Ask Hackaday: Is There a Common Mechanical Parts Library?

Like many stories, this one started on the roof. This particular roof is located in Michigan and keeps the rain and snow off of the i3Detroit hackerspace. Being an old industrial building, things up on the roof can start getting creaky, and when an almighty screech started coming from one of the rooftop vents as it swiveled in the wind, Nate, one of the group’s coordinators, knew it was time to do something about it.

Previous attempts to silence the banshee with the usual libations had failed, so Nate climbed up to effect a proper repair with real bearings. He dug into the unit, measured for the bearing, and came down to order the correct items. That’s when it struck him: How many should I order? After all, bearings are useful devices, not just to repair a wonky vent but especially handy in a hackerspace, where they can be put to all sorts of uses. Would extra bearings be put to good use, or would they just sit on a shelf gathering dust?

That’s when Nate dropped us a line and asked a question that raises some interesting possibilities, and one which we couldn’t answer offhand: Is there a readily accessible online library of common mechanical parts?

Not All The Parts

My first reaction to Nate’s question was, “Sure there is — it’s called the McMaster-Carr catalog.” The voluminous gold and green dead-tree version and its online equivalent are valuable references for anything and everything mechanical. But McMaster’s stock really represents more of the universe of parts as opposed to a well-defined list of commonly specified parts. If a part exists, it’s likely in the catalog, but just because it exists doesn’t mean it’s generally useful. So much for the wisecrack.

It turns out that Nate has something else far more interesting in mind, and it’s based more on his experiences with electronics parts ordering. He’s thinking about something along the lines of Octopart’s Common Parts Library, Macrofab’s House Parts, or Seeed Studio’s Open Parts Library. These repositories allow designers to see what components are currently being used for various applications. The aim is to ease the design process by limiting choices to just a few widely available parts. Start there, and if your needs aren’t met with some of them, branch out, spending extra time on only the parts that actually call for it.

Say you’re working on a circuit and need an op-amp. Searching through every op-amp available can be a daunting task. Having a concise list of the devices broken into a few general categories helps to ensure you don’t get bogged down with each part choice when your design constraints aren’t very particular. Pick your bandwidth, pick your package, and move on with things. In the case of Contract Manufacturers, the added benefit is that you won’t need to source the parts for the production run and their familiarity with the components helps avoid problems when the pick and place machines are fired up.

Pick an op-amp, any op-amp. Source: Octopart CPL

An equivalent library for mechanical parts would be a boon to all designers. The line between electronics and mechanical design has always been a blurry one at best, and with more and more products interacting with the real world through servos, steppers, and actuators, specifying mechanical components is just another part of the job for electronics designers. Making that process easier would be a win for both the designers and the suppliers that will get the business through quick and simple one-click ordering.

Where Do We Find the Common Mechanical Parts?

And yet, to the best of my knowledge, no such Common Mechanical Parts library exists. I could be wrong, of course, and we’d love to hear from the readers if they know of any such services. But we’d also like to brainstorm what just such a library would look like, and how it would be executed. For my money, I’d start with the McMaster-Carr catalog, since if they don’t have it, you don’t need it. One must assume that they have an API for their catalog; after all, Fusion 360 has a searchable catalog of parts so that you can drop CAD drawings for McMaster parts right into your mechanical designs. A little Googling around suggests that the API is private, though.

So what do you think? Would a Common Mechanical Parts Library be useful? Does one already exist? What would you want to see in terms of features from such a product? Sound off in the comments section below.

And thanks to Nate for the tip on this and the useful discussion.

45 thoughts on “Ask Hackaday: Is There a Common Mechanical Parts Library?

  1. Common to what?
    Rapid prototyping has different requirements than car clubs which are different from instrument builders.
    Without at least a few tasks in mind you’re gonna waste time / money on parts you don’t use. It’s why most of the time ‘electronics starter kits’ are a waste of money. You get 100 10pf caps and after a couple months of work you still have those original 100 but you’ve ordered another 100 2.2k resistors and a couple dozen LED assortments. The same happens with bolts and screws.
    Without a filter for the goals any list is going to become unwieldy.

    1. It is an interesting analysis yours, we all have a drawer with components decades old that came in a starter kit, but never used, but that is the purpose of starter kits, because you don’t know what you need when you start, that is something that you figure out with experience.

