Talk:Research scorecards
From Wise Nano
I vote for leaving out the cost consideration by Matt 08:14, 19 Feb 2005 (CST)
re system/product cost:
Is this question really that relevant? We all know what happened to Charles Babbages's Analytical Engine because of denied funding. Also we know that personal computer systems have had exponential growth in capabilities for decades while maintaining a relative constant, if not declining, price.
Personally, I feel it is best to leave out this cost consideration. If you ask about the possible data bandwidth or how much bigger the equipment is in proportion to its products, these questions will be meaningful at any time. But cost of system and products is a wholly different matter, and depends on many more factors than just the technical specifications. Price of system and products can change rapidly even though the specifications change little or not at all, because the value of the product is determined mostly by how much value the consumers believe it has. Is a tabletop nanofactory more useful when it has a low price? Arguably so, but does its usefulnes increase or decrease when the market discovers its (the nanofactory´s) value and the price accordingly rockets up? Or, viewed from the other side around: If the nanofactory produces colored diamond gems that could today sell for huge prices, will its usefulness decrease when people quickly realize that colored gems aren´t that valuable, after all, if they can be manufactured by the hundred thousands for next to zero cost? Look here for the "cultured diamond" matter, because that already is a very real question.
Cost as in production cost, not price by Chris Phoenix, CRN 14:36, 19 Feb 2005 (CST)
You are quite correct that the price of an item depends on a lot of factors including the consumers. What I was asking about, though, is the production cost. That will depend on technology and regulation, and (to a decreasing extent) the cost of labor.
A general-purpose rapid prototyping machine that cost $500,000 and made products costing $1000 per kg would be useful, but not revolutionary. But if the machine's products cost $1 per kg, and it was capable of building a copy of itself in reasonable time, that would be revolutionary. That's what I was trying to get at.
(no topic) by Matt 15:53, 19 Feb 2005 (CST)
A general-purpose rapid prototyping machine that cost $500,000 and made products costing $1000 per kg would be useful, but not revolutionary. But if the machine's products cost $1 per kg, and it was capable of building a copy of itself in reasonable time, that would be revolutionary. That's what I was trying to get at.
I still believe it doesn´t matter much. Here´s why:
While being just out of reach of most private individuals, such a machine would be affordable to most companies, and easily to every government, in numbers. Even if it cost millions, it would still be affordable to most governments and large companies (there´s no way IBM would pass up on that one). Obviously, the goal is to improve the former, expensive machine into the latter, powerful machine. Assuming it is possible at all, then if an organisation is dedicated to achieve this, costs will hardly matter, because no matter how high the costs of R&D, they will be constant. Spend them once and you get the powerful machine. Once you have a machine of the latter type, costs of replication and production (which will be essentially the same) will sharply drop to almost energy+material costs, so by then, costs won´t matter, either.
Of course, one has to consider that high initial costs can be discouraging and can lower the perceived total usefulness. The main reason why you have problems quantifiying this problem is that, unlike the other factors, it involves a lot of human motivation, regarding cost estimates as well as estimates about probability of success and payoff. These variables are so numerous and partially unrelated to the actual technology, the question of costs deserves a score card of its own.
Material cost may be expensive by Chris Phoenix, CRN 09:26, 20 Feb 2005 (CST)
I chose my numbers carefully. If the product costs $1000 per kg, and the machine weighs 500 kg, then a machine-built machine will cost at least $500,000. There'd be no advantage in using it to build copies.
If the machine could be converted to use $10/kg materials, then a machine-built machine would cost only $5,000.
Hm... most of the weight of the machine would be in relatively cheap/simple components. If the machine could build all of its components, then it might be worth building just the expensive stuff (printhead, computers, maybe drive system) and using just a bit of labor to snap together the housing, plug in the wiring harnesses, etc. So even at $1,000/kg, the partly-machine-built machine might drop its cost by an order of magnitude. That makes it a significant advance; but then there's still the problem of having to pay $1000/kg for any product.
(One might think that computers are cheap, so it's better to buy them off-the-shelf than machine-build them. But a fast 3D printing system will need a huge data rate, which implies specialized hardware--and lots of it. I worked on a 2D color printer (EFI Fiery) that shipped in 1991. It had a custom 25-MHz MIPS board and 64 MB of memory, and cost $50,000.)
