One other thing to try, since you didn’t specify. If the head is bouncing during a print, while it’s laying down a long line, then it’s your row height. Too small of a row height will cause the head to drag and therefore skip, and therefore “wave”. So check row height. The best troubleshooting guide for good calibration is on the reprap site. I use it all the time:
http://reprap.org/wiki/Print_Troubleshooting_Pictorial_Guide
I think there was a case of a small-ish mercury spill in cargo bay that did not take the bird down but resulted in the hull having to be scrapped or something like that.
http://www.iasa.com.au/folders/Safety_Issues/others/mercurycalling.html
There are cases of hull losses or expensive repairs after corrosive cargo leaks. (Ship whatever you want, fake the paperwork, but for the sake of the airplane gods, pack it well!)
Also, never ship mercury in glass containers. Avoid even carrying larger ones in pockets. The thing sloshes, and impacts the walls and can easily cause a breakage. A quarter century ago I almost witnessed this happening in a friend’s pocket (chemistry school is fun!); I only saw the escaped mercury balls and heard the story.
Well, off to the annual industrial machinery fair with me!
A bit more specific than @shaddack’s explanation (but I had to look it up and can’t vouch for it):
Gallium attacks most other metals by diffusing into their metal lattice. Gallium, for example, diffuses into the grain boundaries of aluminium-zinc alloys or steel, making them very brittle.
Todo: update the article with gallium and perhaps other metals. And add some more liquid/solid embrittlement-related cases… Cannot do it now.
I always mix up gallium with thallium (that’s the favorite of the poisoner in The Young Poisoner’s Handbook).
Indeed, when my father was at school his class did a science experiment to find the boiling point of mercury. Open beakers of the stuff over Bunsen burners. The whole class. In pairs.
(I’ve never thought of it before, but I wonder what sort of thermometer they used to measure the temperature?)
I’ve done the gallium/aluminium can thing and watched the can basically turn into aluminium paper with my son as a chem lesson.
You can use water afterwards to answer that question - the aluminium product essentially dissolves (pls don’t ask me for the equations) while the gallium gathers on the bottom. I’ve warmed it up to gallium’s melting point and the gallium coalesces and you can do it all again.
It’s (a) fucking awesome; and (b) safe and easy.
If it’s not safe, please don’t tell me.
A very handy phrase I will steal and use!
HANDS OFF THE SUBBIE! Besides, I’m more of a taint than an asshole, really.
Ok then serious question. Is there a way that one could liquefy a gallium/aluminum combination, PRINT it with a 3D printer, harden it and then boil the gallium back out of it leaving the aluminum. Or something like that. ???
Ousting gallium out of a solid solution with aluminium is likely to be a very difficult to impossible task.
But maybe there could be some sort of printing using some chemical change that’d deposit metal?
What would certainly work, and could be excellent for jewellery, is extrusion of art clays. There are modeling clays based on gold, silver, bronze, and possibly other metals, that after burning in a kiln (sometimes in bed of activated carbon to provide reducing environment) yield a solid metal piece. (There is some shrinkage but it is quite predictable.)
What about dual-extruder machine that’d extrude such metal clay from one, and carbon-based support from the other one (with the same shrinkage coefficient)? Then we would end up with a charcoal object that after thermal treatment yields a metal object.
What about a paste or powder or wire that’d be laser- or microplasma-melted immediately after deposition?
Great ideas. I would love to build a home brew SLS. The laser part would be the easiest aspect. The harder part is constructing the gas tight chamber with integrated electronics. Even harder than that would be sourcing the uniformly fine dust and perfecting the delivery mechanism. Laser sintering at home. It would be revolutionary. I’ve seen SLS titanium parts. They are unbelievably cool.
Join the club!
Quite expensive, though. A good laser can be pretty costly.
The gas chamber I’d say is easier than the laser. You have to wash it with argon (maybe even CO2 would work for some metals) but then just keep it at slight overpressure and the inevitable leaks will translate from a major problem to a minor cost of gas; the chamber wash would be way more costly anyway. If you want seal problems, work in vacuum.
That’s my major worry.
Smaller particles are less of a problem than larger ones; sifting may be needed as preprocessing for cheaper powders to remove off-specs grains that could cause problems in layer deposition.
