3D Metal Printer Projects?

Dear all,

Is there any work being done on metal 3d printing? Like, for example, an open source EOS system [www.eos.info] if anyone knows about systems like this, I would be grateful if you could point me in their direction.

From what I have researched, there are two main methods of metal 3D Printing, metal deposition where you spray the powder through a tube and it gets melted by a laser, and laser sintering where the metal powder is spread on the base and the laser melts the powder directly.

They both have their challenges, but I think laser sintering is more accessible, mainly because you don't have to feed the metal through the moving parts.

The requirements would be 1 - a high power laser, 2 - a method of spreading the powder evenly each layer, 3 - a gas proof chamber for the argon or helium.

Is anyone doing research in this area?

Many thanks.

... here's the related forum: [forums.reprap.org]

I'm on this, but actually more busy with other projects

In the next issue of RepRap-magazine I'll post an article about laserdiodes an their properties/usability, in the following issues maybe some more info and examples using them in different applications ... including SLS ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I'm working on it but on my own tack.

I have decided to investigate hypersonic powder deposition as a building method.

[en.wikipedia.org]

With this method you could deposit plastic and metal all with one system. The problem is creating a collimated stream of particals. That I have been working on and am now building the test rig.

And yes its a little ambitious and nobody has done it for this before but I feel it has promise.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

... have you tested this for plastic?

With powder particle velocities of up to 1km/s you probably only can 'draw' on a solid surface ... for filigrane or thin-walled structures the jet would be more abrasive/cutting than adding!

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I'm with the other guys - its on the backburners, I've only showed that you couldn't melt aluminum to itself because of high thermal conductivity issues.

One way to do it is to put a lot of metal material in a print of another material and 'cook' the non-metalic part off. Expect a lot of shrinkage, and difficulties resulting from that. One method i came across at the RapidPro conference/exposition used that method in combination with stereolithography.(edited: screwed up, kindah used to markdown)

The lost wax methods mention is obligatory, afaik the skills required doesnt go all that far from 'regular' casting skills.

Both those methods require a second method after the printing. Unlike selective laser sintering/melting(SLS/SLM), or this 'hypersonic powder deposition' you are proposing. There was also a thread about electroplating.

Actually, how positive they were about electroplating surprised me a bit i was all like rough calculation Fe+2, charger 2⋅1.6⋅10^-19 C/ per pass(56⋅1.7⋅10^-27 kg)=3.3⋅10^6 C/kg ~ 1kAh/kg not that much since 40kWh/day for US population. (I shouldve compared with that, i wouldnt have been so wrongly negative.)

Edited 1 time(s). Last edit at 03/08/2013 07:42AM by Jasper1984.

I need to get it all built first.

And get a small beam. But I will be testing after that. Probably short tests as I don't think my compressor is up to the supply requirements but enough to prove the concept.

As far as I can see as long as the jet hits a surface at right angles then it will join. Multiple layers can be built up and are by the current tests. Plastics have been tested already and join easily as the velocity required is lower.

If you directly sprayed a high velocity jet on to a soft material it would probably cut it. But I intend to test putting a layer of the high velocity material (Metal) down at low speed first so it joins to the plastic but not to itself. Then the the next layer at a high velocity to join to the previous one which would hopefully stop it cutting the soft material beneath and the particals join to the unbounded metal instead. If that works and its a big if then you could build parts of metal and plastic combined. Otherwise it would have to be all similar materials per object.

And obviously I need to find a suitably support material.

But even if that fails being able to repair delicate aluminium casting by building up damaged or broken arias would be very useful to me so its worth it either way from my perspective.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

Jasper1984

Isn't that how shapeways does it?

That may be a good option. To build the part and stick the metal dust together with a binder then cast it together filling in the gaps with a little extra material than runs in the space left by the binder in the mould.

I have a small furnace capable of 1200c so its an option.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

This is ideally what I think we should be looking at... DMLS

The issues in bringing this to the hobbyist are as stated before.

