3D Laser Printer

Working with easy replication in mind, [Peter] is building a 3D laser printer. The majority of the machine is made from laser-cut acrylic held together by parts that are inexpensive and available at your local hardware store. In the end this will lay down a layer of powder, use a laser to fuse the powder together in the outline of your choice, then repeat. This is known as selective laser sintering which is sometimes used in commercial rapid prototyping and, like a lot of other cool technologies, came into existence as a result of a DARPA project.

Sorry folks, this is not a fully functioning prototype yet. [Peter] is searching for the right laser for the job and a source for the powder. If you’ve got a solution please lend a hand and let’s see this project through to completion.

43 thoughts on “3D Laser Printer

  1. its cool and all, but If I were doing it sourcing the laser and powder would be at the top of my list

    as of right now you have a lot of time and effort into something that does nothing

    best of luck to you

    1. agreed i recently ran across scrap nylon sls powder from filastruder its seems to be in good enough quality to still print with although ive yet to actually test it with a laser looking to test it with a 2w 445nm laser still worried the layers might not fuse at such a low wattage now seeing that some sls printers are using 50w fiber lasers and the powder is just perfectly white

  2. How strong will the printed parts be?

    This 3D printer seems way superior to the reprap type printers, most likely faster and with no need for overhang support.

    I’ve been thinking about building a reprap but now I’ll wait for a while and see how this turns out.

  3. @osgeld

    His first post actually begins with the search for ABS powder and a laser. I haven’t read everything yet but it looks like he built platform (is that you would call it?) knowing harder problems lay ahead.

  4. Yeah, basically he built the fun part first and saved the dull researching and experiments for later. Designing and building the machine to do the work involves straight forward solutions that have been addressed time and again. Getting the whole laser and powder bit working is a lot trickier, but not nearly as immediately satisfying.

  5. I would say the things to consider are absorption of the laser wavelength and thermal conductivity. A C02 laser would be well matched to many polymers. It might be possible to use a diode laser for polymers that are tinted such that the absorb the wavelength of the diode.
    Materials with low thermal conductivity would be good for lasers that run in a continuous mode.

    For a material like Al which is reflective to long wave IR and has high thermal conductivity a shorter wavelength laser that could be pulsed with high peak power would likely be necessary.

    I would say trying different lasers and materials is going to be the fun part of the project!

  6. I thought about doing a very similar thing, but I wouldn’t use a laser…

    I would use a cheap DLP projector from ebay. Color quality would not be an issue since I’d be using as a focused light/heat source. I might also hack it to use a more inexpensive bulb also. With the right optics one should be able to focus an XGA (or higher) resolution projected image down to a pin-prick of hot light (which could be highly controlled). This way one could move up from being able to print one pixel at a time (with varying degrees of accuracy) to a DPI rating comparable to a high quality ink-jet printer. The hard part then would be to find a way to stabilize the projector assembly, as vibration would become your enemy. Having the optics stationary would be best. The color of the powder one uses would also be an issue (the more opaque the better), since you wouldn’t want to over-fuse and loose the clarity of the print.

    Software would also be a breeze with a projector method, since it would simply act as a monitor showing a zoomed-in version of the layer currently being printed (in solid white), moving across as the print arm moves, etc.

  7. We had a commercial model at a place I used to work at. When it was working correctly, it did some very cool stuff. It also required an enormous amount of maintenance and cleaning. Used enormously expensive plastic power that ended up on EVERYTHING. Ran in a sealed nitrogen enriched inert environment. Was horribly sensitive to the temperature of both the part bed and the preheated powder before it was sintered. And the process to make metal parts involved a kiln after the initial sintering of the part.

    No way would I build one that worked like this. The cool factor no way comes close to offsetting the huge PITA that thing was.

    I still wish the guy good luck

  8. @pɹɐoqʎǝʞuǝʞoɹq No, one would not want to use aluminum powder for something like this. Sounds like a good idea at first, but aluminum is flammable if there is enough oxygen (thermite anyone?). Putting it in dust form and pointing a laser at it is just asking for a fire, assuming you could ever get the laser hot enough for the aluminum to fuse before it distributes the heat.

    If you want to make an aluminum printing device you’d have to use a chemical reaction instead of heat. Printing just the right amount of an aluminum solvent might work, but finding a way to print it through an ink-jet head without dissolving the head would be a challenge.

  9. @wdfowty
    I searched online for how to turn text upsidedown when I was looking for a name for my Gizmodo profile.

    @Max I was thinking black anodized aluminum powder in a nitrogen environment. Since there is laser etching which partially melts/evaporates the material, I figured anodizing would be redundant, thus didn’t mention it the first time.

  10. There are a number of reasons why reprap doesn’t do this by default. Inkjet print heads are expensive and proprietary, they are imposable to get cheaply or to print off. Inkjet print heads clog easily giving you a much shorter lifespan. Inkjet print heads a very detailed but very slow in the quantities we need. Inkjet print heads require much smaller steps, meaning we need either more expensive steppers or a smaller build area. powder deposition tends to be much weaker then filament deposition. making structures with powder would be very hard.