    2. I agree only to a point. In my job I need to design mechanisms trying to reuse as many “common” parts as possible; we contract the rest to machinists and other kinds of shops. But many times the parts that we needed in fact already exists: it’s common in some sort of industry that we aren’t aware of. Please note that I live in Argentina, where it’s very difficult to source things if they are only marginally uncommon. So, knowing *where* to look for a part would be extremely useful for me, narrowing down my search to something more sensible. For example, certain parts are common associated with house water piping, or wall insulation, or RC cars; other parts are bread-of-the-day in the food industry, or jewlery making, etc. Hey, even knowing that certain screw sizes are more common than others is very useful!
      So yes, “common to what” is a most relevant question, but instead diminishing the value of such a catalog, it points out possibly one of its best selling points if done well.

      1. It happened to me. Did a reseach for days, found no existing solution, started new design from ground up, did additional reseach for that, found out the whole thing exists in a different industry under a different name.
        Different industries with different history can have the same solutions, but one’s name can be weird and unnatural for another. Sometimes even a simple dictionary could be helpful. (* is having wet dream of Wikipedia articles being more detailed.)
        For example the same device can be called multiphase inverter, VFD, ESC, motion controller, motor driver or something else depending on who you ask.
        Bonus questions: why is BLDC called Direct Current, and what’s the difference between that and PMSM?

        Oh, how much time could I save if names were consistent.

    1. That’s completely unhelpful.

      It’s like saying “Digi-Key has electronics”. Right, but they have 8 million parts. Whereas the Common Parts Library suggests about a hundred that would be worthwhile to have on-hand, because they show up in a ton of hobbyist projects.

      McMaster likewise has a half a million parts. I’m not going to order ten of everything.

      But I would like to have ten rollerblade bearings on-hand, because they show up constantly in hobbyist projects from the internet. I know that one. What others don’t I know yet? Is there a list of those?

      1. Maybe we misunderstood the question being asked, but I don’t think Dan is asking what 100 parts should you have physically on hand. He’s asking if there is a library of common CAD models out there that can be used in design so you don’t have to scratch design off the shelf or standard parts.

        1. Does it make sense to separate the two? If you have to express ship the perfect parts for a project is that any better than using multiples of the mediocre parts you have on hand? All the big manufacturers standardize what they use because even if it’s not ideal, the bulk discount makes up for using more of them to get the correct performance. It also cuts down on tooling or process requirements. You can bend sheet metal in a few minutes to create the perfect bracket or wait a few hours to 3D print it. Likewise having 4 different screw sizes means you need to swap tool heads or parts coils which adds production time.

          1. So why does Tesla use different size bolts on the left and right tracks for the front seats in the Model S? Makes much more sense to use all the same size bolts to mount the seats to the floor, like the rest of the vehicle industry.

          2. “Does it make sense to separate the two?”

            Yes! Few of us ‘hackers’ will have the space or resources to purchase the whole library. Likewise our individual interests, needs, skills will mean that even among the smaller, common library of parts many will not apply to us as individuals.

            “… the bulk discount makes up for…”

            Totally N/A for hackers/hobbyists/diyers or whatever we each choose to call ourselves.

            “If you have to express ship the perfect parts for a project is that any better than using multiples of the mediocre parts you have on hand?”

            This absolutely depends on your goals. Do you want a one-off for yourself or are you going to keep producing these things long after your parts-on-hand have run out? Is this just for you or are you planning on open sourcing the design?

            In a one-off project for yourself finding a way to use parts from the junk you have on hand is a virtue. It’s an extra challenge to be proud of conquering. It saves money. And it keeps resources out of landfills.

            Expecting someone else or even your own future self to find an equivalent to that funny part you took out of an old lawnmower engine from 1952 and was only in production for a year… not a very good idea.

            “You can bend sheet metal in a few minutes to create the perfect bracket or wait a few hours to 3D print it.”

            True but if you are designing for an audience, one familiar with 3d printers and unlikely to have the shears, brakes and skills with metal working than you do… know and design for your audience.

          3. @Hairless Ape
            >>Yes! Few of us ‘hackers’ will have the space or resources to purchase the whole library.

            Definitely, but if there’s a parts library of the 10,000 most common parts from 747’s to arduino shields, doesn’t it make sense to have some of that list on hand? As others have mentioned not everywhere has the option to express ship something from shanghai or the nearest amazon depot. Obviously you’re going to focus on the parts useful to your projects.

            >> …many will not apply to us as individuals.
            That is the point I was dancing around. There are many parts that are very common across multiple industries or hobby niches but they don’t make sense for rennaisssance hobbyists to keep on hand or the ‘common’ parts might be prohibitively expensive.

