Hacker News Reader

Comment by bognition | Hacker News Reader

While the Matty's inventions may seem silly and frivolous, they signal something huge. He prototyped 350 physical inventions in 3 years. Thats roughly 3 days per item.

When compared to how hard this was a decade ago, the cost to build physical prototypes has effectively fallen to 0. This is both in terms of the money and time. I really believe we'll see a renaissance of physical goods in the next decade.

Previously you'd need to find a factory and contract with them to manufacture a minimum number of items. It would cost several thousand dollars at least and take weeks if not months to get your product.

Once you have the physical item in your hand, you can start the testing process which feeds into the next iteration cycle.

View on HN

>> it would cost several thousand dollars at least and take weeks if not months to get your product.

Or you made it out of wood. Or metal. Or clay. Or carved it by hand out of a block of plastic. Building and testing three or four versions of a tool is just a normal day for a shop carpenter or metal worker. Many of this guy's "inventions" are just jigs for holding things. Youtube is full of metal shops fabricating and testing jigs for strange use cases.

I don’t know what carpenters you know who make 3 or 4 jigs a day, but that does not match my professional experience in carpentry at all. Maybe you get this impression from YouTube but that is very different from everyday work, it is theater.

I used to work in film/tv/live entertainment. The shops pumping out scenery and stage devices are probably not normal, but every day was some sort of innovation. If you are building a different product every day, creating new jigs to make those products is a daily task. A new bit of metal to secure a bit of equipment to camera rig, aka a "camera cluster", was a very normal daily thing.

Yeah but Hollywood carpenters are in another class. They hire Han Solo to build doors in Hollywood.

Luke looks at door: "What a piece of junk!"

I have a friend who was one. He'd get a kick from that comment.

This is what he does in his spare time, these days: https://www.youtube.com/@jimkehoe6495

That looks fun, I especially like that he put a NEMA stepper in rather than a servo. I do the same thing for planes, where weight is a consideration (no turret there, though), and having no weight constraint is another level of fun.

I just want you to know that at least one person understood the reference here.

(The story is that Harrison Ford was “discovered” while doing odd jobs and minor carpentry for some high status person in Hollywood)

That is very cool. But yes, it is extremely unusual. That is, like, the opposite of what most businesses want, which is to standardize the tasks and labor. A cabinet shop has a set of jigs that they use over and over and over.

Ok but your experience is totally unrelated.

This thread is about making prototypes which are one time jigs. The reason businesses aren’t making 3-4 jigs a day isn’t because they can’t but because they don’t want to.

But if you’re prototyping, you are literally trying to create jigs.

And this sub-thread is whether you could create jigs before 3D printers existed and you definitely can. Of course, it’s great that now you have both options, speaking from someone with a woodworking shop and a 3D printer, but I wouldn’t say there was such a substantial innovation in physical prototyping.

The purpose of the movie set example was to illustrate that, if you want to, you can create constructions out of wood really quick, not that it’s normal for a business to do.

If anything, the biggest innovation in at-home prototyping was how cheap microprocessors got. Because building circuits is way more painful and expensive than just chucking in a microprocessor.

GP said 3 days per item, not 3 items per day.

Still, there is truth to their comment. It's at least become a lot more accessible to get the tools to do that sort of work with Amazon. 3d printing has also become more convenient and accessible. YouTube sheds a lot of light for a lot of people into how things are made. I think setting up an online store has probably become a lot easier, too. These things are big changes of the last 10-15 years or so.

I'm a partner in a plastics manufacturing facility (injection molding, thermoforming).

3D printing has been big for us. No, we don't make money directly from it, or not very much. And it's extremely expensive per unit.

But it allows us to design a part for injection molding, and then iterate it many times until we are happy with the functionality of it.

Once we commit to an injection mold, it might be $15k or $30k to get it made, and it'll take 6 months. But we know that the fit will be nearly perfect by the time those first shots come out. Then we can roll with making thousands or hundreds of thousands of parts per day at the efficiency of scale.

The way that 3D printing is used by hobbyists looks fairly different than the way it's used in industry. In industry, we're mostly still using traditional design-for-manufacturing principles for the prototype: for example, we're designing it so that it could be injection molded and ejected properly from a mold, even though the 3D print doesn't need that. That's because our intent is to merely confirm what we see in Solidworks in the real world, and to show a customer an idea in their hands in the shape and form it'd actually be produced at scale.

