The Metalbot Project: an open source initiative...

[metalbot]

Hi everyone!

I would like to discuss the pros/cons and engineering challenges one would face in order to create a 3D printer capable of printing in metal.

I have created an open source initiative (linked to below) that deals with this subject, there is quite a bit of information I have amassed on the subject, so it is worth checking out...

http://www.metalbot.org


I am sure most people here have heard of 3D Printers/printing, perhaps some of you've had the opportunity to use them.

From what I have learned, the two limiting factors with 3D printing have always been a - strength of the printed parts and b - part resolution. Hopefully it is possible to solve both of these problems by designing printer that works through a process called 'laser sintering' (although a few other avenues are being pursued).

I would be more than happy to answer any questions to the best of my knowledge!

Best regards,

Jethro.

[MacAttak]

The RepRap forums or other printer-focused groups would probably be a better target audience for you.

Also, I'm betting Stratasys still has all (or most) of the key aspects of the sintering process locked away under patent protections.

[mikeselectricstuff]

Also, I'm betting Stratasys still has all (or most) of the key aspects of the sintering process locked away under patent protections.

That doesn't stop someone publishing an OS design, just selling it

[MacAttak]

That's a pretty good point. But still, I think the RepRap forum and maybe Soliforum and other places like that would probably be a better target to focus efforts on.

[Corporate666]

The original post is spam - the same message has been posted exactly the same - word for word on lots of forums I frequent.  The guy is not interested in having a discussion, but rather just driving people to his website.

[amyk]

I think Dave's said that he doesn't mind people advertising their sites on here.

That being said, this guy's site isn't exactly Google-friendly either; visiting it without JavaScript capability (like Google's indexer would do) results in a completely blank page with no content, and even with JS enabled in a slightly older browser I see absolutely nothing too.

You've basically cut off everyone except those with the very latest software from being able to use your site, no wonder you have little visitors...

[ResR]

It's a spambot program making this topic most likely. I would be suprised if metalbot answers to your questions at all. Most likely the site is infected with some malware that accidentally prevents it loading correctly to your browser. Better do a antivirus scan (as thorough settings as possible) on your computer just to be safe. Edit1: Google search "metalbot forum" gived at least 4 matches with word-to-word post.

[EEVblog]

It's a spambot program making this topic most likely. I would be suprised if metalbot answers to your questions at all. Most likely the site is infected with some malware that accidentally prevents it loading correctly to your browser.

It's way too on-topic to be automated bot spam (which is very hard to get through onto here). Looks genuine enough.
It is not uncommon for people to post the exact same thing in multiple forums.
Whether he comes back at all?  It does seem more advertising than a genuine desire to discuss on here, but certainly benefit of the doubt.

[GeoffS]

Site works fine for me - although there's not a lot of content.
Perhaps he's just trying round up some electronics engineers to help with that side of the project so he's come to the place the best ones hang out 

[metalbot]

@ - MacAttack, interesting that you should mention reprap as this is where the idea originally grew out of. For those interested, here is the thread that started it all. My name was Hazel1919 in the thread. 
 
http://forums.reprap.org/read.php?1,190528,page=1
 
Many companies use a process called metal infusion, where the metal powders are bonded using a binder and then put in a furnace and infused with another metal etc... This produces weaker parts and is inferior to laser sintering which is the process we want to use. It is interesting to note that some case studies have show sintered titanium parts are stronger than even forged equivalents.
 
The two main companies that use the sintering process are EOS and Arcam and these are way out of reach for the hobbyist.
 
@ Corporate666, I have posted the original post on a few other forums because I am willing to discuss this subject on a few other forums. It surprises me how very few people know about 3D metal printing, this is the reason I have started the community - to bring all the knowledge into one organized space.
 
@amyk - This is disturbing to hear. I have heard this from a few different people and am racking my brains for a solution. Unfortunately I am very new to website building and although this is not why I came here, if you have any tips or suggestions, please send them my way. One option I have been seriously considering is scrapping the information site and building another WordPress site from scratch.
 
The forum and wiki (where most of the good info is kept) should work well on any browser. It is just the information site that is hassling me. I will link to them below because you can only access them via the info site...
 
Wiki - http://metalbot.org/wiki/index.php?title=Main_Page
Forum - http://www.metalbot.org/forum/index.php 
 
There is not a ton of content at the moment but that will change over the next few weeks. There are a few people forthcoming that will be documenting their experiments at Metalbot so that will be very interesting.
 
@ ResR - There is no danger to your computers, don't worry, the website is NOT infected with viruses. It is a problem with my HTML5 site. As mentioned before the wiki and forum should work.
 