The delivery mechanism will be quite a challenge.
Crank up the power and have selective laser melting. Full-density parts, FTW! 
I saw quite some such things at an industrial fair yesterday. They are becoming quite common in the industry.
If you’ll go this route, I’d suggest starting with gypsum as powder and water droplets dispenser instead of the laser. (You could possibly even add CMYK dyes and have colorful printouts.) That way you’ll solve the powder-related issues, and have something that may make money to pay for the laser and gas for Mark 2.
Another interesting and cheap material for Mark 1 model is powdered sugar. You can sinter it with ease with even a low-power (few watts) laser diode - inexpensive, easy, and you can dispose of the failures (or not-needed-anymore objects) by eating them. Once it works, you can upgrade to more difficult to source and more expensive polymer powder. Then once this works crank up the power to eleven and add shield gas and go full-metal.
So much meat here!!!
I had thought about ganging laser diodes, since 1W types are cheap. Develop an auto focus mechanism that relies on visible pilot beams. That way I could use it for a group of cheaper 40w co2 tubes. More replaceability, like a bunch a cheap light bulbs. I like this idea better than a big fat expensive 1kw tube or a yag or something.
As far as the motion electromechanics for laser cutters, I helped a friend put his together. It’s just like 3d printers & CNC, with the hardest part being aiming and proper focus.
For a home SLS, I’d figure out a cheap way to ramp up the power into the 1-2kw range. Then figure out a way to send the light into the gas chamber through a porthole rather than having everything inside. The motion electronics: piece of cake. The details… Lots of trial and error. I like your idea of starting with sugar.
With laser diodes you don’t need visible pilot as the cameras you’ll have to use for the autofocus are seeing in near-IR anyway.
I like that. Laser beam combining is however rather tricky because you have to get the beams in-phase, otherwise you can end with destructive interference eating all the power. Or you can use noncoherent mixing…
Military works on directed energy weapons. Some details could be useful here.
Could the lens actuator assembly from a CD/DVD head be of use here? (Not on itself, but as a knockoff?)
Beam combining.
Easiest part, I’d say. A window from zinc selenide or (perhaps better) silicon for CO2 lasers, a regular glass with antireflective coating may do the job for near-IR or visible (think frequency-doubled Nd:YAG). If your combined beam is polarized, a window in Brewster’s angle may be helpful to reduce stray reflections.
For CO2 lasers, it is said that hard drive platters can do a good job here.
For solid-state, you want to go the DPSS way. Laser diodes are way more efficient for pumping than flashlamps.
True that. Though for the speed you may want to use a mirror on a galvo (pair of mirrors, actually) instead of classical xy positioning of the mirror. With a high power laser you could cut down the time to scan the layer significantly.
You can also consider going with vacuum, and using an electron beam for selective melting. Would be a col thing to do. DIY vacuum tech is grossly underrepresented.
Sweet, isn’t it? 
Absolutely yes. Durrh, on my part.
Also, not even cameras. Just an array of pickups: the laser fires at it and the array signals the motors or screws to turn to pull it all on-center.
This is probably what will bring SLS to the everyman. That issue of interference is significant though. I bet there is a way to control the power feeding the lasers, or synchronize it in such a way (imagine pistons firing in sequenced sets) instead of trying to analyze the phase of the light and adjust with a feedback mechanism and software. Without knowing anything about it, it’s probably way more economical never to try to analyze light and do everything on the backend electrical.
That would be the simplest place to start. Invent the parts that need invention; but if it’s already been done, then use off the shelf or scavenged as much as possible.
I truly admire people who have MADE their own CO2 lasers from hardware store parts. One woman on hackaday.io posted a project where her goal was to spend less than $100 and only use parts from her local harware store, and she did it: sputtered her own glass, whole nine yards. I think that is AMAZING. But unnecessary. 40W Reci tubes on ebay are $150. If they were $1500, I would say make my own! But if they are $150, that is basically an expensive light bulb and if it breaks, my financial world isn’t over.
OK, next question. How would one measure the precise amount of laser needed to sinter and then full melt a metal? What I’ve read pegs SLS in the kw range. But what if that’s unnecessary? How would we calculate the amount of energy needed in a specific surface area, say a 3D printer nozzle sized area of about 0.25 to 0.5mm diameter? And then, how to translate that requirement back to determine the wattage of laser needed?