            >>This absolutely depends on your goals
            I completely agree, but the target audience I see using this list most often are Hackerspaces trying to stock parts for a couple dozen people and people sending small runs to some board house. You may only need to build single digit quantities of a project but production time (price) adds up when you have to swap reels or tool heads. Your lawnmower example is a good one because sometimes parts are ‘common’ simply because that year model broke quickly and you can find them at the thrift store or recycler for scrap prices. How many HV projects start with ‘find and old CRT display to gut’?

            >> designing for an audience
            Absolutely, as I said, the tooling you have access too informs your design choices. I agree someone that enjoys 3D printing is unlikely to have a brake or shears, I think that’s a gross oversight though as you can buy or make both of them for under the same price as a good low end printer. Door hinges and some hardwood boards will produce nice bends in light gauge metal. Or a set of hand seamers for small enclosures. I understand space is a premium even if you have a 10,000sqf ‘space to work in, that doesn’t mean you need aircraft repair equipment to work on sheet metal.

            All of this goes back to my first point of ‘common to what?’.

          1. I don’t know of a list of hardware you should keep around. I know I roughly follow the formula usefulness + (available space)/(size) + (budget)/(cost).

            With unlimited budget and larger space it would be awesome to have every screw that mcmaster carries on hand. If you have a realistic budget and smaller space, maybe you just carry M3 through M8 pan heads.

            You can see why electronics assortments are common though. Cost and size are so small the other factors almost don’t matter, you just buy the resistor and capacitor assortment to have around. Coilcraft gives away assortments of their inductors for free.

      2. There’s no reason you can’t just make a ‘shop standard’.
        Keep M6 or 1/4″-20 bolts, and some deck screws in various lengths. Unless you’re building trebuchets or other big projects you’ll probably be fine. Another thing to consider is raw stock. 2×4’s, 1/2″ plywood and 1″ square tubing are pretty invaluable. There’s maybe a dozen things that are truly universal without specific goals in mind. The list also depends on what processes you have access to. If you have a plasma cutter or CNC router your options expand.
        Dan Gelbart has a great video for what he finds useful in prototyping.

        1. Yep, for a carpenter, it would be a 2×4, or a 4×8 sheet of plywood. For a welder, it would be (I don’t know I’m a lousy welder). For one friend’s Dad, it was slotted angle iron…
          But if it was on-hand/in-stock at the local makerspace, it will probably be the first choice, and that includes surplus materials (old ducting for sheet metal, PVC tubing scraps for structural, construction dumpster contents…

        2. Sure. A shop standard is useful to your shop. If your concerns end at your shop’s walls then that is good enough. Open source hardware however could benefit from the majority of it’s shops being on the SAME standard.

  2. Solidworks tries to do this with its toolbox of standard fasteners, dowel pins, ring clips, etc. but I have never liked using it. There is something about parts that are not tied to reality with a manufacturer part number. I prefer mcmaster models for screws, nuts and other off the shelf hardware since you know it is an actual part you can order. They say they are reference only but I have never had one with a major mistake in it.

    Trust is a big issue with parts libraries too, GrabCAD is supposed to be a place you can quickly get models, but without quality control the models tend to be just some approximation made by someone with a ruler and eyeballs, not something you can trust to design around.

  3. My personal process for finding an electronic part to use goes as follows:

    1. Go to digikey
    2. Search for part requirements using filters
    3. Sort Price -> Ascending and punch in a small quantity so I don’t have to wade through pages of parts with a 10000 MOQ

    Works for me.

  4. It’s not useful because there is too much variety in mechanical parts. I think if you did a survey of what people actually use, you’d find that across all electrical designs, something like 95% of them contain at least 1 resistor, at least 1 capacitor, and at least 1 IC. But for mechanical projects what do you have? Screws? Even they aren’t always in a mechanical design.

    I’ve worked at a number of tech companies and about the only mechanical part I see stocked in each of them is screws. For electrical parts you generally have a kit of resistors, capacitors, wire crimps, fuses…

    Even if you expand the mechanical list to include other common things like springs and bearings, one person/company A is typically going to work on projects at a different size scale than person/company B.

    And by their very nature, shape is what makes mechanical parts do the thing they do. In electronics projects, the PCB is usually the only electrical part that you can’t order from a catalog. But how often do you see a mechanical design that was ordered directly from catalog numbers? Almost never.

      1. Why not standardize on the 1/4W 5.6k resistor and 10uF electrolytic cap for casual electronics builder ? Because it doesn’t work in each and every cases, even for an amateur.