I'm not compelled by the idea that 3D printing is going to make a big difference by allowing very small runs of parts (except for prototyping in the manner above). That's because like everything, there is a power law when it comes to parts. Nearly all the revenue comes from medium and higher volume products. Sure, it's useful in niche applications, and for things like jigs, but as much as people love to talk about the power of long tails, it rarely is as good a fit to the real world as the power law.

But the idea that 3D printing is making a big difference in improving the design and innovation with those medium to higher volume products made with conventional manufacturing methods is absolutely true.

Why does it take so long to make molds? Has prototyping reduced the number of molds you have made? Has that decreased the wait time, or are more people making molds and so the wait time is now increased instead?

I've worked in the plastics industry. A lot depends on the kind of mold of course. A mold can be somewhat complex when all is said and done, and made from a combination of parts that have to be ordered, or materials that are difficult to work such as hardened steel. Producing multiple pieces that fit together correctly and aesthetically can be an art. Whether the mold actually works in your machine without problems depends on the quality of the mold maker.

And it could boil down to the fact that mold making is a mature industry with a lot of established supply chain and lead time expectations, and nobody is trying to rock that boat.

Of course you don't want to spend all of that time to get a mold that you can't use because you screwed up on your end and forgot to add a needed feature. The world has not moved away from the benefit of a physical object that you can hold in your hand and see that it's actually right. And the 3d printed prototype can also let you check other things while the mold is being made, such as the fit of things like circuit boards, buttons, and connectors.

Molds are surprisingly complicated. They weigh 1,000 pounds or more; some molds we have weigh 10,000 pounds. They're made of high grade tooling steel that must be annealed to the right hardness and ductility. At each injection, they are subject to pressures of 25,000 pounds per square inch, and this happens every ten seconds 24/7 for the duration of the run (which can be weeks or months). Over time, plastic cuts into the steel, like water on rock in a canyon. So it takes lots of milling time, because tool steel can't be cut fast, like wood or aluminum.

The molds also have to be machined to high tolerances. Take vents as an example. Vents allow the ambient air inside a mold to escape as plastic fills, but they must not allow the plastic to escape out the mold on the same route, so they must be machined to a tolerance of around 3 tenths of a thousandth of an inch. That in turn requires high precision CNC machining over fairly large spans, and with many subassemblies to make sure the two halves fit together just right.

Some kinds of molds can be completed faster. Usually smaller parts, or parts where a base size of a mold has already been completed and you are just using inserts. Or if the mold is not expected to last very long, it can be made out of weaker material, which also cuts faster.

I think there are some interesting software opportunities to improve mold build lead times. I suspect some of the design of the molds could be more procedural and automated than they are. That would help cut down the engineering time. And some people are exploring 3D printed molds, which is a cool idea, but currently they are very weak and not useful for production because nobody yet has figured out a way to 3D print high grade tool steel.

3D printing has actually likely increased the number of mold designs we're doing, because it makes it easier to innovate and iterate, and easier for a customer to feel comfortable with a design and move forward with it. It reduces customer utility and engineering risk, and so cuts the total cost of product development.

> I'm not compelled by the idea that 3D printing is going to make a big difference by allowing very small runs of parts

On the contrary, that future has been here for years. I print a ton of stuff for me (I'm printing one now!), they aren't prototypes (as in, I won't eventually make them "from something else"), they are the final part.

3D printing has enabled me to have custom parts, to my exact specifications, in record time, for pennies. It's amazing. Of course it's not going to make a difference to your revenue, the entire point is that, with a 3D printer, I don't have to come to you at all.

Don't get me wrong - I love 3D printing! I use it myself to make enclosures for electronics projects, or to replace parts that have broken on my old sailboat, which happens often. Notably, I don't turn to my business to injection mold these parts, because it makes no sense to make a mold for one part! Without 3D printing, I'd probably have someone mill custom parts, which is a lot more expensive.

But I'm a realist. Look at the total poundage of all resin sales to the 3D print market. It would be generous to call it a rounding error in comparison to conventional resin sales. People like you or me are not the norm. The market for conventionally produced high volume parts is enormous compared to hobbyists 3D printing stuff at home.

That's because most people consume the same small set of products repeatedly, and they care about cost, and some of my machines produce at a rate of ten parts per second with part weights 90% less than what's feasible with 3D printing. It's just the way markets work: most volume is accounted for in a small percentage of products, and engineers will always be able to get much more efficiency out of a specialized machine designed for a small number of high volume items than a general purpose machine that can produce anything.

