure your model can be produced on any machine (printed, cutted, milled).
>>>
>>> In practice:
>>> Each fabrication process has its limits and your 3d model needs to compensate them.
>>>
>>> * FDM-prints schrink after printing. Overhangs >45° require support. Supports leave support marks
>>> * Milling: Removing wood or metall from a block of material creates tension, your workpiece will bend. You need to compensate for that, e.g. by running a prototype slightly greater than your model. Some CNC-machines only have 2.5 axes, so you need to remodel overhangs.
>>> * sintering: you need additional holes to let metal powder run out of your model
>>> * Injection Molding: you cannot use cubes, the model needs to be releasable in a machine. You need additional injection tubes.
>>> * laser cutting: the heat generated by your laser will generate tension within the material. The cut out pieces will generate tension. Your workpiece will bend on the laser bed. Good drivers may compensate tension by chopping your cutline into segments. Sometimes the machine operator needs to treat the worki piece with heat within the cutting process. The machine operator needs to know how to put your piece onto a raw supply material to compensate for in-material-tension.
>>>
>>> > Are there other approach for this?
>>>
>>> Create one model per fabrication process. There is no general solution.
>>>
>>>
>>> > Would be very interested to have a discussion that have a similar problem.
>>>
>>> Parametric models:
>>> I wanted to try SLA printing. The model were some credit card size gears. The operator rejected my 3d model, it could not be smaller that 2mm in each dimension. I forked my own model.
>>>
>>> Color models:
>>> I have a FDM pirate treasure chest that emulates the looks of wood by surface elevation/texture. I want to replace the texture with a multi color FDM print. I remodeled the box.
>>>
>>> Production processes:
>>> I have been thinking about CNC milling a guitar. I intend to play on that instrument.
>>>
>>> I have talked to some carpenters and some mill operators about the process. Sourcing the wood in such small quantities will be hard.
>>> Arranging the milling process to compensate for bending is very hard. There is no guarantee that I will get a flat guitar neck, just educated guesses.
>>> Milling from two sides will be hard.
>>> I will need to postprocess by hand to remove tool marks.
>>>
>>> You cannot operate a CNC mill without being a carpenter first.
>>>
>>> I have given up on that idea.
>>>
>>> Marcus
>>> _______________________________________________
>>> OpenSCAD mailing list
>>> To unsubscribe send an email to discuss-leave@lists.openscad.org <mailto:discuss-leave@lists.openscad.org>
>> _______________________________________________
>> OpenSCAD mailing list
>> To unsubscribe send an email to discuss-leave@lists.openscad.org <mailto:discuss-leave@lists.openscad.org>
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email to discuss-leave@lists.openscad.org
JB
Jordan Brown
Wed, Sep 4, 2024 10:36 PM
On 9/4/2024 3:27 PM, Bob Carlson via Discuss wrote:
"Unfortunately $variables have been broken, “
Can you elucidate?
A couple of years ago there was an incompatible change made to $
variable behavior and how they interact with use<> files. It's
invisible to most people, but it broke a particular pattern that
Nophead's library uses extensively.
Personally, I think we should make a larger and more incompatible
change to that behavior. The current behavior, before and after that
change, imposes a huge performance penalty on use<> files. The
transition would be painful, but there would be a significant
improvement in performance. Or introduce a third library-inclusion
mechanism with the better behavior, leaving use<> as a legacy.
(The differences are around when and whether global variables in use<>
files are evaluated.)
On 9/4/2024 3:27 PM, Bob Carlson via Discuss wrote:
> "Unfortunately $variables have been broken, “
>
> Can you elucidate?
A couple of years ago there was an incompatible change made to $
variable behavior and how they interact with use<> files. It's
invisible to most people, but it broke a particular pattern that
Nophead's library uses extensively.
Personally, I think we should make a larger and *more* incompatible
change to that behavior. The current behavior, before and after that
change, imposes a huge performance penalty on use<> files. The
transition would be painful, but there would be a significant
improvement in performance. Or introduce a third library-inclusion
mechanism with the better behavior, leaving use<> as a legacy.
(The differences are around when and whether global variables in use<>
files are evaluated.)
BC
Bob Carlson
Wed, Sep 4, 2024 11:45 PM
Can you point to an Issue in Github that contains the details?
Seems like a new inclusion mechanism would be the cleanest way forward.
And I am still trying to be patient waiting for the object oriented features to be added that have been in the works for years.
-Bob
"Unfortunately $variables have been broken, “
Can you elucidate?
A couple of years ago there was an incompatible change made to $ variable behavior and how they interact with use<> files. It's invisible to most people, but it broke a particular pattern that Nophead's library uses extensively.
Personally, I think we should make a larger and more incompatible change to that behavior. The current behavior, before and after that change, imposes a huge performance penalty on use<> files. The transition would be painful, but there would be a significant improvement in performance. Or introduce a third library-inclusion mechanism with the better behavior, leaving use<> as a legacy.
(The differences are around when and whether global variables in use<> files are evaluated.)
Can you point to an Issue in Github that contains the details?
Seems like a new inclusion mechanism would be the cleanest way forward.
And I am still trying to be patient waiting for the object oriented features to be added that have been in the works for years.
-Bob
> On Sep 4, 2024, at 15:36, Jordan Brown <openscad@jordan.maileater.net> wrote:
>
> On 9/4/2024 3:27 PM, Bob Carlson via Discuss wrote:
>> "Unfortunately $variables have been broken, “
>>
>> Can you elucidate?
>
> A couple of years ago there was an incompatible change made to $ variable behavior and how they interact with use<> files. It's invisible to most people, but it broke a particular pattern that Nophead's library uses extensively.
>
> Personally, I think we should make a larger and *more* incompatible change to that behavior. The current behavior, before and after that change, imposes a huge performance penalty on use<> files. The transition would be painful, but there would be a significant improvement in performance. Or introduce a third library-inclusion mechanism with the better behavior, leaving use<> as a legacy.
>
> (The differences are around when and whether global variables in use<> files are evaluated.)
>
