Okay, this is a sort of "best of" idea from the library research and thinking I have been doing.
Neil gershenfeld and friends have a patent on a printer that can use" digital materials" essentially really small blocks that fit together and come apart : [
www.freepatentsonline.com] . Interesting. However the problem is that the object would not be very strong or dense with the method shown. However you could modify the shape of the blocks they are using to make it nearly fully dense (except for the small gaps between blocks because they do not fit perfectly).
Then, after you are done building the object, use so called isostatic pressing combined with ultrasonic vibrations to bond all the blocks together. [
www.faqs.org] . Build the object, and every voxel has to be filled by a solid, in areas you want a void it would be water soluble or something for easy removal. Then basically encapsulate the built object in some plastic or other material, and surround it with fluid, pressurize it to compress the object then the build plate base can be the sonotrode. Apply ultrasound and your gravy.
The extremely interesting thing here is that the connections bonds could be 100% of the strength of the material the blocks are made from even if the ultrasonic bonds are not perfect. This is because the bond in a GIK block is a larger surface area than a cube if you see what I mean. It's like a dovetail joint suppose you want to join 2 beams together at the ends - if the bonded layer has a tensile strength of say 50% of the rest of the material, but it is being pulled on at a sufficiently low angle due to the dovetailing at the end of the beams, where they join, rather than at an angle perpendicular to the bonded face and the area of the bond is larger than the cross section of the two beams, then the joint can be de facto stronger than the material even.
Awesome. To the uninitiated it may sound like a simple idea and like it may have little benefit over say DMLS, however this is one of the only ways, and currently looks like the best way, to overcome the 2 biggest problems for good full strength material printers that can print to tight tolerances (like 8 microns) and make pre assembled assemblies:
1. thermal distortion/ residual stress that occurs with most melt deposition in most materials
2. temperature incompatibility between materials - e.g. depositing metal on top of plastic would burn the plastic.
Bonding the blocks is a real issue and thank goodness for ultrasonic welding or although explosive welding or other welding methods that depend more on pressure than heat might similarly be used maybe, as mentioned in the second patent.
I'm sure there would still be problems especially achieving good bonding throughout a large object, and also any distortion that may occur during the welding has to be kept down. These could both be helped a great deal by using larger blocks, or blocks of different sizes, small ones to do the fine features at the interfaces between materials, and the large ones for the interior of volumes filled with homogeneous materials. The surface finish might be an issue, if you tried to use squarish blocks, and simply approximate a curve with a large number of very small blocks. You don't really want your surface to have more roughness than a micron Ra or so ideally, to print decent bearings, ideally even smoother, which requires very small blocks that take longer to print with, are harder to make and harder to handle. For hydrodynamic bearings like used in an engine, you need smoothness that is too high to do with any blocks of size that are likely to be practical, a small fraction of a micron.
Another thing I can think of is that the welding requires that mutual movement can occur between blocks, however if it works for powder I assume I could work as well or better for blocks, plus although the movement in normal ultrasonic welding is on the order or 15 microns it may be practical to make it much smaller?
So you might want, instead, to decide on all the different curves you want to print, and come up with a set of exterior blocks so to speak, which have different radii of spherical curvature on one side, positive or negative, and can interlock with adjacent curveblocks, or squarish blocks on the other sides. Then you require a larger variety of blocks but they can be much larger. Something we could do right now is decide on the number of different blocks, that would be easy but we need to know the radii of curvature normally used in various economically important objects, and maybe would have to restrict the set so only certain defined radii can be made but that may not imposing any serious limitations for product designers. There may need to be angle blocks too. The blocks can be rotated when desired, etc. to further reduce the variety of different blocks. You get the idea. Figure out a good block set that will fit together well.
Then making the blocks... MEMS mass production technology can help here, microscale sintering or casting into micromolds might do it.
Edited 1 time(s). Last edit at 08/31/2011 10:56PM by GreenAtol.