        A quick search on 123roulements (french supplier of bearings) shows 118 references for OD 22mm, ID 7mm, H 7mm bearings, aka 608, 15 of those by SKF. Which one do we choose, the ZZ or the 2RS ? If I need axial loads, why use a 608 when any angular contact bearing would work way better ?

        The problem is not the bearing, it’s which one to use in which case. Standardizing around parts is just gonna end up with suboptimal solutions. Mechanical engineering (as a hobby, not for mission critical parts) is not that hard, just like electronics or other fields, you just have to have a bit of understanding of forces, tolerances, how to navigate the various suppliers websites and read the datasheets… Do your homework, read some books, try some easy stuff, then buy better tools, and build better stuff.

        As for the original questions : not sure a library of standard parts is useful, but a library of standard designs (with explanations) would be a great tool. I’m thinking about 507 mechanical movements by Henry Brown, for example, but with CAD files. I have somewhere at work a 1980s bearings catalog with standard use cases, great ressource.

        1. 608 bearings are great because they are widely available and much cheaper than other sizes due to how commonly used they are. Just pick a permanently lubricated, sealed bearing. For the vast majority of things you do it works fine. You’ll be able to finish a lot more projects having some parts on hand.

          I standardised on 1.2k resistors because I got 1000 for free and they worked great. Flexibility of design and all.

        2. > a 1980s bearings catalog with standard use cases, great ressource.

          Now THAT is an excellent point. Older semiconductor databooks were good in the same way. Hmmmm!

          I own 507 Mechanical Movements, it’s wonderful. I think of it as conceptually right up there with Graf’s Encyclopedia of Electronic Circuits. Most of the circuits don’t have component values, they’re just conceptual, then you detail them into your specific need. But neither of those helps you stock your wall of parts…

          To your point about specific components: Yes, sometimes you need more than 1/4-watt, but often you need less, way less, and you use a 1/4-watt part anyway because there’s no need to stock less. Have the E6 or E12 value series in 1/4-watt and you’re pretty well equipped for most needs. And once in a while, you special-order something for an oddball need.

          So likewise, a bucket of 608 bearings (thank you for the term!) will fill a lot of needs where the suffix doesn’t matter, and once in a while you’ll order something specific.

          The 3d printer folks also have a family of “vitamins” they use a ton of, specific nuts and screws that a whole ton of projects have standardized on, mostly M3 stuff. Because those folks have those parts around in great quantity, they use ’em in other projects, and pretty much any lasercut box with captive-nut joints will be sized for M3 hardware. So that goes on the list: http://www.desiquintans.com/3dvitamins

          That’s interesting because otherwise, I don’t keep a lot of metric hardware around. Specific stuff for specific needs, but otherwise mostly 4-40, 6-32, and 1/4-20. There’s some M4 stuff for VESA mounts, too.

        3. I remember seeing something like that in relation to Misumi extrusion. As animations and full cad files. Possibly even based on the 108 movements… If I remember correctly it was a Japanese website.

  5. > It’s not useful because there is too much variety in mechanical parts.

    I’m sure people said that about electronics, too, but then we got Common Parts Libraries and they really do help.

    > contain at least 1 resistor, at least 1 capacitor, and at least 1 IC.

    Right, but which values? That’s like saying most mechanical things have at least one threaded fastener, one rotating joint, and whatever. Okay, fine, but some of those are more common than others. Unfortunately, it takes a lifetime of experience to know which those are. I’m hoping to find someone who’s condensed that experience into a spreadsheet.

    Thankfully for electronics, the EIA came up with the standard value series, so instead of trying to stock every possible value of resistor, you can have 10, 12, 15, 18, 27, 33, 47, 68, and 82 of something. Do that across a few decades and you’ve solved resistors for most needs. Standard values and preferred numbers solve a lot of problems.

    In fasteners, we have standard values, called screw sizes. I have M4 and M6 screws, but there is no M5.333 screw. This is good.

    But it gets weirder when you get to raw materials. All sorts of stock is made in standard sizes, plus weirdo sizes made for special requirements. The problem is that the catalogs show them all, without hinting “these are the ones you’ll use if you’re sane, don’t order that one we only carry it for Bob and nobody else has it”.