@ Dave (I assume) - Thank you for the benefit of the doubt. I would like to thank you for your wonderful video you made regarding open source hardware:
 
It gave me some good practical pointers. As always, very well put across... thanks!
 
@ GeoffS

Perhaps he's just trying round up some electronics engineers to help with that side of the project so he's come to the place the best ones hang out 

 
Well put! The reason I have come to EEV and other electronics forums (all about circuits) is that there is a lot of electronics that needs to be bashed out.
 
The electronics associated with Metalbot can be split into two categories. 
 
1 - Powder System Electronics
 
So this is the electronics that will drive the valves, powder spreaders and Z-Axes. It is basically timing electronics (perhaps arduino would be suitable for this purpose?). There will be a lot of mechanical interfacing (for example with ball screws). The whole thing requires precision (as the powder layers will be up to 30 microns in thickness).
 
2 - The Laser System
 
So this would mainly be interfacing the laser scanner head with 3D Printing software that slices the CAD model and feeds each cross section into the scanner which directs the laser beam down onto the metal powder. 
 
One idea we have is modding a laser engraver (as all of the equipment and hardware is al ready there) instead of starting from scratch. Unfortunately the software that runs laser engravers is very closed so we need to find ways around this.
 
This video gives a good overview of the process, you can imagine what type of electronics will be required.
 
I am sorry for the very quick overview regarding the Metalbot electronics. I can provide a more in depth explanation tomorrow if required. 
 
Sincerely,
 
Jethro Hazelhurst.

[Corporate666]

It seemed like spam because I saw the same thing posted on lots of different forums all from a new user on their first post - anyway...

The challenges with an "open source" metal printer is not the platform or electronic control.  There are few areas of electronics that have more open source hardware and software available than motion control - so that part is comparatively simple. 

What is not simple (or cheap) is the laser.  The commercial SLS machines are using 200-500 watt YAG or Ytterbium lasers.  It's not something you can do with a cheap Chinese CO2 laser.  Also, starting with a laser engraver won't help because the SLS machines scan the laser, they don't move it in X/Y as an engraver does.  The laser point size at the powder is tens of microns... moving it in X and Y would be much too slow. 

The cost of plastic 3D printers has come down because of all the motion control stuff being easily available and dropping in price.  Combined with the ease and availability of stuff like Arduino and R-Pi, it lets lots of folks get into the market.  The other part of 3D printing is just melting controlled amounts of plastic at a controlled rate.  That's really easy.

But SLS requires a very high power and very expensive laser.  There is also not much (that I know about) in the way of open-source or widely available laser scanning systems with the precision required for scanning a laser beam for sintering.  Those are the two tricks that would need to be solved for any sort of SLS machine.  Not to mention the questions about safety (both in the laser power supply and laser beam itself).

[metalbot]

Hi Corporate666,
 
Spot on. No.1 Problem = laser
 
Commercial systems use high wattage (the EOSint M280 standard model uses a 200w fiber laser and has a focus of 100-500um check http://www.eos.info/systems_solutions/metal/systems_equipment/eosint_m_280) We are looking at Fiber Laser engravers (not CO2 lasers which have 10x the wavelength and do not affect metals).
 
We are looking to use a laser engraver similar to that in the link below...
 
http://www.metalbot.org/forum/viewtopic.php?f=8&t=29
 
These lasers have an 'engraving' depth of up to 300um, so we will broaden the 10 micron beam to prevent vaporization of the powder (100-30um layers). The reason EOS use 200-400w lasers is for production speed. They use relatively coarse powders (Arcam have a shocking part resolution of 300um!)
 
So if we are going to go cheaper > we need to go smaller laser > we need to go smaller powder layers. This has benefits and drawbacks. The two main pluses are using smaller powder layers means that the sintered material will be denser and therefore stronger. It also means greater part resolution.

Main drawback is build time. However this is not a major concern for us hobbyists...yet.
 

The laser point size at the powder is tens of microns... moving it in X and Y would be much too slow

 
I am not quite sure but I think you mean using an X-Y Plotter would be too slow. We are going to use a 'laser scanner' which is two mirrors connected to high speed motors called galvos. This allows us to scan a large surface area very quickly and is the same technology used in the EOS. Depending on the angle of the mirrors you can plot X-Y coordinates very quickly and are how YAG laser engravers work.
 
The trouble is that the software and design of the scanner and indeed laser itself is not open and so hacking would be difficult.
 
Jethro.

 

[Corporate666]

Hi Corporate666,
 
Spot on. No.1 Problem = laser
 
Commercial systems use high wattage (the EOSint M280 standard model uses a 200w fiber laser and has a focus of 100-500um check http://www.eos.info/systems_solutions/metal/systems_equipment/eosint_m_280) We are looking at Fiber Laser engravers (not CO2 lasers which have 10x the wavelength and do not affect metals).
 
