3-D Printing Progress: Unobtanium might be more Obtainable soon!

[Reynard]

There have been a number of threads concerning the emerging technology of 3-D printing, especially as it pertains to antique car parts. This article reviews a new company attempting to develop 3-D printing into an easier, faster and cheaper process. It's written so that the process, present and future, is made more understandable.

 

A few forum members have already experimented with the current technology and find it has promise, but is limited by high costs, time requirements and the need for a high level of knowledge needed to achieve acceptable results.

 

Hopefully there is a steady stream of breakthroughs that lead to the results that are being forecast. I feel this has great implications for the old car hobby.

 

"A startup founded by SpaceX veterans aims to realize the potential of a technology whose big promises have never quite come through."

 

https://www.bloomberg.com/news/articles/2023-02-01/3d-printer-by-ex-spacex-engineer-aims-to-refresh-technology?utm_campaign=bw&utm_medium=distro&utm_source=MSN

 

[Gary_Ash]

The 3D printing is the easy part.  The bottleneck is taking an old 2D drawing or a 3D scan of an old part and converting it to a useful, accurate part that can be 3D printed.  As long as you are willing to pay $100/hour or more for the design service, the 3D printing file can be created.  But, it's still a long way from a final part, cast, machined to size, and plated.  You want a thousand parts?  No problem with costs.  One part?  Expensive.

[joe_padavano]

3D printing of metallic parts is making rapid inroads in the aerospace industry. Costs are still high, but they are coming down. Software packages are automating more and more of the programming. What's really amazing are the new structural optimization packages that minimize weight of the part while maintaining strength and stiffness requirements. The results often have a biological look to them.

 

To be honest, I'm more interested in the 3D sheet metal forming technologies that were the rage about ten years ago, but seem to have vanished now. Ford (for one) was playing with a machine that was essentially a huge dot-matrix printer with steel rods instead of ink jets. The rods were controlled to form a piece of sheet metal (as in a fender or door panel). Imagine the possibilities for patch panels.

 

[Terry Harper]

9 hours ago, Gary_Ash said:

The 3D printing is the easy part.  The bottleneck is taking an old 2D drawing or a 3D scan of an old part and converting it to a useful, accurate part that can be 3D printed.  As long as you are willing to pay $100/hour or more for the design service, the 3D printing file can be created.  But, it's still a long way from a final part, cast, machined to size, and plated.  You want a thousand parts?  No problem with costs.  One part?  Expensive.

So true! Gary hit the nail right on the head!
 

At the moment I have about $1,500.00 of my time and talent invested in scanning, reverse engineering and development of the shop drawing for this part. (See image below) To have it 3D printed in aluminum $1,999.00. ( not going to happen!) A 3D print in PLA to serve as a pattern for a lost PLA casting of this particular part is $30.00 if farmed out to a commercial 3D printing service. Regardless, 3D printed metal or a casting, it will still need machining. 
 

Picture a 3D direct to metal print as being similar to a casting or forging - it still requires machining to be a finished part and to meet design intent. (Fits & tolerances) I have not estimated my time for processing for CNC, setup nor machine time but another $400.00 to $500.00 should cover it. In this case it’s pro bono work for an organization that is dear to my heart. 

 

Add that all up and if you have to pay someone to do the work one little part (in this case less than 5 inches in diameter) can be rather costly. As Gary stated…. The printing is the easy part.

 

while all the above may sound rather negative it’s not. It’s just reality. The technology is amazing and I love working with it and all it offers but there seems to be a common misconception of the 3D scanning and 3D print technology providing near instant and cheap useable parts at the push of a button or two. Whether it current 3D printing technology or a massive production printing technology as described in that particular article, there is still and always will be considerable and costly work that needs to be accomplished before the “print” button is pushed.

Edited February 4, 2023 by Terry Harper (see edit history)

[edinmass]

Been there, done that…….with help from people here commenting. Most parts are not inherently easy or well suited to three D printing. It takes talent and time, both expensive to purchase. Unlimited funds make things easier, but very few projects don’t have a budget.

[Studemax]

One of my nephews has already printed plastic parts for his dad's Corvette using engineering files found on the internet. I suggest you join a group or forum for 3-D printer owners and ask around.

