Ask HN: Small-batch builds of smaller phyiscal objects
I'd guess there's gotta be some services (in China?) that got this need covered. Anyone has any experience with this sort of thing? Any tried and tested companies?
Thanks!
I'd guess there's gotta be some services (in China?) that got this need covered. Anyone has any experience with this sort of thing? Any tried and tested companies?
Thanks!
Typically, machine shops specialize in some form of manufacturing type (milling / lathing metal, for example) or specific niche (metal stamping). If your parts require different manufacturing processes, you may potentially need multiple shops to get that done.
I found that Chinese machine shops are going to be both cheaper (about 8x in my experience) and faster lead time than North American shops. The exception will be that if you need something highly specialized (e.g. super tight tolerances), you may find it easier to get it done locally. I'll assume you'll be going with a Chinese place below. For finding Chinese places, I use Alibaba with MOQ set to 1.
Machine shops are always trying to expand and get bigger contracts with bigger customers. They dislike low order quantities and low volumes, and I found that I need to continuously cycle machine shops every couple of years as they move to larger customers. The other thing is that with manufacturing, there's a setup cost (flat cost per batch) plus a per-part cost. At single and low double-digit volumes, setup cost is going to be a lot. Make your part easy to manufacture - easy to clamp, made from readily available stock, use more readily available machining (e.g. 3 axis and not 5 axis, straight lines doable with a manual mill vs CNC), reduce manufacturing steps.
I found that cheaper machine shops (say, 2x cheaper than others) will tend to have a lower ability to understand what I'm trying to get done (will take a lot more & simpler communication) as well as a higher probability of the part being out of spec.
The more proficient you are at knowing what you want (obsessively defined specs and tolerances on everything you can think of) will get you much cheaper results. The less you spec out, the more they need to think about your design and your needs, and that's quite expensive from their side. My process is to have a drawing PDF that I send to 10 manufacturers for quotes, get 5 quotes back and select from those.
Say, if I have a part like [1], what's the best way to get it done? I'm now leaning towards laser cutting the flat bottom (the geometry allows for that) and then welding the top "clutch" after getting it lathed. Is there a better/cleaner/cheaper way to do it? The part should be metal.
I'm trying to figure out what the application is here. It looks like you want to attach something to a shaft, and trying to use a collet with a set screw. Why not put the set screw in your bottom part? Or perhaps use a shaft key? What kind of forces and what kind of load is this thing taking?
But if I had to answer your direct question: spec some screw holes, welding has downsides for what I think the application is (cost, axis alignment).
Force-wise - about a pound of load per "arm", give or take, not much really.
Re: screw holes - noted, thanks.
[1] https://www.amazon.com/GRILLJOB-Stainless-Rotisserie-1-Pair-...
[2] https://www.amazon.com/-/en/25437-R%C3%B6sle-High-Quality-Ba...
I also realized that my original suggestion of "set screw on the bottom part" is also probably hard to manufacture.
China sells tons of products they cloned from something pre-existing. I know because I own some of it and I've seen lots of western companies making use of it too. Here are a couple of examples of products from Amazon that existed long before China started producing clones.
https://www.amazon.com/BIG-RED-TAM82012-Hydraulic-Carrying/d...
https://www.amazon.com/BILT-HARD-Vertical-4-Stroke-Compliant...
All that being said, my best advice is to refine your design locally first, before farming out mass prooduction. Lots of local machine shops can fabricate simple custom metal parts --- they exist just for this purpose.
The cost angle rules out local machine shops as they are (really) expensive. Hence the quest for the online services.
In other words, cheap typically means automated mass production.
Otherwise you're literally just fucking around with your design software for your own personal fun, with (clearly) zero idea of how physical objects are actually made.
I suggest you google "machine shop online"
They have a fairly powerful platform -- instant quoting from uploaded design files, 2D/3D views, DFM feedback, good turnaround times, etc -- and the pricing isn't half bad for prototyping... but might not be as good for large-scale runs compared to bigger shops where you can nail down volume pricing discounts.