    One step would be to see which items are available from the most suppliers — if McMaster has it and Grainger has it and Fastenal has it and MSC has it, and if they all have it *in stock* and ready to ship, then it’s probably a pretty common part and maybe I should have it too*. But if only 2 or 3 of the usual suspects have it, or if it exists in their catalogs but nobody actually keeps it on hand, then it’s probably not very commonly used out in the world. This is basically what Octopart does for electronics, and I don’t think anybody’s currently doing it for mechanical parts.

    * I should have it for two reasons: First, something around my shop might break and I might open it up and discover that, hey, the broken part is a common thing that I happen to have on-hand. Or, I might download plans off the internet and discover that the designer used commonly available parts, and I happen to have them, and I can build the thing without having to tweak the design, or if I order them, I can order extras and they’ll be likely to come in handy for future stuff.

    Second, and more importantly, is that if I make my own design and share it on the internet, someone else might download it and make their own, and maybe they’ll have the parts already, or maybe they’ll order ’em up to make my thing, they might order extras, and then they’ll have them around for future projects.

    Right now, if I design something that needs bearings, I’ll try to use a rollerblade skate bearing, because I now understand that they’re super common and cheaply available. But if I’d designed something a year ago, I wouldn’t have known that, and I might’ve picked an oddball size that works for my application and maybe somebody had surplus they were getting rid of cheaply, but going forward I’d be saddling myself and others with legacy special-snowflake parts. That’s what I’m trying to avoid, you see?

    1. Yeah the ‘if it can be found in stock in three places..’ advice is good.
      On a lot of things I tend to go with “find two very similar parts. If one is 4 times the cost of the other, it must not be common”
      Like you can get virtually any value resistor in any standard package (not just the common series values) at all the standard tolerances down to 0.1 %. But a 1.234k ohm resistor in 0805 package at 0.1 % cost a LOT more per unit than a 1.2K at any tolerances you want.

  6. The closest industry has been able to get is the set of preferred number series:

    https://en.wikipedia.org/wiki/Preferred_number

    specifically the logarithmic series.

    Instead of choosing materials and dimensions from whatever a specific calculation says, stick to a reasonable set of standard values. It isn’t hard to stock a shop with values that land close to an E6 series (each about 120% the previous one), whether you’re talking about resistors, wire, sheet metal, pipe, bolts, bearings, or gear teeth.

    Design for those values first, and only choose new values when you have a specific need for them. Err toward working a size too large if you don’t have to min-max your design for production in quantity.

    Working that way, you can accumulate an inventory of parts and materials that all work together, and can build a shopping list of things to buy going forward. You’ll also develop a collection of standard designs that you can scale up or down for a specific job. The math of both electronics and mechanics makes that possible.

  7. The AST bearing catalog is basically the best thing ever. It has the dimensions, load ratings, and CAD models of basically every standard-sized bearing ever, and is super easy to navigate:

    https://www.astbearings.com/catalog.html

    My usual strategy is find the appropriate bearing on AST, then search Amazon/Ebay for it first. For deep-groove and thin section metric bearings, they’re usually incredibly cheap in almost any size, as long as you don’t need high precision bearings.

  8. Some people here is missing the point. Of course M8 bolts won’t be useful for you if you focus on miniature RC airplanes… But that’s not the point. A standardized list is a great tool that can be tailored to many needs. For instance, if you grab a part from that list, you could give it feedback that you’ll be using the part for a 3D printer hack. Then the site can use that feedback to present others a list of parts that are popular in the 3D printing world, and it will include your part!

  9. Another guy who uses the McMaster catalog for downloadable 3D models as well. My usual rule of thumb for figuring out which parts are the commonly available ones is that they’re often significantly cheaper from both the next size up and next size down.

  10. I happen to live about 20 minutes from McMaster-Carr, so I’m spoiled and can pick things up via “will call”, but that aside…

    I think a catalog of parts which “ranks” parts according to popularity and availability would be VERY useful. For example, a “A” part would be available from your local hardware / big box store, a “B” part from the usual online vendors, and “C” parts from specialty shops, etc.

    This would like have the commoditizong affect where the most common parts would become even more popular and available, driving down the costs.

    This would also be a big help for new people, students, occasional hobbyists, etc., who don’t want to fall into a trap where they design in some weird part that looks great in CAD, but costs a ton to ship, if you can even get it.

    I could even see something like a “McMaster Express”, sort of a store within the store, for selling parts in small volumes (analogous to Ali Express).

    1. “I think a catalog of parts which “ranks” parts according to popularity and availability would be VERY useful. ”

      I agree, but then, I can see $ony or Apple deliberately telling their engineers to avoid the higher ranking parts…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.