We are looking to use a laser engraver similar to that in the link below...
 
http://www.metalbot.org/forum/viewtopic.php?f=8&t=29
 
These lasers have an 'engraving' depth of up to 300um, so we will broaden the 10 micron beam to prevent vaporization of the powder (100-30um layers). The reason EOS use 200-400w lasers is for production speed. They use relatively coarse powders (Arcam have a shocking part resolution of 300um!)
 
So if we are going to go cheaper > we need to go smaller laser > we need to go smaller powder layers. This has benefits and drawbacks. The two main pluses are using smaller powder layers means that the sintered material will be denser and therefore stronger. It also means greater part resolution.

Main drawback is build time. However this is not a major concern for us hobbyists...yet.
 

The laser point size at the powder is tens of microns... moving it in X and Y would be much too slow

 
I am not quite sure but I think you mean using an X-Y Plotter would be too slow. We are going to use a 'laser scanner' which is two mirrors connected to high speed motors called galvos. This allows us to scan a large surface area very quickly and is the same technology used in the EOS. Depending on the angle of the mirrors you can plot X-Y coordinates very quickly and are how YAG laser engravers work.
 
The trouble is that the software and design of the scanner and indeed laser itself is not open and so hacking would be difficult.
 
Jethro.

As for that chinese laser, maybe it will work, but I would be amazed.  Even the one which claims <20W is only 5-10% of the power that the commercial ones use.  It may not only be a limitation of speed... I am not sure what the thermal characteristics of the powder is, but it's possible you cannot heat it up enough with only 20W because the energy is dissipated before the critical temperature is reached. 

As for the engraver, you mentioned starting with a laser engraver rather than starting from scratch, I thought you were proposing to use one of the Chinese CO2 laser engravers as a ready-made motion platform.  I am familiar with galvos but I am not sure how accurate they are - considering you are dealing with a 0.001" beam size that needs to be super accurate over a not-insignificant distance. 

I wish you luck with it, but I think this is a very very difficult project.  I think that even the big money SLS metal machines have serious drawbacks.  So much so that many of them perform intermediate machining steps on the sintered metal.

[metalbot]

As for that chinese laser, maybe it will work, but I would be amazed.  Even the one which claims <20W is only 5-10% of the power that the commercial ones use.  It may not only be a limitation of speed... I am not sure what the thermal characteristics of the powder is, but it's possible you cannot heat it up enough with only 20W because the energy is dissipated before the critical temperature is reached. 

 
Good point about about the critical temperature requirement. There are 50w and 75w variations of the laser engraver now we're talking 25-35 power percentage of the commercial systems. Now when considering that EOS use 100um layers, we are looking to go down to <30 micron layers and can buy 3um titanim granules. 
 
Another idea we have kicking around is using a preheating laser with a wide beam and a smaller spot laser in the center of the preheating beam that pushed the temperature beyond melting point (localized heating). 
 
So the inherent problems we face with the actual process are:
 
 
    Warping of the printed part due to shrinkage after cooling
    ?A powdery finish
    Removing powder from enclosed spaces
 
Jethro.

[metalbot]

Just a quick update for the interested... there are some top notch ideas going around!

We have decided, for now at least, to put a "laser scanner" on the back burner and go with a Gantry. So some form of CNC software in combination with a slicer will be used.

Check out some preliminary designs for the Metalbot printer here... http://www.metalbot.org/forum/viewtopic.php?f=4&t=111

If you have google sketchup you can download the model here... http://sketchup.google.com/3dwarehou...74&prevstart=0

Best!

Jethro.

[thefatmoop]

This is a pretty off the shelf solution and probably won't work well.

Try taking the icing addon for some of the 3d printers and put PMC (precious metal clay) in it. It's a clay that jewelers can quickly make jewelry with. Then bake the PMC and the clay substances will bake out and all you're left with it a metal object.

Downside is that I don't think it's super strong, and it tends to shrink a little... and the stuff isn't cheap

[FJV]

Commercial systems use high wattage (the EOSint M280 standard model uses a 200w fiber laser and has a focus of 100-500um check http://www.eos.info/systems_solutions/metal/systems_equipment/eosint_m_280) We are looking at Fiber Laser engravers (not CO2 lasers which have 10x the wavelength and do not affect metals).

There are several things that happen when a laser beam cuts a metal.
- Reflection (a percentage of the beam energy is reflected)
- Absorption (a percentage of the laser beam energy is absorbed and turned into heat)
  This will be the percentage that will do the fusing.
- Transmission ( a percentage of the laser beam energy "shines" through the material)

These percentages are determined among other things by wavelength of the laser beam and the complex refractory index of the material.
This results in the fact a CO2 laser will cut glass, but a NdYAG laser wil not cut glass. (for instance)
Also polarisation of the laser light plays a role.

