Thoughts on RapRep
Jul. 20th, 2008 09:31 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
fivemackRapRep seems to be attracting a lot of blog interest at the moment; people are calling it a revolution in manufacturing, and I think this misses the point entirely.
Look around you for things made out of hard plastic which are three-dimensional: Tupperware boxes, tags on key rings, your mouse, your headphones, the fastenings on your luggage, your hairbrush, your toothbrush, the cap on most bottles you own.
For each plastic component of every such thing, there has in the past existed a pair of objects made out of very hard steel in the exact shape of the inner and outer faces of the component. Someone bought a block of quite expensive steel, from some company you've never heard of like Virat Special Steel, and programmed a machining centre - a box the size of a small shed and costing somewhere between what a house in a back-alley in Manchester cost in 1980 and what a two-bedroom house in central London costs now - to mill out the relevant shapes with artfully-designed rotating tungsten carbide knives designed by models in which the very hard steel is a fluid through which the knife cuts like the keel of a boat; the cutters come from companies like Iscar, who I'd heard of only in the business pages where it was praised as a sensible acquisition for Warren Buffett, without a mention of sintering or of titanium-nitride vapour deposited coating. The cutters cost a hundred pounds and last a couple of days.
They cut awkwardly-shaped holes with an EDM machine, which uses electricity and graphite electrodes to push an electric arc through even obdurate materials; the machine costs somewhere between a new SUV and a new Ferrari. Then they installed the pair of moulds in an injection-moulding machine, which is fairly cheap in comparison to the machinery that made the moulds, which heats them, runs molten plastic between them, presses them together to stamp the plastic to shape, pulls them apart to let the part fall out, and repeats, until either they'd made enough parts or the mould wore out. The mould wears out: get another block of die steel from Virat and make another mould.
This is a capital-intensive industry, and it's an industry in which the minimum sensible quantity of a product is the number that can be made in the lifespan of a single mould; the mould plus that quantity of output costs about a year's disposable income of a reasonably well-paid software person and fills a shed.
The additive technologies, of which RapRep is the most visible, let you make things out of plastic which vaguely resemble the products of injection-moulding - the finish is generally dreadful - without having to deal with the capital cost; if you'd wanted to make things out of wood, or of textile, or of cast metal, or even of machined metal or machined plastic, the capital cost for the tools has been small enough that they can be afforded out of disposable income.
For the last twenty years, software people have always owned the means of production - the computer they use at work probably belongs to their employer, but they have a (generally better) computer at home, which can do anything the computer at work can do, and which cost somewhere between a cheap and an extravagant holiday. This gives a self-image that they're not employees, in the Marxist sense that an employee is a person that a person with (injection-moulding-machine shaped) capital is obliged to get to work for him in order to get an income stream out of the capital. RapRep is giving them the possibility of making a wider range of real things while remaining not employees, either of a person with capital, or of the large expensive machine which they personally own but are obliged to operate in order to pay the interest on the loan that bought the machine.
I think what we're seeing is the reactions of computer people to noticing the vast ecosystem of manufacturing, and thinking that what manufacturing needs is a Version 2.0 because that's so often the popular answer, and indeed fairly often the right answer, to computer problems. But a cheap slow computer gets the same answer as an expensive fast computer, it just takes longer; a cheap machining tool doesn't give you products with the same surface quality as an expensive machining tool.
The person turning up with the raprep'd mobile-phone cover, the colours not quite matching Hello Kitty's official colours and the pattern of the plastic threading visible on the surface, will get asked 'did your mum make that' in an incivil tone - I don't think that the pure-bling pure-intellectual-property industries have much to worry about.
Look around you for things made out of hard plastic which are three-dimensional: Tupperware boxes, tags on key rings, your mouse, your headphones, the fastenings on your luggage, your hairbrush, your toothbrush, the cap on most bottles you own.
For each plastic component of every such thing, there has in the past existed a pair of objects made out of very hard steel in the exact shape of the inner and outer faces of the component. Someone bought a block of quite expensive steel, from some company you've never heard of like Virat Special Steel, and programmed a machining centre - a box the size of a small shed and costing somewhere between what a house in a back-alley in Manchester cost in 1980 and what a two-bedroom house in central London costs now - to mill out the relevant shapes with artfully-designed rotating tungsten carbide knives designed by models in which the very hard steel is a fluid through which the knife cuts like the keel of a boat; the cutters come from companies like Iscar, who I'd heard of only in the business pages where it was praised as a sensible acquisition for Warren Buffett, without a mention of sintering or of titanium-nitride vapour deposited coating. The cutters cost a hundred pounds and last a couple of days.
