We often get asked 'at what point does it become cheaper to mass produce and item rather than 3D print it? at which point we have to ask what do you want to make, in what material and to what level of complexity or customization? To approach a need and simply switching method of manufacturing misses the power of 3D printing.
A recent blog post by 3sourceful compared the cost of manufacturing two items using Shapeways 3D printing and Protomold to make injection molded parts.
"In this case, we prices out two different parts. One, a very small bracket (~1cm^3) and one a larger jig (~50 cm^3). To compare, we obtained quotes from Shapeways and Protomold. And for simplicity, we just assume the cheapest material from each. We then plotted out the total cost of production for different quantities. As we would expect, the tooling costs of the molds resulted in 3D printing being cheaper at lower quantities in both cases. But, in the case of the larger part, the cost of the 3D printing material meant that over 100 units, Protomold became the cheaper solution. Where, for the smaller part, 3D printing was cost effective over 1000 units."
This is great for a simple equation for comparing the cost of a simple part required in bulk like their bracket, but if there is any level of complexity in the part or strict tolerances, the price to injection mold is likely to quickly increase, where as the price to 3D print would likely stay the same or may actually be reduced if the complexity is in the form of meshed or perforated features. The larger item might not have taken advantage of the density discount on Shapeways that can dramatically reduce the cost of large parts.
The comparison does not take into account the upfront investment required along with cost to warehouse and distribute the injection molded parts with the liability of predicting sales and holding inventory that may not sell. Customization and/or fast iteration is also an incredibly powerful advantage of 3D printing not so easily posible with injection molded parts. If you want to make 1000 components that are very similar but not the same, the cost to 3D print remains the same where as with injection molding you will need to invest in 1000 different molds, or at the very least, 1000 different mold inserts which would then need to be manually changed out after each part is manufactured.
Same too with fast iteration, if you want to modify your design in any way to optimize your design there is no additional cost or delay with 3D printing where as with injection molding you would need to retool in most instances, adding greater cost that would need to then be amortized across the sale of your product.
We have heard of examples of people making 3D printed molds to then take to injection molding small runs. Depending on the material you may need to treat the surface to ensure smoothness/mold release and to seal any porosity.
There's also an issue in that ProtoMold's tooling is aluminium and works within very specific (albeit ever growing) constraints, so what you have here is a comparison between 3D printing and non-traditional-shorter-run-moulding-processes, not your toolsteel-run-a-billion-parts-off-tooling. In the 1,000 part range, that's not going to make much difference, I know.
The comment about "upfront investment required" doesn't really scan out either because Protolabs/Protomold store your tool, and you order more parts as and when you need them. They load it up in the moulder, shoot some more plastic and ship you the parts. There's no massive warehousing involved or investment in the tool. It's paid for with that first order and spread out "per part".
On a personal note, the whole "for simplicity, we just assume the cheapest material from each" worries the hell out of me. If you're not using the same material, then you've got two entirely different parts, made in entirely different ways, with entirely different performance characteristics.
They just happen to be the same shape. Which is about as useless for a comparison as a chocolate Thermos flask.
There is a problem with Shapeways in that the only business models that work are very small parts or extreme specialty parts. It would be great if Shapeways partnered with a firm that would offer injection molding based on the print models people upload to Shapeways. That would offer a production pathway to people looking to create a general purpose product like a toy. Currently scaling up requires starting all over again with a new manufacturer that does molding.
IMHO the problem is that you have to design for moulding or design for 3D printing. Given completely clean CAD for a 3d printable part is only the first step to get a set of CAD for the plastic moulding. (Having gone the other way moving low volume life expired plastic tooling to 3D print the same btw is true that way around and in some cases cannot be done due to wall thicknesses and accuracy requirements).
3D printing masters for lost wax casting and for white metal or pewter casting works for a lot of shapes but again some care is needed so it will come out of the mould. Resin is somewhat easier than plastics as beyond the obvious "can you get it out of the mould" concerns the wall thicknesses and the like are not disssimilar.
I'm finding 3D print is actually cheaper than going to white metal/pewters for small parts, but metal has advantages. If the FUD part itself is very delicate you can make a lost wax brass master and cast that in pewter or when you need weight low down then metal castings give you weight, which is one reason I'm still using metal for battery boxes on some of our model railway kits.
Interesting numbers. Ones that need ramming down the throat of all the "3D print piracy panic" press people and lobbyists.