Talk:Working face aggregation
From Wise Nano
Manipulator vs gripping arms by 4.229.18.94 18:58, 15 Oct 2004 (CDT)
You probably want the gripping and manipulator arms to be identical. That way they can alternate, one gripping the block it just placed while the other is placing the next block. It also increases the level of functional redundancy.
That was me by Brett Bellmore 07:48, 16 Oct 2004 (CDT)
About the two manipulator arms alternating; Couldn't log in last night.
Scaling an replication of working face aggregation nanofactories by Brett Bellmore 08:17, 16 Oct 2004 (CDT)
The working face of the nanofactory can consist of smaller units, seamlessly tiled together. In this way, the cross sectional area of the array can be increased at need by simply extruding more tiles, and adding them to the edge. And tiles which have become too damaged to function can be replaced.
Conceivably this could be done automatically; A tile along the edge of the array would extrude first a new tile, and then a placement mechanism holding the tile. Then the placement mechanism would install the tile, and finally be disassembled, and it's component blocks stored for future use.
Which raises the issue that this form of nanofactory can function as more than a factory. If blocks are not chemically bonded, it can disassemble as well. And by simply extruding the apropriate surface of mechanically bonded blocks, it can assume any surface finish or function desired. Keyboard, display terminal, counter-top, floor or wall or ceiling.
(no topic) by Darius Bacon 10:57, 19 Oct 2004 (CDT)
The article refers to 200nm blocks, and then in the next paragraph to 100-micron precision. Is one of these a typo? If not, I don't understand it.
Brett Belmore's note about extruding tiles reminds me of how bacteria expand their cell walls.
Not a typo by Brett Bellmore 15:53, 19 Oct 2004 (CDT)
200 nm refers to the size of the blocks, and 100 nm refers to the positional tolerance in placing them, assuming that they're assembled on a bias. The idea is that if you assemble out of cubes, but with the "lattice" at an angle to the working face, you will be presented with a series of corner shaped pockets which would help guide the block placement.
This assumes, of course, that whatever fastening technology you're using doesn't require more precise placement. Some of the proposals, such as interlocking groves, would definately require much closer tolerance placement.
Solution in that case by Brett Bellmore 16:10, 19 Oct 2004 (CDT)
Of course, the solution in that case would be an end effector on the placement arm, which pilots off of neighoring corners. In that way, the arm would only need sufficient precision to make sure the pilots engaged the intended corners, and enough compliance to allow the piloting to fine tune the block placement.
Something similar is already done in some industrial applications on the macro scale.
A serious question, of course, is how to handle voids and partial blocks. One solution would be to have "filler" blocks, which the same outward features, but made of something which could easily and automatically be removed afterwards. If the product being made has even a little heat resistance, for instance, you could make the parts of the blocks which are to later vanish out of a substance which would sublime at elevated temperature, or alternately be disolved by some solvent which wouldn't harm the permanent features.
An alternate solution would be to construct in any voids a space filling mechanism designed to remove itself subsequent to the product being finished.
"Not a typo" by 67.114.230.235 16:26, 20 Oct 2004 (CDT)
OK, I thought that's what you meant. I went ahead and fixed the typo that's not a typo. :-)
Vacum vs argon filled assembly enclosure. by Brett Bellmore 13:01, 1 Nov 2004 (CST)
It's my understanding that most of the energy consumption of the nanofactory is in making the nano-blocks, not assembling them. If the enclosure manufactured around the product is made of blocks which are capable of being disassembled and stored for further use, then the product enclosure can feasibly be made thick enough to support air pressure without buckling, and thus be a vacum enviroment rather than inert gas filled, as the energy cost of building and disassembling the enclosure would be small.
Specialized blocks capable of powered sliding motions relative to each other, in one direction, (Normal to the surface of the nanofactory.) while maintaining a gas tight seal would greatly facilitate this. I picture a sequence where the product enclosure is extruded at the same rate as the product, while maintaining a vacum. when the product is finished, and the bottom of the enclosure extruded, a small portion of the side of the enclosure would slide down, allowing air to enter. then the majority of the wall would slide down, allowing access to the product. On product removal, the rest of the wall together with the lid would be retracted and disassembled. At no time would the vacum integrity of the system be compromised.
Feasible density of arms on working face by Brett Bellmore 15:24, 1 Nov 2004 (CST)
Some of your comments, about expecting half of the blocks on the working face to be held at any given time, lead me to think that you think there will be essentially as many arms per square milimeter of working face as there are blocks. This does not strike me as feasible.
100 nanometer blocks will require a hole in the face of the nanofactory somewhat in excess of 100 nanometers on a side to emerge from. That hole will have to be surrounded on all four sides by structure which supports the assembly arms, as well as by whatever system is delivering the blocks to the working face. Therefore the hole spacing must be greater than the spacing of the blocks on the working face.
How much greater? In large measure this is determined by engineering requirements we can't yet quantify, such as the need to deliver power, and extract heat, from the face of the factory. However, we can say that assembly would be seriously complicated were the spacing of the holes to be other than an integer multiple of the spacing of the blocks in the working space.
So, assume that the ratio is 2-1. (I think it will likely be greater, perhaps 3 or 4 to one.) The number of holes per square millimeter will be a quarter as many as the number of blocks. Assign each hole two arms, which alternate at placing and holding. You now have half as many arms as blocks, and only a quarter of the blocks are held at any given time. And this is a best case situation.
A secondary consideration is that the blocks, as you propose orienting them, cover each other as they are installed. This also constrains the mechanism, prohibiting you from placing a block immediately next to a block which is being held. So it's not even possible to hold as many as a quarter of the blocks on the face. (Unless you do something tricky, such as placing the blocks in some kind of wave motion, with only a small portion of the arms working at any given moment.) A safe assumption would be that there are 16 blocks on the face (4x4 spacing) for each delivery hole, and pair of installation arms. So you're holding one sixteenth of the blocks at any given time. I would design around this assumption.
Installation arm by Brett Bellmore 15:39, 1 Nov 2004 (CST)
Assume the blocks are delivered on standardized pallets. The installation arm can be a hollow extension of the track system, capable of parallelagram motion. If it's comparatively long relative to the x-y travel, it can get by with two axis of activation.
The arm would be designed to fit within a standard pallet while collapsed, and the face of the machine would have power and data connections around each hole. Arms would themselves be standard blocks which, when delivered to a hole which doesn't have an arm, would expand to lock into it. Indeed, I'd design all the tracks in the delivery system this way; Straight tracking would simply be arms which didn't move, while switch points would be arms which retained their mobility.
The z axis of travel would be supplied by the pallet, which would also store addressing data telling the system where each given block was supposed to go.
Note that this system does require a route for the empty pallets to return by.
Damaged installation arms can, if they retain that capacity, collapse themselves, and transit the same routes that the regular pallets take to recycling. If they are too damaged to do this, a tool can be delivered on a standard pallet to see to their disassembly in situ.