[prewarnut]

9 hours ago, Gary_Ash said:

The 3D printing is the easy part.  The bottleneck is taking an old 2D drawing or a 3D scan of an old part and converting it to a useful, accurate part that can be 3D printed.  As long as you are willing to pay $100/hour or more for the design service, the 3D printing file can be created.  But, it's still a long way from a final part, cast, machined to size, and plated.  You want a thousand parts?  No problem with costs.  One part?  Expensive.

I'll second that. Actually I DIY'd the design work to replicate a 18th century part to get laser cut in 2-d. It was over 100 curves and cuts. It took me 15 hours and learning 3 different CAD programs to produce a DXF file I liked. Once done it was probably a 10 minute machining job. I am publishing an article on this (submitted) to the journal of my other hobby. FYI the technology has now progressed as some laser cutting machines are 3-d and can cut bas-relief just from a photo file if I understand correctly. The software will interpret shadowing to calculate depth. Now, lost crank esucteheons, data plates, vent panels, missing double eagle gold coins (ohhh wrong category) can be replicated - even without holding the part as long as overall dimensions are known or decided upon. In a couple years we'll all be replicating '34 Dietrich Packards from Pebble photos on the winners block - maybe not!

Edited February 4, 2023 by prewarnut (see edit history)

[Dandy Dave]

Sounds like in some cases it is still easier and cheaper to make a wooden pattern and have the part cast from that.

[Terry Harper]

23 minutes ago, Dandy Dave said:

Sounds like in some cases it is still easier and cheaper to make a wooden pattern and have the part cast from that.

Absolutely! It’s all about choosing the correct tool for the job at hand - not trying to set finish nails with a spike maul or drive rail spikes with a tack hammer.

Edited February 4, 2023 by Terry Harper (see edit history)

[Billy Kingsley]

This technology is still so new and so few people are working with it, but imagine in 10 years time when the technology progresses...it's going to be a lot more advanced and prevalent. 

 

It is currently beyond my knowledge on how to do it but I could see myself learning it in the future. 

[edinmass]

It’s not new, I first looked at it ten years ago.

[bryankazmer]

I saw a powder bed printer in the late 1980's.  It was emerging technology at the time.

[Terry Harper]

37 minutes ago, Billy Kingsley said:

This technology is still so new and so few people are working with it, but imagine in 10 years time when the technology progresses...it's going to be a lot more advanced and prevalent. 

 

It is currently beyond my knowledge on how to do it but I could see myself learning it in the future. 

Billy,

I teach this stuff to high school students as part of an engineering technology course. Believe me you can learn it on a hobby DIY level. if you have an interest I would suggest for entry level 3D modeling software looking into AutoDesk Fusion 360 which is available free for hobbyists. Even as a subscription with all the features activated it’s fairly affordable. It’s intuitive and there is a ton of support on YouTube etc. We use SolidWorks which is a different animal. However, I use the manufacturing tools in Fusion for CNC programming.

 

entry level 3D printers are relatively cheap these days. Our first printer was close to $10,000.00 our most recent printers were less than $600.00 and have far, far more capabilities and way more reliability.

 

3D scanning… well you get what you pay for in regards to accuracy and utility. In regards to accuracy - the more zeros you add to the right of the decimal point the more dollars you add to the left and I can tell you it’s not a 1:1 ratio either! However, unless your deep into reverse engineering or doing Quality control and assurance you don’t need to consider it at this point.

 

 

 

Edited February 4, 2023 by Terry Harper (see edit history)

[californiamilleghia]

1 hour ago, edinmass said:

It’s not new, I first looked at it ten years ago.

i saw it 30 years ago , then it was a vat of resin and  lasers , the lasers hardened the resin , then it moved the base one layer and hardened that , on and on ,

 

that has always been the problem as each layer is a step and you can feel  that on the outer edge , 

[jp1gt]

The beauty to 3d printing is that once the scan is done it can be shared (for a cost). Supposedly they have plastic that is easier to print that is as strong as steel.