    That being said I know of a reprapper who is working on a reprapable printer that uses this method and commercial laser printer toner, which you can often dumpster.

  11. @Max I doubt the DLP could handle dissipating the energy from a light source that can be focused down to sinter aluminum or melt plastic. However, you could pump lower power UV through the DLP and irradiate a UV curing plastic, generating one slice at a time through raster rather than slowly through vector means.

  12. While I have no basis for this, I thought I’d chime in with my two cents. I work in a chemistry lab and we use IR spectroscopy for identifying and analyzing compounds. Liquids are easy to do, but for the solids we take a small amount of the sample and sprinkle it with either KBr or CsI powder. It is then compressed into a glass and an IR beam is shot through it.

    My take is that you could use a plastic or acrylic powder mixed with some CsI and while you wouldn’t want to pressurize it, a laser may be enough to excite the electrons and cause binding. The reason I suggest CsI is because it is absorbs but doesn’t reflect light from as far as 5000 cm-1 to 200 cm-1. It would not be affected by the laser, but the proper plastic would, causing it then to bind and form a solid glass/plastic layer.

    Like I said, I really have no evidence or basis for this. I haven’t done any research into the structures of powders and I’m probably wrong anyways. But I thought I would just share something. I’m not a polymer chemist also, so I don’t know much about synthesizing plastic polymers.

  13. Hi

    I am Ashraf

    I am very very very distinguished in solid geometry ,At primary and secondary school , my records in it was full mark spontaneously without any effort from me in study it is my talent and it is the gift of God for me

    3d modeling is my interest and hobby

    I love to use solidworks software

    so

    If you send me PHOTOGRAPHS OF any PRODUCT and want a 3d solid OF it I can do perfectly for you THAT CAN BE MOVED BY online shopper to see the product details and as it is “solid “made by solidworks it can be 3d printed very easy

    Also if you have a 3d model or graphic & want to have A 3d solid of it I can do perfectly for you

    I want to attach some photo of some of my 3d solids I draw purely in solidworks and I can do more difficult than these . with the help of God

    P.S I AM EGYPTIAN LIVE IN UPPER EGYPT I PREFER TO WORK FROM HOME

    God bless you

    thank you

    good bye

    Aِِِِِshraf

    ashraf349202@yahoo.com

  14. @FredP I couldn’t agree more, seems like they’ve done the easy stuff first while totally neglecting research into what’s actually required for the core part of the project. I think the power of the laser required has been wildly underestimated too.

    @chango You right about the problems with the DLP being able to take the power from a laser source. Then again I see no advantage of using a point light source on a mirror matrix.

    The DLP projector does however work great for proper MSL (rather than SLS) when using a resin with appropriate photoinitiator. Unfortunately the cost of such resin puts this technology out of reach for you average hobbyist/hacker/consumer.

    Good luck to the guy though, I think he’s gona need it!

  15. someone suggested using an array of ps3 optics in a line to combine the low power DVD writer laser beams into a single beam.

    would be a PITA to align but should work..

    You can get multi-watt IR diodes on ebay now, but they are extremely dangerous to work with and you need proper 808nm full power rated safety goggles and gloves.

  16. I seem to remember seeing something like this at the Engineering expo at UW-Madison about 10 years ago. It was a huge machine, and they said it took 3 days to fab a sprocket.

    I imagine it’s faster now, but if you can’t find the consumable component(powder) it would be pretty well dead on arrival.

  17. I’m not sure about a source for the powder, but you might ask around at some injection molding plants, they may have sources.

    You could probably pick up a relatively high-power laser pointer from Amazon (http://www.amazon.com/Laser-Flare-AixiZ-Illuminator-650nm/dp/B001HSSGEG/ref=sr_1_2?ie=UTF8&s=electronics&qid=1270249603&sr=8-2)

    Or just get a diode: http://www.amazon.com/POWER-Burning-Cutting-Laser-Diode/dp/B0036IRETS/ref=pd_sbs_op_6 (OTOH, IR lasers scare me.)

  18. @ulrich lukas , what about using intersecting bluray laser beams? one on its own can’t do anything to the plastic but three when correctly focussed (using ps3 lens actuator) should work.

    someone else suggested using a homemade lens made of existing non-ideal lenses held together with uv reactive (note this needs heat AND uv, or it won’t work) glue off Ebay.
    this is also a useful way to make diffraction gratings out of old HP printer positioning strips (glue together with the lines at right angles in the centre) for other experiments.

    Another worthwhile hack is to obtain a grey gradient filter from Jessops, shield the surfaces with easily removeable clingfilm then cut it along the gradient axis into strips. A simple LED/phototransistor pair will then be sufficient to measure the location to within 0.2mm or so.

    Yet another hack is to obtain a spare touchscreen and have your laser module set up with a pen attachment touching the surface. this is not the most accurate method but has the advantage of being easily calibrated and repeatable.