1 - Affordable high wattage lasers
2 - A method of spreading the powder after each layer effectively
3 - An inert atmosphere for the melted metal to prevent burning (especially for Titanium)

The infrastructure and software is already there (the x,y,z plotting), we just need some heavy duty modifications.

I am sure that it is possible to find inexpensive CO2 lasers.

As for spreading the powder after each layer, perhaps EOS uses an ultrasonic transducer of some method of vibrating the fine particles into an even layer.

The inert atmosphere does pose a problem for the hobbyist though.

How powerful laser is required?

A sealed chamber filled with argon would be best as it would use a very small amount and prints can take a long time. But anyone with a TIG welder has argon. Its not that hard to obtain.

Looking at the price of Co2 laser tubes if it needs lots of power for metal then it going to be really expensive and the life span of these tubes is not spectacular either.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

... with CO2 (10640nm) you'll need more energy than with diodes (808nm or 975nm) or NdYAG's (1064nm) for the different absorption rates at different wavelengths ... CO2 is some decades worse for metals

And it's too depending on the focus size or energy density ...

I've successfully brazed gold-tin microspheres to gold surfaces with 2-4Watts of 975nm with a spot of 0.15mm, but needed more than 20Watts of CO2 for the same material with a spot of 0.2mm diameter!

Made some fusing too with iron dust and 1Watt@808nm with a spot of 0.05mm and 9Watts@975nm and a spot of 0.2mm, what's nearly the same energy density ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I see. I'm not up with lases so forgive my ignorance. So its really a ratio of watts per mm squared for a selected wavelength?

But what about fusing to an object witch had quite a bit of mass to conduct the heat away? Its one thing to weld to a thin plate but to weld to a large lump of metal needs much more power. Up to a point of course.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

I really can not believe that there are no open source 3d metal printing communities! Who is willing to start one? I do not believe that this tech is limited to $500k EOS machines!

Here is a rough sketch of how I imagine a home built printer to look...

EOS state that they use Fiber lasers in their machines.

[www.youtube.com]

Here is a 9W 1100nm Fiber laser on Ebay... clearly where the expense will lie.

[www.ebay.com]

Edited 1 time(s). Last edit at 03/09/2013 09:03AM by Hazel1919.

... there were some early starts some years ago, but didn't continue.

I've made some related development for companies (so not open source) and started some forums and threads for this, but it seems, the 'criticall mass' (or count of people following) isn't sufficient ... could change soon.

Since some years now I'm offering IR-laserdiodes with 5 or 9Watts of power and attached 0.1mm fiber for relative cheap (below 100€) - don't confuse fibercoupled (pigtailed) diodes with fiberlasers!

But in 3 years I've only spread 7 of this 'toys'

For my own development I have a fiberlaser with 50Watts (original price some +30k€) and diodelasers upt o 25Watts and the possibility to combine many single diodes to kW-heads, but busy with other projects, so only slow progress ... but can change soon, as I'm rearranging ...

For melting/sintering powder vs. thermal loss in bas materials - for melting a small spot on a solid metal piece you'll need really high energie densities (typically >50Watts) or 'tweaks' with preheating the part.

Melting dust is another matter - here energies of some Watts are enough.

With plastic it's complete different, as the heat loss in the solid is nearly neglectible - mostly you have to reduce the power to avoid evaporating the plastic in the spot ... but you'll need dark materials and/or absorbing additives or CO2-lasers, as the wavelengths of the IR-diodes will pass through most light/shiny/opaque materials ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

One thing that I miss in topics like this is: What properties do you want to achieve in printed object? Metals greatly vary in physical properties (melting point, heat and electrical conductivity, mechanical strength, etc). So which property do yo want to get / what metal do you want to print? Printing metal just to print metal doesn't make much practical sense to me. Then, especially with metals, material alone isn't all that matters, look, for example at steel working how many different technologies are there and their impact on product properties.

I'm going for aluminium for starters as its good for making lots useful of parts. But as my ideas are a little out there its worth thinking on other lines.