[SC38dls]

Terry, I first tried to read your post without my glasses on and thought it said a million five hundred thousand!  I thought is he nuts!! I had to put the glasses on just to be sure and realized my error. You do know how to get a guys heart going. 
dave s 

Edited February 4, 2023 by SC38dls (see edit history)

[Terry Harper]

52 minutes ago, jp1gt said:

The beauty to 3d printing is that once the scan is done it can be shared (for a cost). Supposedly they have plastic that is easier to print that is as strong as steel.

Its all about the correct material for the application. Material properties is a whole world in itself! 

[avantey]

When I was working in industrial design for Eastman Kodak we were fortunate to have good management and better budgets that allowed us to get many new state of the art technologies for prototyping.  In the          mid -90's we got sterolithography and very early 3d wax printers.  From there we tried corn starch, sintered metal, 3d specific plastic "inks" and eventually all kinds of new plastic resins that started simulating production materials like abs, polycarb, etc.  Today there are catalogs of materials for all types of applications.  Never tried much with metal printing as I left there in 2007 as that was just starting as a big thing. 

 

As for making parts there was no limit on what we could model, particularly in Solidworks.  SW also had the best file translation for generating .stl or other files for the printers or cad to cad translation.

 

When reverse engineering the scanning and sewing softwares were not very advanced so you developed your own techniques to merge the different direction scans and get some kind of model.  Never needed to do a lot of it as ID is the start point of product design with a blank page, not remaking a competitor's product.  That's still a very time intensive task I think.

 

Everything has evolved light years since then but I sometimes wish I had kept my cad skills up and stayed involved. From EK I did a lot of contract work in more mundane production driven places.  My 10+ years in ID were a blasted and learned so much about a product life cycle and new product R&D.

[bryankazmer]

46 minutes ago, Terry Harper said:

It's all about the correct material for the application. Material properties is a whole world in itself! 

And in particular the properties of 3D printed specimens are usually considerably lower than the data sheet properties tested from conventionally prepared samples

[joe_padavano]

5 minutes ago, bryankazmer said:

And in particular the properties of 3D printed specimens are usually considerably lower than the data sheet properties tested from conventionally prepared samples

That's not necessarily true anymore. In aerospace there is a lot of work being done to characterize the materials properties for printed metallic parts and more importantly to understand the repeatability and number of internal inclusions and voids. As with any sintered part, fusing the material into a homogeneous part takes some processing. Process control is essential. Also, keep in mind that most printed parts are replacing castings, were mechanical properties weren't that great to start with. Having said that, we are printing high strength/high temperature parts like turbopump impellers, turbine wheels, and rocket engine combustion chambers. You can rest assured that materials properties and production consistency are the utmost requirements.

[Gary_Ash]

One of the many 3D printing services is i.materialise.com for lower cost parts that don't need the utmost in mechanical properties, though they are quite good and many materials available.  I've used them for a number of projects.  However, their industrial division at materialise.com offers top-of-the-line mechanical properties in stainless steels, titanium, aluminum, inconel, and other metals, as well as ceramics and plastics.  Sizes for metal parts up to 500 x 280 x 315 mm (19.7" x 11" x 12.4").  If you elect "performance grade" processing, the properties are very close to machining from a billet.  Of course, you will pay a bit more for this.  See, for instance, the specs on 3D printed type 316L stainless steel:

https://www.materialise.com/en/industrial/3d-printing-materials/stainless-steel-ss316l

 

For a lot of things, the hobby/commercial grade parts from i.materialise.com are more than good enough, just don't come with certificates of compliance.

[bryankazmer]

In plastics, the last I saw was the best processes giving about 80% of the physical properties of injection or extrusion, with the less expensive processes considerably less.  But that may be enough for a given application.  My point was to be sure to refer to values generated from the process you intend to use, not from more general data sheets.