Another option is to just learn to do it your self. Many of the maker spaces will have have classes on metalworking, welding, aluminum cutting, CNCs, laser cutters. I also TA as a 2.5 day weekend workshop on the central coast that will teach you to woodworking and welding.
Let me know if you have any other questions.
chineese sheetmetal/cnc milling/3d printing (including metal): https://www.pcbway.com/
But if you can find a local fabricator that is probably best as they can help you understand if you design is manufacturable as is or if there are some simple tweaks that would make it cheaper/easier.
Galvanized steel flashings from Home Depot etc. can be hand formed with tin snips, a hammer, and a block of wood. Or you could buy a cheap brake from Harbor Freight, Amazon, etc. Same with a drill press.
Doing it yourself will speed up development and eliminate most if not all of the need for CAD drawings.
Not to mention it will avoid the mental overhead that your question embodies.
Or not, good luck.
I have, basically, the design of the whole thing figured out. I can 3D print it from resin, but that's pointless as it's a load-bearing gadget that needs to withstand high temps. Just need to fabricate it.
Got most of that figured out as well, pretty much all parts can be laser-cut from a sheet metal, then welded together. One bit I don't know is how to make something like this - https://i.imgur.com/RsambTE.png - 15mm long, 7mm in diameter. If you have any ideas, I'm all ears.
So I guess your local machine shop is another option.
Which reminds me that your local welder can build you a custom grill.
Whether they will is the real question. A lot of shops in my area had some rules around sizes and quantities.
It’s been a bit over a decade but I used this to make small runs of metal parts for a store I worked at. Their basic CAD is good for simply parts and you can easily adjust settings to find the price vs material/finish goals you’re after.
all you need to do is look for 'machinist' or 'tool and die' or 'metal fabricator' and make sure its a place that takes on small random work and not some high margin specialized manufacturer. probably call them and get used to rude dismissmals.
show up and shop and ask for the guy in charge. for this kind of work its often done under the table, so offer to pay cash. its very likely that the floor manager or a senior fabricator is going to do this in their spare time and not even put it on the books.
you're gonna pay too much. probably a couple to a few hundred depending on complexity.
you also sometimes ask a metal supplier. they have relationships with these kind of people. they often aren't that friendly though
- OSHCut for sheet metal cutting and bending.
- 3DHubs (very very recently became Protolabs Network) for aluminum machined parts.
In general, the task you're now on could be considered "design for manufacture" - you have the basic concept, but now need to figure out how to go from general design to manufactured product.
Design for manufacture includes keeping in mind the process/technology that will be used to produce the part: - you can do some amazing things with laser cut and bent metal, but you're limited to a single thickness per component; - lathes are extremely precise and great for making round/concentric items, but the part is limited to rotational symmetry (barring fancy machines and/or moving the part to a mill for further work); - mills are great for all sorts of things but really shine in "2.5d" work - where you have different profiles milled down to different depths, but from a single direction. Mills also can't handle overhangs (without a 5-axis, multiple setups or very special tooling) or features in different directions without multiple setups (which add time/complexity/cost/risk of misalignment of features).
For one-off or small run products, you can get away with being very inefficient in the process - starting with a piece of stock and milling 90% of it away for instance.
For the one part I saw you post a picture of, that could be done through either SLS or similar metal 3D printing (as someone else said), or one of the "official" ways to make a gear is to cut it with a custom single tooth cutter - where you run a special cutter along the length of the gear, rotate the gear one tooth, and repeat, cutting away the space between the teeth each time. Clickspring on Youtube has a bunch of videos covering the topic. I don't know who could do that as a service though.
If you post more pictures / sketches of the project, people here might be able to give better guidance for specific parts and operations.
Any local machine shop?