If lasers really are too costly and considering they have been around for donkey's years they probably wont get cheep any time soon. Perhaps some other less perfect approach could be tried? And I'm not the kind of person to shy away from crazy two card short of a deck ideas.

To collimate light ideally you want infinitely small point.Without a laser it is not easy so how small powerful point of light can we make?

Well how about an arc lamp. It has a small but powerful light source in the arc. Now I'm not thinking of the old carbon arc lamps you tend too. In modern arc lamps you put tungsten electrodes in to a glass tube and fill it with the desired gas.Tungsten is cheep. Glass tube is cheep. But apart from argon gasses are a bit more difficult. Hook it up to a tig welder and you have an arc lamp with lots of power. Do not try this without a welding shield on. I'm sure much trial and destruction would be needed to get it all working but there's nothing too extraordinary or impossible here.
[en.wikipedia.org]

So you have a smallish point of light. So where can you get cheep and relitvly small parabolic mirrors to collimate it? How about old camera mirror lenses. If your lucky you can get one thats full of fungus really cheep. The trick to take them apart is to take the screws out the bank and leave the front ring alone. Then you have a nice small parabolic mirror or two. Bit of isopropanol later and its almost good as new. They have a convenient hole in the middle too. Now you have parallelish beams of light but in a very big ring. A couple of lenses later and you have a much smaller but still way to big beam of straight light. Focus it on the surface and you have a small powerful point of destruction or creation.

Or a disaster.

If it worked it would not be as tiny as a laser but that just means slightly less detail. Above a certain curvature the part could still be accurate. And if it all ran in a box filled with argon then there is no need to make a separately filled and sealed tube if argon was OK for the arc.


There are loads of old lenses and stuff you can pick up and take to bits to make optical things. Surly there must be a way to make something effective that wont cost thousands of $$$. As long as the point is small and regular it can make a acurate component. Maybe not a super detail model but just look what we have done with low detail printing in plastic.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

... there are common arc lamps used in beamers with best focus of maybe 5mm and a halo of more than 20mm.

Other types of optical heads with special small high energy halogen bulbs and parabolic mirrors are used for 'light soldering' - here you can receive a spot of maybe 3mm diameter ... but such a head will cost around 3k€-5k€ with a driver for maybe 2k€ too!

With a comercially available 9Watt-lasediode for around 350€, optics for maybe 20€, constant current electronics and coolers for maybe 50€ you'll get much higher energy densities (9Watts on 0.1mm), than with arc lamps!

And for even higher accumulated energies you can combine/add ten of this 9W-heads (for the same privce as a single 'light-soldering-head') on a spot of maybe 0.3mm ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Thought it wouldn't work.

But I'm far more up with gas systems. Not that for a complete system mine will necessarily work out any cheaper or safer. It would however negate the need for inert gas.

Urrrr dread to think what would happen to someone's hand if it got in the way of metal dust traveling at 1000m/s. Nasty.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

Madkite Wrote:
-------------------------------------------------------
>...
> Urrrr dread to think what would happen to
> someone's hand if it got in the way of metal dust
> traveling at 1000m/s. Nasty.

... think about non-toxic metal or ceramic powder 'implanted' inside or beneath the skin - a really nice methode for CNC-tattooing

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

As Miso states, it depends on what you want as the end result, if low resolution is OK, then an idea I posted some years ago may be worth revisiting," ballbearings", a similar size as found in ballpoint pens, fused together by an arc, if you have seen pnuematic silencers the finished product could look similar to them, the power supply is relatively low tech and would not require exotic technology, then maybe some sort of infiltration to smooth and densify, but if you want high resolution then
this is not the answer.

---

Random Precision

Lowery the velocity, then a metal tattoo!!! Someone else can test that.

At the speeds I'm aiming for it might be stopped by the bone. I don't intend to find out the hard way.

Of course the nice thing with it is to slow it down all you have to do is lower the temperature. Though it means the lowest speed would be about 300-350 m/s without refrigerating.