[californiamilleghia]

5 hours ago, Gary_Ash said:

One of the many 3D printing services is i.materialise.com for lower cost parts that don't need the utmost in mechanical properties, though they are quite good and many materials available.  I've used them for a number of projects.  However, their industrial division at materialise.com offers top-of-the-line mechanical properties in stainless steels, titanium, aluminum, inconel, and other metals, as well as ceramics and plastics.  Sizes for metal parts up to 500 x 280 x 315 mm (19.7" x 11" x 12.4").  If you elect "performance grade" processing, the properties are very close to machining from a billet.  Of course, you will pay a bit more for this.  See, for instance, the specs on 3D printed type 316L stainless steel:

https://www.materialise.com/en/industrial/3d-printing-materials/stainless-steel-ss316l

 

For a lot of things, the hobby/commercial grade parts from i.materialise.com are more than good enough, just don't come with certificates of compliance.

Have you had anything done by them ?

and maybe a photo , what type computer file you gave them and approx price

 

Thanks 

[Gary_Ash]

Here are some parts that were printed direct-to-metal in bronze-infused steel by i.materialise.com.  The arm with the C-shaped opening was about $100 each, the smaller thin-wall domes with many small holes were about $27 each.  Both parts were tumble polished.  The holes in the arms were printed undersize and drilled to finish dimension later, but this is how they looked when I received them.  I sent them an STL file to print.  I recently had them print a 14" x 5" x 3/4" part in polyamide/nylon which was used as a sand-casting pattern at an aluminum foundry, price was $135.

 

The current price for an Ender 3 printer from Creality is about $189.  It will print 220 x 220 x 250 mm in very good quality, easy to use, a good machine to learn on.  My personal rule-of-thumb is to use an outside service until you have spent 10-20% of the price of a machine, then buy it and be money ahead in the long run, plus the time saved waiting for part delivery.

 

[alsancle]

7 minutes ago, Gary_Ash said:

Here are some parts that were printed direct-to-metal in bronze-infused steel by i.materialise.com.  The arm with the C-shaped opening was about $100 each, the smaller thin-wall domes with many small holes were about $27 each.  Both parts were tumble polished.  The holes in the arms were printed undersize and drilled to finish dimension later, but this is how they looked when I received them.  I sent them an STL file to print.  I recently had them print a 14" x 5" x 3/4" part in polyamide/nylon which was used as a sand-casting pattern at an aluminum foundry, price was $135.

 

The current price for an Ender 3 printer from Creality is about $189.  It will print 220 x 220 x 250 mm in very good quality, easy to use, a good machine to learn on.  My personal rule-of-thumb is to use an outside service until you have spent 10-20% of the price of a machine, then buy it and be money ahead in the long run, plus the time saved waiting for part delivery.

 

Very cool!

[edinmass]

28 minutes ago, alsancle said:

Very cool!

Yes, very cool.Especially because Gary made them for my car……….so Gary receives the very seldom given three thumbs up…….👍👍👍

[californiamilleghia]

2 hours ago, Gary_Ash said:

Here are some parts that were printed direct-to-metal in bronze-infused steel by i.materialise.com. 

 

 

Thanks for the info and photo , 

 

Did you have to add shrinkage to the STL  file ?

[Gary_Ash]

No, make the STL file in the final size and i.materialise.com will adjust it for shrinkage.  For the bronze-infused steel, the Z shrinkage during sintering is different than the X-Y shrinkage (about 20%), but they take care of it.

[jp1gt]

How did you get the STL file? I am waiting for the day that scanning a part will be down to earth.

[Billy Kingsley]

It's new for us plebians! 

 

I'm on my third smartphone and I consider them new, too. 

[Gary_Ash]

OK, boys and girls, here is a lesson in Computer Aided Design 101.  The purpose here is to indicate the difference between creating a 3D object in a CAD program and generating an STL file for 3D printing.  There are many CAD programs in the world, most of which create and save files in some proprietary format.  Better CAD programs can open files in formats created by other CAD programs and sometimes save a file in an alternate format, but generally only a limited number of alternates.  When 3D printing got started, it was necessary to have some file format that the 3D printers could use even if the design came from any CAD program, hence the birth of STL files about 1987.  STL usually means "STereo Lithography" but it can also mean "Standard Triangle Language".

