The implications of consumer 3d printing

[leady]

I've always found it amusing that a good number of people have a view that complex, expensive and imprecise 3D printing will replace precision mass manufacturing in terms of cost and ubiquity.

[Tyndmyr]

*shrug* It's almost a fad in terms of tech. Probably, at least in part, because it demos really well. Looks hella impressive. Nobody wants to hear about technical problems, just possibilities.

[ijuin]

As has been said in this thread, printing won't be cheaper than mass production, but the lesser infrastructure requirement will make it attractive for one-off items or other small production runs in which there are not enough units to justify the cost of dedicated molds or dies.

On the consumer end, it will cost more for an end user to print an item than it would cost a large manufacturer per unit, but when the costs of marketing, distribution/delivery, and middleman and retailers get included, the retail cost can be an order of magnitude greater than the manufacturer's production cost, so the end user may still be saving over retail. For example it might cost $2 for a manufacturer to produce and retail for $10, but cost $7 to print.

[slinches]

There is also an advantage that 3D printing can produce parts that would be impossible (or at least insanely difficult) any other way. Things that have a bunch of internal passages that couldn't be drilled from the outside or that have an internal lattice structure for reduced weight and nearly identical strength are possible. Also, with small complex parts, production on a 3D printer can be cost competitive for reasonably large production runs (100s to 1000s of pieces) since you can print dozens at a time.

It really is a cool technology. It just isn't the Star Trek replicator-esque panacea that it's made out to be.

[Tyndmyr]

Well, those things are only impossible to produce in one piece. Such things exist all the time. Say, an oil filter. All kinds of crazy internal structures, you can't build that in one piece with traditional manufacturing...so they use multiple, and it's cheap and easy.

[Zamfir]

Also, with small complex parts, production on a 3D printer can be cost competitive for reasonably large production runs (100s to 1000s of pieces) since you can print dozens at a time.

At least in my experience, many powerful computer-controlled techniques are often better suited to small to medium size production runs, than for true one-off designs (prototypes excepted).

People have to spend time thinking up the complex shape, analysing it, drawing it, checking it for errors, perhaps make a test model. And the people doing the assembly or installation have to learn the intricacies as well. For truly one-off designs, you want to minimize that complexity if you can get away with it. Also for budgeting and planning purposes: at the decision point, you can't quite take the potential savings from complexity into account.

Which often means that the really complex parts are off-the-shelf, hammered into the design even if they are not the perfect match. While the one-off parts are as straightforward and simple as possible, even if it's not optimal. Straight lines, right angles, oversized whenever possible.

I see something similar with friends with 3d printers. They sometimes design a purpose-fit part to fix some problem at home, but that's more a hobby to use the printer. In the time it takes to design and draw the part, you could typically cobble together some board and screws or whatever else fits.

[ijuin]

That works when engineering tolerances are loose, but when you need aerospace-grade precision, you will need to get an exact fit rather than just taking a grinder to pre-existing parts.

[Tyndmyr]

Aerospace-grade precision is likely not generally the result of 3d printing. It's pretty normal for me to do some hand finishing to make pieces fit. Now, higher end printers do exist, and they are getting better, but even consistency wise, you have to be fairly careful with 3d printing. Temperature, humidity, etc can seriously bork a print job, and even when you control for that, you can get significant variance.

[Zamfir]

That works when engineering tolerances are loose, but when you need aerospace-grade precision, you will need to get an exact fit rather than just taking a grinder to pre-existing parts

Don't worry, modifications and repairs of aircraft employ all the regular tools of sheet metal working. Blade saws, grinders, even nibblers. You don't need super tolerances everywhere, as long as you have a good grip on what tolerances you do need, and how to guarantee them. That might well be 'qualified worker with a grinder and a caliper'.

[slinches]

There is a huge difference between the commercial and consumer grade printers when it comes to accuracy. The parts I was talking about are actually for aircraft turbine engines and the weather has no effect on the machines we use (fully purged enclosures and meticulous base material handling practices). Tolerances in the thousandths of an inch are the norm and these machines can meet those requirements. Usually, the only finish machining we have to do is to remove build supports and clean up the interfaces.

[wumpus]

Aerospace-grade precision is likely not generally the result of 3d printing. It's pretty normal for me to do some hand finishing to make pieces fit. Now, higher end printers do exist, and they are getting better, but even consistency wise, you have to be fairly careful with 3d printing. Temperature, humidity, etc can seriously bork a print job, and even when you control for that, you can get significant variance.

The first place I've seen a full-blown computerized milling machine (block of aluminum goes in, complex turbine part comes out) was making parts for aerospace. I suspect they will stick with subtractive methods due to initial investment, strength, and pure accuracy. Largely in that order (assuming that accuracy can be increased to at least aerospace tolerance). I won't claim that the output of said mill might not have needed adjustment and polish, merely that what looked to my layman's eyes was a complete part and that nobody was directly controlling the mill while it did its thing.

Does anybody know about success in 3d-printing molds for low/mid/high volume production? My understanding is that the high volume stuff will likely use subtractive methods (and steel molds), but that lower volume might be an option (although sintered metal might be required).

[Tyndmyr]

Aerospace-grade precision is likely not generally the result of 3d printing. It's pretty normal for me to do some hand finishing to make pieces fit. Now, higher end printers do exist, and they are getting better, but even consistency wise, you have to be fairly careful with 3d printing. Temperature, humidity, etc can seriously bork a print job, and even when you control for that, you can get significant variance.

The first place I've seen a full-blown computerized milling machine (block of aluminum goes in, complex turbine part comes out) was making parts for aerospace. I suspect they will stick with subtractive methods due to initial investment, strength, and pure accuracy. Largely in that order (assuming that accuracy can be increased to at least aerospace tolerance). I won't claim that the output of said mill might not have needed adjustment and polish, merely that what looked to my layman's eyes was a complete part and that nobody was directly controlling the mill while it did its thing.

Does anybody know about success in 3d-printing molds for low/mid/high volume production? My understanding is that the high volume stuff will likely use subtractive methods (and steel molds), but that lower volume might be an option (although sintered metal might be required).

CNC machines are very different from 3d printing. They are indeed quite good...the additive stuff introduces a lot of additional complications.

And of course, there are a wide range of options...better accuracy can be had for a great deal more money, and (usually) a slower print time. Laser sintering is a good deal different from extruding ABS, for instance.

I haven't messed with 3d printing molds yet, oddly enough. My injection molder needs maint at the moment, but it DOES work to 3d print the original, and make a mold off that more traditionally. I 'spose you could print the mold, but that's usually going to be a lot more printed mass, and a larger size object means you have some tradeoffs due to the required pressure.

[slinches]

Actually, aerospace has been doing prototype parts (and even limited production runs) using stereolithography and soft tooling for decades. In that process, they print the original and use that to create the soft tool which is in turn used to make plastic or wax molds to be used in standard casting processes. There are some trade-offs in accuracy and consistency/tool longevity, but you can get fully finished cast metal parts in weeks rather than months or years.