I would suggest you'ld look into calculating these 3 factors to arrive at the optimal material/ laser wavelewength /laser energy combination.

At higher laser intensities there will be plasma forming and coupling, but for your application this may not be the case.

Also make sure no metal vapors are generated or that such metal vapors are dealt with, because you don't wanna cause metal poisoning.

Some idiots on this forum think that all you need are safety glasses when operating a laser, of course that all depends on what you're doing / cutting with that laser.

Also don't automatically believe a damn thing I say and check everything I posted, because I am not immune to the arrogant know it all engineer effect.

Getting advice from a credible expert in the field of lasers would be a good idea. A laser manufacturer might point you in the right direction to such an expert.

At least you have a lot of interesting questions to ask.

[FJV]

Also because you're using the laser on powder, you get multiple reflection.

And don't get set on the beam delivery system just yet.

You could end up with a mirror galvanmeter system with a flat field lens, to use a bit of jargon.

Also you could do tests first by having someone fire a laser of the desired type at the desired powder. (For instance a laser engraving company)
A failed test will prevent a lot of moiney being sunk in an unworkable design. (failure is always an option)

And you may also wanna look into Q switching and beam caracteristics.

That is about all from the top of my head, if I remember correctly.

Have fun with it. 

[metalbot]

@ thefatmoop, selective laser sintering currently provides the best printing strength and resolution for both plastic and metals objects. That is why it is the primary application currently used in industry, commonly referred to as additive manufacturing. It also does not require any post processing such as the baking off of binders or infusing the "green" parts with a different metal such as bronze. Selectively melting powders is a technical challenge mainly due to finding a suitable laser... but it is the best way to go.
 
 
@FJV - Thank you! Great suggestions and info...
 

There are several things that happen when a laser beam cuts a metal.
- Reflection (a percentage of the beam energy is reflected)
- Absorption (a percentage of the laser beam energy is absorbed and turned into heat)
  This will be the percentage that will do the fusing.
- Transmission ( a percentage of the laser beam energy "shines" through the material)

 
So CO2  lasers have pretty much been discounted because they have a wavelength approximately 10,640nm which does not work well with metals. On the other hand, laser diodes, Nd:Yag lasers and Fiber lasers all operate at around 1000nm which is ideal for metals.
 
For us as hobbyists, the most affordable and accessible solution is that of the laser diode, but unfortunately the beam quality is not that good for high power laser diodes.
 
The best bet so far is fiber coupled laser diodes (with a fiber core of apprx ~ 100um) and a 1:1 focusing lens. There will be some interesting experiments related to this in the up and coming weeks  .
 
Safety is high on our documentation list. Metal and inert vapors, radiation hazards all need to be thoroughly looked at. It is all going into the wiki...

Getting advice from a credible expert in the field of lasers would be a good idea. A laser manufacturer might point you in the right direction to such an expert.

 
Funny you should mention this, it was on my to do list this evening .

And don't get set on the beam delivery system just yet.

 
For Metalbot V1.0 I am pretty sure we will go with a conventional X-Y plotter. It is just too complex to get a galvo scanner and optics going for now. But a galvo scanner is the ideal option, still though, on the back burners.

Also you could do tests first by having someone fire a laser of the desired type at the desired powder.

 
There are a few people over at Metalbot.org who are actively experimenting with metal powders and BIG lasers... should prove very interesting
 
All the best, 
 
Jethro

[FJV]

For Metalbot V1.0 I am pretty sure we will go with a conventional X-Y plotter. It is just too complex to get a galvo scanner and optics going for now. But a galvo scanner is the ideal option, still though, on the back burners.

You could have 2 linear moving mirrors (x/y axis) under 45° to the laser beam and a small focussing lens.
This may save weight that you have to move around. (Less weight is often less problems)
You also reduce wires to moving parts.

Also you could combine several lasers with a single lens (all lasers combine on the same lens focal point).

[KTP]

I don't know if Metalbot is still active here, but I am going to get a 1mJ IPQ doped fiber laser and could perform some tests for him.  I calculate the peak power at 20khz to be around 10 kilowatts, focused down to probably 10um or so.  Average power is low (20 watts).

[metalbot]

Hi everyone...

@KTP, sorry for not checking back sooner... I can give you my Email if it is not too late! 

A quick update... It has been a few months and the Metalbot project is slowly coming together.

Check out what is going on: http://www.metalbot.org/

We have had loads of great ideas and there are quite a few very interesting projects unfolding at the moment (take a look at the projects section).

We are slowly working towards a preliminary design and there are a lot of options now so if you guys have any ideas, please send them our way!

All the best,

Jethro.