 

When creating in CAD, objects can be drawn as "primitives", like a cube, a cone, a torus or donut, a sphere, and many other shapes like a threaded hex head bolt.  In the simplest view in CAD, a wire form of the shape is seen.  Solids can also be displayed as shaded or colored objects and smooth surfaces are seen.  When the CAD drawing is "saved as" an STL file for export, the CAD program redefines the objects using a large number of small triangles to approximate the surfaces.  Generally, the triangles are small enough that the resulting edges where the triangles meet are not seen in the 3D printed object.  The CAD program does this automatically, just hit the "save as STL" button.

 

Most of the current 3D printers want a little further processing of an STL file.  Using a "slicer" program with some info about the setup of the 3D printer such as the nozzle size, desired layer thickness, etc., the slicer will generate a .GCODE file which tells the nozzle where to go to make each layer on a Fused Deposition Modeling (FDM) printer.  This is the kind of printer that uses extruded plastic filament to create a 3D print one thin layer (e.g. 0.12 to 0.4 mm) at a time.  But, like the universality of STL, GCODE language can also be used to operate other types of 3D printers, laser engravers, and milling machines.  There are many slicer programs out there, most of them free to use.  I use Ultimaker Cura to convert STL files to GCODE for my Ender 3 printer.

 

Here is an example file showing a cube, sphere, cone, and torus.  

 

Class dismissed!

 

 

The four objects shown above as wire forms.  The computer draws these very quickly on the screen. 

 

The four objects shown as shaded solids.  Generating this view takes a little longer.  A photo-realistic view of complex objects can take minutes.

The CAD file converted to an STL file by approximating the surfaces with many small triangles.

 

The solid cone as seen in a preview window in the Ultimaker Cura slicing program about half way through simulated 3D printing.

 

Edited February 8, 2023 by Gary_Ash (see edit history)

[avantey]

Thanks Gary, making me want to jump back in!   On top of the useful results 3d modeling is just plain fun!

[Gary_Ash]

Example #2:

I used the Structure Sensor 3D scanner attached to my iPad to scan my daughter's two dogs, Buster and Poppy.  They cooperated by sleeping on the floor while I scanned.  From the scans, I was able to generate STL files for each of the dogs, made a GCODE file, and printed out 3D models of the dogs.  Here's Buster.

 

Buster as an STL file with lots of triangles of various shapes.  The triangles were generated automatically by the scanning software.

 

The 3D prints of the dogs.

[Terry Harper]

Excellent explanations Gary!

[BearsFan315]

technology is emerging everyday !!

last 20+ years i had been doing sketches, drawings, models. from paper, to CAD to 3 design. we strictly do all our new products in 3 models. and straight to the foundry from there. we design the model based on the process to be used.

 

we also have been going back and modeling older products from hand drawings. it can be a trial and error process. also have scanned in existing parts for modeling. then tweak, add material, modify features, etc... 

 

i did design work for years and spent last few years managing and validating model and developing the process and techniques we use to develop these models. there are a millions ways to get there, and it all comes down to your intent  

 

i love staying on top of the technology and learning and figuring out how to implement into our day to day business. 

[jp1gt]

If someone has made a part and needs it hydrochromed let me know. I have finally mastered the process. I have some parts that are broken and need to be duplicated also. I would like to share the process to others so we can get things done at a decent price.

Edited February 6, 2023 by jp1gt (see edit history)

[drhach]

The ting that compliments 3D printning is 3D scanning. As Gary said, that's what's making the difference. 

 

I've seen a few of this guy's videos.

 

 

Once this becomes accessible, I think think the game will change for those who want to adopt the technology.

 

My wife had her knee replaced not too long ago. On the surgery table, they opened her knee, scanned it and 3D printed a replacement. 

 

 

 

 

[Dandy Dave]

So if I am reading this info correctly, once the part is programmed into the system the next step in printing would involve a wire feed welder of sorts that would feed the proper material to produce a part. Sound like we are already there, or not quite?

[Gary_Ash]

2 hours ago, Dandy Dave said:

So if I am reading this info correctly, once the part is programmed into the system the next step in printing would involve a wire feed welder of sorts that would feed the proper material to produce a part. Sound like we are already there, or not quite?

Well, Dave, maybe not quite a wire feed welder, but metal powder and a laser can do the job.  See this video:

https://www.youtube.com/watch?v=yiUUZxp7bLQ