Purpose of the scorecard? by Matt 11:48, 20 Feb 2005 (CST)
Maybe we´re talking past each other. The scoring table states:
0-60 points: The work is not very relevant to general-purpose manufacturing.
70-120 points: The work may be relevant, and may develop an enabling technology.
130-200 points: The work is likely relevant and may significantly advance the field.
210-260 points: The work is highly relevant and may produce breakthrough capabilities.
As far as I understand, this scorecard evaluates research regarding its relevance for leading to mature MM, as opposed to direct commercial application. Is that correct?
Proposed additions - production rate and energy efficiency by Tom Craver 15:22, 20 Feb 2005 (CST)
Chris:
I think it's worth asking where the hypothetical $1000/kg would come from. It shouldn't be labor. Eventually any "pure profit" component, such as that due to a monopoly, would disappear, one way or another. Any capital equipment can be reproduced with the system, quickly driving down costs if that were the main cost source. Of traditional cost inputs, that leaves input materials, energy, and land. I may have missed some, but at least those.
Land roughly corresponds to size of equipment, and a system that can't fit conveniently into a store - let a home - isn't going to be revolutionary. Size of equipment is variable - e.g. lithography can take a factory or be done in a lab, and could probably be miniaturized to a desktop system if that were of value. Probably we can assume that most systems of interest can somehow be reduced to under 2 cubic meters, which is also small enough to be widely used. So we could roll that assumption into another scoring.
If we take the MM system to include everything back to the point of cheap bulk feedstock chemicals, it won't be physical raw materials that create high costs. So that leaves energy per kg of product - kw-hr/kg - as an interesting measure.
A rough cut at the scoring might be something like:
Production rate for a system 2 cubic meters or smaller:
1kg/month or less - 0
1kg/week - 10
1kg/day - 20
1kg/hr or faster - 40
Energy efficiency for a production system:
1MW-hr or more per kg of product - 0
100kw-hr per kg of product - 10
10kw-hr per kg of product - 20
1kw-hr or less per kg of product - 40
Scorecard purpose; second scorecard? by Chris Phoenix, CRN 22:48, 20 Feb 2005 (CST)
Matt, you're right about "relevance for leading to mature MM." But that's ambiguous. "Leading to" implies that we don't care so much about cost. "Mature" implies that we do. I'm not trying to quibble; I really don't know whether to put in something about cost or not.
Here's why I'm leaning toward it: There are lots of MM systems that could be designed and built. But some will lead pretty directly to the inexpensive "flood of products" while others will not. A system that had to be programmed via DNA strands would be pretty slow. Of course that is covered by question 1. But there's no question that asks about feedstock costs. If the system builds DNA... I've been told that DNA precursors cost $10,000 per kg, IIRC. There are only a few things that are worth building at that price. Unless a DNA-building system were just a stepping stone to another system (and that step could be taken quickly), then it wouldn't be revolutionary.
Tom, I think that answers your question too. Without knowing the cost for feedstock (and if extremely high purity is required, even simple chemicals can be expensive) we can't know how generally useful the system will be.
I like the direction of your questions. Let me try to recast them even more generally than you did:
Time for the system to build its own mass of product:
- More than a month = 0 points
- Less than a month = 10 points
- Less than a week = 20 points
- Less than a day = 40 points
Energy and raw materials cost for a kg of product:
- More than $10,000 = 0 points
- Less than $10,000 = 10 points
- Less than $500 = 20 points
- Less than $25 = 40 points
And one more:
Improvement vs. today's technology (all that apply):
- Logic operations per watt improved by more than 4 orders of magnitude: 10 points
- Electromechanical power conversion density improved by more than 3 orders: 10 points
- Tensile strength improved by more than 1 order of magnitude (vs. carbon fiber): 10 points
I'm reluctant to add too many questions, though, for two reasons. First, I want the scorecard to be quick to use. Second, I don't want lots of questions to dilute the score. Especially not overlapping questions.
Thoughts?
Chris
Not sure you want to use $$$ by 68.2.106.58 02:15, 21 Feb 2005 (CST)
Chris: I tried to avoid using $$ costs. E.g. if a system could make solar power essentially free, and process raw materials from dirt, the input costs might be zero. But if 1MW-hr/kg of energy is needed, the system won't be very practical for most uses. The example of DNA is interesting, but I suspect that there is a high skilled labor cost in that price that would go away if there were a market for many tons of the stuff.