I remember watching a video a while ago where they were talking about how it's better/cheaper/easier to buy a part out of a catalog than custom machine whatever you are needing because chances are really, really high you don't actually need a custom part. IIRC they were talking about bearings but I'd imagine gears would be the same.
In this case, yes, any local machine shop could theoretically get or make the gear cutters, get a piece of the stock the correct diameter (/turn down from larger), mount it in a mill with the stock in a rotary fixture of some sort, then cut the teeth. Again, Clickspring has a bunch of videos on the process, including different ways to make the cutter. All that being said, you'd have to be on very good terms with your local machine shop folks, or be willing to pay for all of that new tooling and setup stuff, to get any random shop to make the part.
Going with someone like SDPSI might be better for this specific part, as it's in the realm of what they do.
Could also look at starting with "timing pulley stock" - it's basically the part that was shown elsewhere, but without the internal keyway cut. https://www.torontogear.com/bar-stock/ - starting with _that_ - yes, any local machine shop should be able to cut it to length, mount it in a mill and mill out the slot (with a small enough endmill).
One more recommendation here, based on what Uncle said about buying stock - and this is a part of design for manufacture in general - is to design around standard parts. Look at SDPSI/Toronto Gear for common gear specs (IIRC "module" is the main property defining a gear and how they will mesh with other gears - so stick with one module size), pick one and change your design to use those off the shelf parts.
With regards to the specifics, here's one of the parts - https://i.imgur.com/1oqjQ7r.png - basically a flat (bottom) piece that can be laser-cut and a cylindrical clutch-like bit with a cutout for a "flat-leaf" bolt. The idea is to pass a rotating axis through the central hole and to affix the part to it to have it rotate with the axis. Needs to be all metal as it's a high-temp (BBQ) setup.
That being said - to make that part, if concentricity isn't _critical_, you could easily split that into two sections. As you said, the bottom flat section would be laser cut (/water jet) out of sheet material; you could add some mounting holes around the central axis, then threaded holes in the rod mounting boss (or through holes for very long bolts and nuts, but that's better for prototyping with limited tapping/etc. capabilities).
That's exactly it. Well-deduced. Secondary rods are just skewers.
> If so, be careful about your gear ratios.
Yep, got that covered. The gear ratio is 16/11.
First, once you get several projects under your belt, you should start integrating "design for manufacture" into your main design process - that way you don't waste time designing some custom part, then searching for that specific item / someone to make it, when with some minor adjustments and off the shelf part would be _significantly_ cheaper and readily available.
Second, for designing with available parts, I love McMaster-Carr - they have great specs and drawings for just about everything, and 3D models you can import into most CAD software for most parts as well. They are more expensive than other sources for most individual parts, but the confidence you have with ordering from them is great, and for one-offs and prototyping they're invaluable.
They appear to be between versions right now, but to their credit, they are the only KS project I've ever backed that has shipped a high quality product on time and on budget. I anticipate big things from this team.
They were relatively cheap and fast (I haven’t done any prototyping in five years though)
They have about a 10% full-rejection rate; so order at least 10% more than you need
Also sometimes they just mess up and get the part out of spec. Then just complain and they'll make it right
After getting more funding we switched to US based rapid proto shops that charge 3-4x, but have a better experience.
(DISCLAIMER: I only spoke to the ME guys talking about SunPe. I never used them myself)
For machined/moulded parts there's shapeways.com (I haven't used them beyond quoting).
I'm sure there's plenty of alternatives for both too, if you search 'x vs' or whatever.
They do stainless from .048" to .120" thick: https://www.fabworks.com/resources/materials/stainless-steel
Have you used them a lot? Curious what your experience with them is if you're willing to share.
I have made several quotes for different personal projects, but those projects have all been tabled for the moment.
Here’s the announcement post on the main social forum for the robotics program: https://www.chiefdelphi.com/t/introducing-fabworks-fast-affo...
Otherwise yeah find a local machinist and if you're only doing a single one-off it shouldn't be too spendy.
An expensive machining job that finds problems early will save lots of money.