Of course, no argument there. But the fact that printing presses produce most of the world's printed paper doesn't mean inkjet isn't useful.

I guess it depends on how "plug and play" printers can become, right now they are pretty fiddly. Still, I think their growth is pretty healthy, even if not every home has one.

cost to build physical prototypes has effectively fallen to 0

ignoring the initial outlay of 3d printing and materials

Previously you'd need to find a factory and contract with them to manufacture a minimum number of items

That's never how initial product prototypes have been made. His first versions may be less janky than historic prototyping, but I think we're risking rewriting history to make 3d printing more grandiose than it actually is.

10 years ago I was at a company prototyping a new widget. We had 3 day turnaround from the local 3d printer for something the size of a cellphone, it cost over $1000, and the quality was garbage.

Once we did basic usability testing with this, we went to a low-volume manufacturing run using silicone molds in Shenzhen and FedExed back, turnaround time about 10 days. Price $15k+. Once we were happy with this we did alu molds in Shenzhen for a batch of 500, and you don't want to know the price.

Today I can get 3 hour turnaround for items the same quality as the silicone molds, just by getting our own liquid printer for about $1k. This really does change things for prototyping.

Also take note of the companies what have sprung up to supply low volume manufacturing at good prices to aid in prototyping and access to specialized machinery.

https://jlcpcb.com https://sendcutsend.com https://www.pcbway.com https://www.knifeprint.com

I use JLCPCB for small runs of fabrication and assembly of PCBs (I just got ten the other day), and love them. I have to wait a few weeks, but the quality is fantastic, and it saves me from soldering all the small SMD parts, for very cheap.

> the cost to build physical prototypes has effectively fallen to 0.

That's a bit of an exaggeration. The cost to build prototypes of soft plastic objects in unusual shapes has significantly diminished. The cost to build electronic prototypes has decreased only marginally, mechatronics, robotics and most types of actuation are still a big boys club etc.

While 3D printers have been very useful for certain types of prototyping, I'd have to say that overall the excitement/hype around 3D printing you saw 10 or so years ago has significantly diminished.

Yes it's definitelyoved out of the hype curve into the are where it's actually useful. Which means it's actually doing stuff people aren't talking about yet or don't need to show off. I use my printers all the time and I'm Im only slightly more hyped now then when I figure out a new web framework, or a cool new way of thinking I about pandas.

Just cause it's not hype doesn't mean it's not useful. How sexy is rail freight? Power grids. I don't see tons of hypes on my towns water supply.

As long as an electronic prototype can be made digital, things improved a lot with wide availability of Arduino, RPi, and ESP32 boards, and very accessible development tools for them.

Many analog things can be modeled and evaluated using a fast enough CPU, which are cheaply available; see software-defined radios, for instance.

> The cost to build electronic prototypes has decreased only marginally, mechatronics, robotics and most types of actuation are still a big boys club etc.

I strongly disagree. Not only have the non-recurring expenses in EE cratered thanks to extensive part libraries and reference designs but the cost of fab & assembly is easily a tenth of what it was a decade or two ago thanks to Chinese manufacturers. Open reference designs with design files are more common than ever, allowing anyone with a little EE knowledge to bootstrap even complex designs. The equipment necessary to solder BGAs at home have fallen to the hundreds of dollars, cheaper than a decent used Metcal.

6DOF robotic arms isn't even a big boys club anymore with several open source versions available off the shelf with fast turn around part availability from McMaster/Misumi and really cheap high tech parts like linear encoders, motors, stages, controllers, etc from Alibaba/express. All of these vendors allow hobbyists to prototype full blown industrial automation, let alone individual products. Open source 3d printers, CNCs, pick and places, and so on abound with plenty of quality open source firmware. People are even starting to tackle linear motors for high precision actuators, the kind you'd find in semiconductor fabs.

Rapid prototyping (both additive 3d printing and subtractive machining) has been readily available for decades. What's "new" in the last 10 years is the accessibility and cost of it.

15 years ago I could place an order with a local rapid prototyping company at 3pm and have the part (3d printed) on my desk by lunchtime the next day for under £100 (or much less for very small parts). But those printers were tens if not hundreds of thousands of dollars. Now for well under 1k you can have a 3d printer yourself.

There are tools/technologies to make creating certain types of prototypes fast and cheap. I don't think you can generalize that to physical goods broadly. If I want to experiment with new battery designs, I'm not doing that at no cost in an afternoon.

Reply on news.ycombinator.com