Things like aluminium need 800 m/s+. Copper works well too so I'm hoping to be able to print circuits. You could insulate the bits not to be soldered with a plastic coat too. And the circuit board could be in 3d so conductors can pass blow or above each other instead of taking long detours.

I need to get the rest of the parts yet. They are not expensive but I'm getting married next month so expensive has a rapidly varying meaning at the moment. There is always something.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

There are a few other engineering challenge to add to the list...

So far we have...

1 - Finding affordable High Wattage Lasers
2 - A method of spreading the powder evenly per layer
3 - An inert atmosphere.

In addition we have...

4 - what to do with the unsintered powder after each run
5 - Isolating all vibrations that the motors and other things will cause.

So I will outline some ideas and questions below...

--- 1 ---

This is where a large percentage of the cost will lie. We need to define what type of of laser will be most effective, and affordable.
EOS, the global DMLS leader uses 200w ?Fiber? lasers. What are the reasons for this? Why such a high wattage for such thin layers?

Here is a very interesting article discussing or, rather, comparing Fiber Lasers (solid state) to CO2 lasers. SSL vs CO2

The article can be summed up as follows, and echoes VDX's comments.

• If cutting materials mainly under 5mm in thickness, then the Fibre laser is more suitable.
• If you need flexibility and quality throughout all types of materials, then a CO2 laser is best.

In light of the fact that we are sintering thicknesses in the micrometer range then a shorter wavelength and therefore SSL laser makes more sense.

--- 2 ---

This is an interesting one. Look at 0:40 to 0:50 of this video EOS powder
As you can see, the metal powder is being dispensed evenly at a certain thickness. You can even see the powder in the center of the arm.

How does this 'arm' work? The question is, does the powder just fall via gravity? Is it shaken? Sprayed?
What exactly is going on inside the arm and how do we replicate that action? Fineness of the layer is crucial I can imagine.

I was thinking of using an ultrasonic transducer in a similar way as this video shows Ultrasonic

--- 3 ---

A gas tight enclosure would have to cover everything, and therefore be quite large. What are the best materials for this purpose? A gas inlet and gas proof door would be needed.

--- 4 ---

Here, I was thinking of vents to the side of the work surface so that when the part is finished, and the work surface rises the excess powder falls into the vents and into containers to be reused. Really, we need to look at the physical properties of metal powder, and how best to manipulate it.

This applies also to the powder dispensing arm... how do we move the powder to the arm and onto the surface?

--- 5 ---

Vibrations from motors, or bumping the machine, I can imagine, will not be helpful... so looking down the road a form of vibration isolation will be necessary.
This, however, can only be properly addressed closer to a final design.

------

I think a key point will be the KISS method. Cut out any unnecessaries, follow the path of least resistance.

I don't believe that DMLS is limited to the high end, and that metal will be the true star of the 3D printing revolution.

Is anyone interested in starting a group to pursue this further?

Many thanks,

Jethro.

Take a look at this:

[www.finishing.com]

[www.halfbakery.com]

Just ordered last parts for test rig. Will take a while as some coming from China but when I have them all it shouldn't bee too long.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

Hi guys, Please excuse the length of posts which can be tiresome to read, but there is so much to cover...

If a self replicating 3D printer aka reprap is a serious goal, metal printing is where it's at, then practically all frame hardware can be printed.

I am fixated on the EOS design, a powder bed, laser cnc, powder dispenser and z axis. To me that is simplicity, the design is in right angles, no complex mechanics. That, I think, is the reason EOS and ARCAM use this method and I am sure they have researched the subject. The fact that two major companies use such a similar process should say something.

The issue I have seen with allot of technology as of late, is that the machine, the tech, when coming from big companies, is grossly over priced.

The closest example I can think of it that of Radio Control autopilots... while commercial autopilots with limited functionality sell for $2k< we have open source autopilots that vastly outstrip that of the commercial ones and cost approx $100.