What are the physical principles and laser-specific technicalities that I need to figure out? I don’t have the terminology in my vocabulary, except maybe the metal-melting part, such as “joules-per-mm^2” to achieve a specific temp on a specific medium. I could probably figure that part out with some work. But for the laser part… I’m lost.
How to get the array on the surface that we want to focus at? Isn’t it easier to focus a low-power beam, and move the focusing elements until you get the smallest spot on the camera?
I’d go for an array of few elements if I’d be aiming for very fast reaction, e.g. how the CD tracking/focusing is done.
Light is easy-ish to analyze if you don’t want too much resolution or speed. If you convert it to an image, a cheap webcam will do.
For noncoherent combining, the best power control strategy is “crank it up to 11” anyway.
The CD head assembly is too small for a laser of this kind. But it can be easily scaled up. The actuators are just a magnet with pole attachments to get the magnetic field into a gap, and a moving coil there. A voice coil assembly, de facto.
I agree!
Homemade allows you using more exotic architectures than what’s available on the mainstream market. You could go for microwave excitation, or for transverse discharge, or route the beam through the gain medium in a way that prolongs its path so you can use a shorter tube for the same power…
Good logic. Ganging a bunch could also be a good power increase strategy for those cheap laser cutters.
I’d go for a calculation first, then try a range of powers on a sample. It will depend on too many variables - powder type (thermal conductivity), grain size (density - thermal conductivity, surface area available for sintering…), shield gas used (thermal conductivity), wavelength of laser (reflectivity of the material…), and so on and on and on.
I’d go for the amount of energy needed to heat and then melt a given volume of metal (thermal capacity times temperature difference, specific heat of melting). Then a coefficient for reflectivity (how much energy the powder reflects away), for heat conductivity (how much will be conducted away to the surrounding material). Then a coefficient for safety, to have some power margin; the more margin, the better.
We will have to focus the beam to smaller area than 0.25 mm, to achieve enough power with weak enough lasers.
The smaller the spot, the less energy we need. The slower scanning speed, the less energy we need (to a degree).
It’s more watts/mm2; also, the speed of delivery of the energy matters here a lot. Too fast, and we ablate the surface and the underlying material gets nothing but a shock wave. Too slow and the heat gets conducted away and we end up with warm powder.
There’s a lot of variables here. I’d take a look at the specs of existing machines and start from there.
Edit: The SLS-for-everyman will probably come from DPSS ytterbium fiber lasers. But that’s just my guess…
fyi, with these, esp when you get over 100w, the problem becomes less “laser power” and more ability to deal with the materials being burned. Evacuation of carbon is a huge problem. Also focal length. Because as you dig deeper into, say a 1" block of hardwood, the more air you have to blow at it to evacuate the carbon and the more the top edges of the wood are going to be burned by the sides of the beam. You could have a big awesome 500w laser, but if you don’t have a real powerful air blower directed at just the right place, and a very tight, long focal length, and some way to move the laser in a z direction during a deep cut, all you’re going to get is the same result you’d have with a 100w laser: a piece of wood that has edges burnt to a crisp. Lasers are awesome, but they also have their issues, like anything else. Thanks for that long response. I need to go read it now!!!
I am reading this now: DIY Selective Laser Sintering FAQ - RepRap
Also, thanks for the tips. I like your idea of a directed beam of plasma on a short-cycle burst to melt but travel fast enough so as not to pierce. The problem is that SLS is powder so it would blow away. But there is nothing saying something couldn’t be built from continuously dispensed metal tape or filament just ahead of the plasma.
Once I have my plasma table built, I may turn to lasers, esp the ability to additively manufacture with metal.
From Proto Labs SLS printing services:
“Resolution
SLS parts are available in one resolution: layer thickness of 0.004 in. with a minimum feature size of 0.030 in. in most materials.
Tolerances
Typically, expected tolerances on well-designed parts are ±0.010 in. (0.25mm) or ±0.0015 in./in. (0.0015mm/mm), whichever is greater.”
I think this is typical of SLS, and you’re right. Resolution is MUCH smaller than 3D printing. The titanium parts I played with had very miniscule resolution.
This topic was automatically closed after 5 days. New replies are no longer allowed.