    #include “$0.02.h”

  19. I see all those peoples on the net experimenting with 0.5watt or higher lasers and i didnt see one of them useing laser googles, are they cracy ? If you use lasers with this high power and even one reflection hits you eye you are blind.

    The first thing you should do is buy a laser google (not cheep) and second make the 3dprinter out of something which the laser can not penetrate. Using acyrlic might look nice, but is very dangerious because nothing shields the laser. Commercial Laser systems are completly enclosed in a black box (or black room) where noone is allowed while the laser operates.

    Hope all those hackers keep it safe.

  20. Hi,

    First of all, in order to get the best resolution, you should try using a pulsed UV laser (because of the low wavelength). Even if they are expensive, you should be able get one easily. For the powder, I have been working on a way to locally change the index of the material used and fusing it should be easy (I have been using silica which is great a great compound thanks to an high the ablation threshold). Therefore, the physics behind this is mainly based on non linear optics that is why pulsed lasers are required (pulse duration less than the picosecond) in order to create a plasma without thermally affected zone. If you succeed into getting enough power for your laser, everything should be easy. If you need any links for publications, just ask me, I ll be glad to send you some on this field.

  21. You can make plastic powder with a lathe by rubbing the side of a cut-off tool on the face of some round stock.

    For production you’re probably looking at the same process except with a cnc and vacuum cleaner involved.

    Screening the resulting dust will get you the desired fineness. Start with ABS pipe in a manual lathe and go from there.

  22. @Max

    You are too quick to dismiss aluminum powder. It may be flammable but only in the presence of Oxygen. A vacuum pump, air tight enclosure, and welding gas solves this problem.

    Even if the aluminum did ignite under these circumstances: it shouldn’t be particularly dramatic unless the powder magazine was recklessly exposed, and the enclosure designed to produce shrapnel.

    It makes it unattractive as a consumer good, however the problem is not insurmountable to engineer around.

  23. hmm.
    has anyone considered using a reverse pulsed electroplating approach?

    someone on the high voltage forums mentioned the use of a syringe needle (dremelled flat)connected to a power supply and pump.

    the idea being that you “etch” away where you don’t want metal, and the pump removes the saturated solution before it can cause problems.

    this approach also might work for making PCBs with a few modifications.

    could be handy for the awkward to make parts with many fine metal components.

  24. Wondering why we need a laser for this. A condenser and focusing lens array, with a 500 watt or so halogen light should be able to get enough energy to melt plastic powder. The trick, which folks used to do with enlargers back in the day of darkroom enlarging, is to use the condenser stage to focus the light of the lamp onto a sheet of blackened metal with a hole in it, to mimic a point source like a laser. All the off-axis rays get cut off, and the problem of focusing the light to a point source is simplified. A point source enlarger could easily focus light enough to resolve the silver grain of the film onto the printing paper, so sub millimeter resolution shouldn’t be a problem at all. Probably would have to cool or at least heat sink the diaphragm and the condenser lenses.

  25. I did some test using the sun, a magnifiing glass and some powder from a factory that makes plasic parts by rotating powder in heated molds.
    The problem I ran into was that partly melted powder was sticking to the melted “part” leaving very rough edges. I think you may need a very fine powder.
    I itried to make finer powder by putting some of the original powder in a blender but I wasn’t able to cut the powder this way.
    I know they use some kind of rotating scissors for grinding up the powder…

    Anyway, I hope you will succeed cuz it;s a great idea !

  26. bothersaidpooh: “has anyone considered using a reverse pulsed electroplating approach?”

    This is called ‘Electrochemical Machining” and is a well established technique. None too fast though, and pretty power hungry for macro-scale parts.

    A non unrelated technique, ‘Electrical Discharge Machining’ is more common in industry but not so quite so easy for a hobbyist to build themselves, but it isn’t exactly rocket science.

  27. I suspect the acrylic powder used for powder coating might work. It’s readily available in all sorts of colors, and it melts and reasonable temperatures (300-500 f depending on the powder). Furthermore it’s not terribly expensive (by the looks of his machine it would probably run somewhere in the 5-10$ department to fill the tank. The only issue might be the strength/rigidity of the finished product. Since it’s designed to be coated on it might be more flexible than desired. (since I imagine a flexible scratch resistant coating is better than a brittle one)

  28. @JohnMc This was done back in 1995. It is called Multiphase Jet Solidification (MJS). Parts are printed using the thermoplastic extrusion method using a mixture that is 60% metal and 40% binder. Afterward, the parts are heated up until the binder melts away and the metal particles sinter into a solid object. The finished product is about 98% solid by volume. There is about 20% shrinkage from the green part to the metal dense part.

    The problem with using a laser to melt metal powders is that most of the harder metals ball up. Most Selective Laser Sintering machines that print metal use a harder material particle that is coated with a softer material like bronze. The laser then melts the bronze surrounding one particle to the bronze surrounding a neighboring particle. Once again, the part is cooked in an oven at an extremely high temperature until the base metal starts to melt. The majority of the finished part is steel or a different base metal, but there is always some of the binding material left in the part.

Leave a Reply to bothersaidpoohCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.