They cut awkwardly-shaped holes with an EDM machine, which uses electricity and graphite electrodes to push an electric arc through even obdurate materials; the machine costs somewhere between a new SUV and a new Ferrari. Then they installed the pair of moulds in an injection-moulding machine, which is fairly cheap in comparison to the machinery that made the moulds, which heats them, runs molten plastic between them, presses them together to stamp the plastic to shape, pulls them apart to let the part fall out, and repeats, until either they'd made enough parts or the mould wore out. The mould wears out: get another block of die steel from Virat and make another mould.
This is a capital-intensive industry, and it's an industry in which the minimum sensible quantity of a product is the number that can be made in the lifespan of a single mould; the mould plus that quantity of output costs about a year's disposable income of a reasonably well-paid software person and fills a shed.
The additive technologies, of which RapRep is the most visible, let you make things out of plastic which vaguely resemble the products of injection-moulding - the finish is generally dreadful - without having to deal with the capital cost; if you'd wanted to make things out of wood, or of textile, or of cast metal, or even of machined metal or machined plastic, the capital cost for the tools has been small enough that they can be afforded out of disposable income.
For the last twenty years, software people have always owned the means of production - the computer they use at work probably belongs to their employer, but they have a (generally better) computer at home, which can do anything the computer at work can do, and which cost somewhere between a cheap and an extravagant holiday. This gives a self-image that they're not employees, in the Marxist sense that an employee is a person that a person with (injection-moulding-machine shaped) capital is obliged to get to work for him in order to get an income stream out of the capital. RapRep is giving them the possibility of making a wider range of real things while remaining not employees, either of a person with capital, or of the large expensive machine which they personally own but are obliged to operate in order to pay the interest on the loan that bought the machine.
I think what we're seeing is the reactions of computer people to noticing the vast ecosystem of manufacturing, and thinking that what manufacturing needs is a Version 2.0 because that's so often the popular answer, and indeed fairly often the right answer, to computer problems. But a cheap slow computer gets the same answer as an expensive fast computer, it just takes longer; a cheap machining tool doesn't give you products with the same surface quality as an expensive machining tool.
The person turning up with the raprep'd mobile-phone cover, the colours not quite matching Hello Kitty's official colours and the pattern of the plastic threading visible on the surface, will get asked 'did your mum make that' in an incivil tone - I don't think that the pure-bling pure-intellectual-property industries have much to worry about.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2008-07-20 10:16 am (UTC)no subject
Date: 2008-07-20 08:25 pm (UTC)The solution to it being hard to get spare parts out to the back of Angolan beyond is roads rather than fragile part-making machines requiring imported lead-screws and circuit boards full of hard-to-procure components.
I don't have any idea what would be enabled in rural Uganda by a source of small numbers of custom plastic components.
no subject
Date: 2008-07-20 11:50 am (UTC)Years ago, consumer goods were made with metal bodies, not plastic ones, and they worked fine. There's no reason why a rapid-reproduction machine couldn't work with sheet metal, cutting origami shapes and folding them up to be riveted shut. They'd just look "uncool".
no subject
Date: 2008-07-20 08:43 pm (UTC)Cutting sheet metal arbitrarily and automatically is a solved problem with XY tables and either plasma or waterjet cutters; but folding sheet metal still seems a very manual task - you get CNC press brakes, but I can't find self-loading ones. The issue there is that the machine that can casually bend sheet metal can't itself be made of the thin sheet metal, the machines I see on YouTube have heavy dies and enormous forged-and-surface-ground hydraulic components, so replication is clearly impossible.
no subject
Date: 2008-07-20 08:58 pm (UTC)no subject
Date: 2008-07-20 09:25 pm (UTC)Net shape or near net shape casting is done, but where you're dealing with materials that have better properties than plastic, you're doing it for applications that have tougher requirements than plastic can meet: else why wouldn't you just use injection-moulded plastic? So they do tend to need a second forging and/or grinding operation to bring them the last milli-inch to the design dimensions.
no subject
Date: 2008-07-20 02:04 pm (UTC)no subject
Date: 2008-07-20 02:28 pm (UTC)no subject
Date: 2008-07-21 07:20 am (UTC)1) I can imagine exactly the same thing being said about Sinclair's MK14 thirty years ago in comparison with, say, a VAX-11/780. The point is that aspiring computer hobbyists couldn't dream of affording a VAX, but they could get hold of a MK14, and then try to improve it to make it do something useful. (A friend of mine controlled his homebuilt radio telescope with one.)
2) Look at the two door handles in my recent post - whilst there are intrinsic limits in the technology that RepRap is using, it's undeniably getting better.
no subject
Date: 2008-07-21 08:48 pm (UTC)Late note on rapid prototyping
Date: 2008-07-25 01:13 am (UTC)no subject
Date: 2008-07-20 02:25 pm (UTC)The word on why and when this technology makes sense
Date: 2008-07-25 02:07 am (UTC)