@ Madkite, that is good news. Where can I find more information on your ideas? It looks very innovative and interesting because you are pursuing it.

@ Jasper, that method I think, personally, is not ideal simply because of the post processing. We ideally want something that just prints the finished product, done.

@ Miso, I want to print in titanium TI-6AI-4V powder with 98< density. EOS claim to achieve strengths at the very least equal to forged parts using the DMLS method.

Fiber lasers are ideal, but expensive so I will window shop around at different fiber laser vendors and report back. What of electron beams? Arcam use electron beams and there is an informative video on the reason behind that which I will try to find.... something about electrons being more efficient than photons for metal sintering. (I also noticed that Arcam use a vacuum instead of argon atmosphere thus eliminating one perishable).

I have a question regarding the inherent limitations of 3D printers and I hope it is clear. Normally, you can not print unsupported structures without supports, but with a powder bed, you do not need a support and can print virtually anything. The question is, how true is that.

Can a part be truly supported in a powder bed? Imagine printing a hollow sphere inside another hollow sphere, how would the powder handle the weight of the un supported inner sphere? I have attached a rough illustration below...

This brings a second issue, that of vibration dampening. If I were to bump the printer half way through printing a sphere in a sphere then that would offset the unsupported inner sphere thus jiggering the whole print. I expect vibrations from the CNC would also offset unsupported parts that lie in the powder bed.

I have drawn an illustration of a machine that I think incorporates all the features we need. I will take a picture and post it soon. I am sure we can come up with a refined, compact and simple machine about the size of a small fridge.

Further, I am making a 3D model on Google sketchup which will be finished soon.

Many thanks,

... with powder-bed printing you don't need support, but you can construct thin beams, holding the inner parts in shape, if you expect displacement ... this beams could be crushed and removed together with the filling powder after the part is ready.

Some problems with powder are warping forces when the melting part changes density while melting and cooling down.

Another issuer is the 'sandy' surface finish, as the melting zone will collect partially not melted particles on the outside.

With a DLM approach, where you blow powder in a hot spot, so it melts without surrounding powder, you'll get perfect smooth surfaces, but you'll need a 5-axis-printer for generating complex shapes and overhangs.

The vacuum with electron beam is essential, as electrons won't pass an atmosphere ... but a vacuum-chamber is more complex and costly than an airtight chamber for inert gas ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Hi Viktor,

You bring up another good point, enclosed spaces need to have at least one hole for the powder to drain. Honeycomb structures would need post processing to remove all the powder.

So issues with the actual DMLS process are as follows.

1 - Warping due to shrinkage
2 - A powdery finish
3 - Removing powder from closed spaces

5 axes printers are, correct me if I am wrong, to complex and therefore not the way to go, I don't want to personally discount anything though.

I would like to somehow document our efforts in a more organized way, this might encourage more people to take part.

Perhaps a wiki would be more organized. I was thinking of something like the wiki structure below...

->DMLS
--------->Overview
--------->Parts List
--------->Frame Design
--------->Axes
----------------->X
----------------->Y
----------------->Z
--------->Laser Selection
--------->Powder Assembly
--------->Electronics
------------------------->CNC, motors
------------------------->Software
--------->Powder Types
--------->Post Processing
--------->CAD Techniques


We need something that brings together and categorically documents every ones efforts, success and failures in a constructive way.

Thank you,

Jethro Hazelhurst.

Edited 1 time(s). Last edit at 03/19/2013 12:54PM by Hazel1919.

Hazel 1919

Unfortunately there is little information and that is hard to find about the normal coting process alone. I'm in a whole new realm of unknown. Where I quite like it. Am making a video with the parts I have as am stuck till the bit come.

But try this for an insight and see attached.

[www.arl.army.mil]

There are lots of factors but its promising from my perspective.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

This is a rough 3D model that I made and am uploading to google sketchup so that you can have a look around. It contains most of the basic parts.

Here is a picture with all the parts labled. We want a highly modular design...