100μm deep is amazing! by cutting steel (and even stainless!) this machine is dramatically overdelivering on its process of marking metals. 100μm to 200μm is plenty thick enough to use it to cut holes in things, and if that only takes 10 passes, you're getting 10μm or 20μm per pass, which is also amazing
that suggests you might be able to manage aluminum foil, which is typically 10μm thick, and doesn't have all that extra aluminum behind the surface to heatsink it. and that would avoid the usual problems with making things from aluminum foil, which is that you can't do anything with it without wrinkling and/or tearing it
an alternative way to engrave or cut aluminum or brass might be to anodize or paint it first, use the laser to selectively strip off that surface layer, and then etch it with acids, bases, or electrolysis where the metal has been left unprotected
the markings in the micrograph seem to be perfectly sharp, straight, and smooth down to the resolution of the microscope, though the finest lines I see are on the order of 150μm across according to the grid (15px on a 104px grid). this suggests that the positioning repeatability is in the range of the specified 10μm, even if (as you say) the messy crater the laser creates is about 150μm across. 10μm out of a 200mm range is astounding resolution, that's 20k × 20k reliably distinguishable positions
the w&m levsha video i linked elsethread https://youtu.be/PAFBkgawH3w?t=2m10s did manage to cut parts out of razor blades (600 passes to get through 100μm) without any noticeable thermal warping, but i think his laser is weaker than yours; he can only ablate the metal oxide, not the metal itself. given that, it's surprising that his can mark brass and yours can't; possibly he's using a grade of brass with less copper and therefore lower conductivity, or maybe there's a relevant reflectivity difference which in effect makes his laser more powerful on brass than it is on steel
i'm guessing the cost to run is probably dominated by either the depreciation of the machine (say, a dollar a day whether you use it or not) or, with heavy use, the lifetime of the laser (say, 20k hours would give 10¢/hour) except that if i recall correctly you're in california where retail electricity is supposedly nearly 50¢ a kilowatt hour, which would be 10¢ an hour for a couple hundred watts. but with solar panels the energy cost would be about 20× less
has anybody reverse-engineered the protocol, i wonder?
OK, you got me curious. I took a candy tin lid, 0.22mm thick according to my micrometer. Using 80% power, 100 mm/s, 40kHz and a loop count of 300, I was able to cut through in about 3 minutes (letter K in your honor, Gill Sans, 1cm wide).
Seen through the protective enclosure (hence the green color cast):
that's super exciting! precisely cutting 220μm-thick steel sheet without warping it is an ability not to be underestimated, particularly when you can cut it to literally any shape you want to an x-y precision of 10μm, subject to i guess a minimum corner radius. 220μm ÷ 300 passes suggests you were vaporizing about 700nm of steel per pass, which seems pretty plausible; it's faster than the w&m levsha results but not that much
how efficient is this? heat of vaporization of iron is 354 kilojoules per mole, which works out to 6.3 kilojoules per gram, plus another 3 kilojoules per gram or so to reach the boiling point. if i estimate your cut width as 100μm and the cut length as 60mm (600ms at 100mm/s, 18 frames at 30fps), that's about 1.3 mm³ of iron, about 10 mg, which should require about 90 joules to vaporize it. this is about 500 milliwatts over 3 minutes, which is a lot less than 14.4 watts, so probably most of the heat is being lost to things like reflection and conduction; maybe most of the focus spot isn't getting hot enough and only the center is actually boiling (though the sparks suggest that some of the iron is being ejected in liquid form)
i suspect it could become a lot more practical with automatic focus
100μm deep is amazing! by cutting steel (and even stainless!) this machine is dramatically overdelivering on its process of marking metals. 100μm to 200μm is plenty thick enough to use it to cut holes in things, and if that only takes 10 passes, you're getting 10μm or 20μm per pass, which is also amazing
that suggests you might be able to manage aluminum foil, which is typically 10μm thick, and doesn't have all that extra aluminum behind the surface to heatsink it. and that would avoid the usual problems with making things from aluminum foil, which is that you can't do anything with it without wrinkling and/or tearing it
an alternative way to engrave or cut aluminum or brass might be to anodize or paint it first, use the laser to selectively strip off that surface layer, and then etch it with acids, bases, or electrolysis where the metal has been left unprotected
the markings in the micrograph seem to be perfectly sharp, straight, and smooth down to the resolution of the microscope, though the finest lines I see are on the order of 150μm across according to the grid (15px on a 104px grid). this suggests that the positioning repeatability is in the range of the specified 10μm, even if (as you say) the messy crater the laser creates is about 150μm across. 10μm out of a 200mm range is astounding resolution, that's 20k × 20k reliably distinguishable positions
the w&m levsha video i linked elsethread https://youtu.be/PAFBkgawH3w?t=2m10s did manage to cut parts out of razor blades (600 passes to get through 100μm) without any noticeable thermal warping, but i think his laser is weaker than yours; he can only ablate the metal oxide, not the metal itself. given that, it's surprising that his can mark brass and yours can't; possibly he's using a grade of brass with less copper and therefore lower conductivity, or maybe there's a relevant reflectivity difference which in effect makes his laser more powerful on brass than it is on steel
i'm guessing the cost to run is probably dominated by either the depreciation of the machine (say, a dollar a day whether you use it or not) or, with heavy use, the lifetime of the laser (say, 20k hours would give 10¢/hour) except that if i recall correctly you're in california where retail electricity is supposedly nearly 50¢ a kilowatt hour, which would be 10¢ an hour for a couple hundred watts. but with solar panels the energy cost would be about 20× less
has anybody reverse-engineered the protocol, i wonder?