Casting Gears At Home

Automatic doors and gates are great, except when they fail, which seems to be about every three days in our experience. [MAD WHEEL] had just such a failure, with a plastic gear being the culprit. Rather than buy a new drive unit, they set about casting a replacement in metal.

The video is light on instructions and heavy on progressive rock, and may be a little difficult to follow for beginners. The process begins by gluing the original plastic part back together, and filling in the gaps with epoxy putty. A mould is then created by setting the gear in a gelatine/glycerine mixture. This mould is then filled with wax to create a wax copy of the original part. The wax gear is fitted with cylindrical stems to act as runners for molten metal, and then a plaster mould is made around the wax positive. Two plaster moulds are made, which are placed in an oven to melt out the wax.

The aim was to cast a replacement part in aluminium. The first attempt failed, with the aluminium cooling too rapidly. This meant fine details like the gear teeth simply didn’t cast properly, creating a useless metal blob. On the second attempt, the plaster mould was heated first, and this kept things hot enough to allow the aluminium to fill in the finer details. With that done, it was a simple matter of some post-processing to remove the runners, clean up the gear teeth and refine the shape of the gear on the lathe.

The resulting part does its job well, meshing properly with the other gears in the drivetrain and moving the gate effectively. Many in the comments have stated that the original gear being plastic was likely as a safety measure, to strip out in the event the gate is jammed. While this may be true, it’s a far more robust design practice to instead use a breakable plastic key rather than breaking an entire gear in the event of a problem.

Casting is quite accessible to the dedicated home maker. It’s a great way to make custom metal parts once you’ve learned the fundamentals! Video after the break.

45 thoughts on “Casting Gears At Home

  1. Typically, manufacturers will put a deliberately weak component in the power train to limit the amount of maximum torque it can supply.
    It would be better to use a shear pin or shear key but also more costly.

      1. The weak component would be the simplest way to get the product safety certified. If it were done with current monitoring the whole electrical and software system suddenly becomes safety critical and the whole thing gets a lot more complicated.

    1. Plastic gears = avoids need for an oil circulation system or reliance on user-directed lubrication. Plus a /decent/ plastic gear is going to be just as strong as a cheap sintered metal gear (ain’t nobody gonna pay for a proper hobbed gear if they can possibly avoid it!) and more durable when it comes to shock loading.

      1. I’m sure real engineering plastics are going to be just as good, if only they would use them…

        Of course the real problem is probably missing electronic torque/current limits half the time. Or unnecessary mechanical stuff in the first place that could have been an extra motor and digital control.

    1. It might have something to do with keeping the detail of the teeth. Correct me if I’m wrong, but maybe the wax causes some shrinkage which needs some extra depth to prevented that? :-)

      1. Aluminum is difficult to cast fine details with because it gives up its heat far too fast during a casting process. It goes solid too quickly as it fills the void of the mold, leaving lots of areas unfilled (hot aluminum pouring in can’t remelt the cold aluminum that has set already). The trick is to heat the mold ahead of time, then pour the aluminum. This allows the aluminum to flow further before setting.

        1. Aluminum is not difficult to cast if you have a reasonable pouring temperature. In the green sand foundry I worked in we had pouring temperatures from 1100F to 1375F. Aluminum does have problems flowing into thin walls and closing, but this can be overcome with head pressure assuming the metal is still molten.
          Aluminum does shrink as it cools. The overall part dimensions will shrink maybe 1-1.3% according to the Wiki listed below. Most importantly, the casting can have internal shrinkage voids because the metal will solidify from the outside in. Internal voids can be avoided by adding risers which have large cross sections so they freeze last and allow molten metal to feed the thicker areas of the casting.
          For this disk shaped part, a riser could be placed above the center of the gear and could feed metal into the center of the gear while the outer edge of the gear freezes. Head pressure can be increased by increasing the height of the sprue cup above the casting.

          This Wiki description is pretty good for more familiarization:

          This page on the American Foundryman’s Society has a good list of standard terms and meanings:

    2. They added some sort of tape to the tips of the gears to account for the metal shrinkage. IIRC you need to add something like 3% to dimensions. A pattern-maker’s ruler has different scales for different shrinkage amount, so the ruler compensates for the material shrinkage.

    3. The guy added thick double side tape to the ends of the teeth and then machined the excess off on the lathe. Then he re-shaped the teeth in a bench vise. Probably with a hacksaw and a file.

      Basically, the casting just acts as a blank and there was a lot of tedious filing and fitting not shown in the video to make the gear teeth mesh properly.

    1. I thought why not a shear pin at first too, but I think I can explain why it wouldn’t work well I this application.

      The reason for the plastic gear isn’t to protect the gear train…it’s there so that it can’t crush a person to death. While it certainly wouldn’t feel good to get stuck in the gate, the plastic gear is ment to strip before it could do catastrophic damage to an adult (would probably still be rough on a kid).

      Why not a shear pin? Shock loads. A shear pin safety device in this application would be very small and shear pins in general break right at their rated load. Imagine if the gate were to hit some sort of movable object in its path…the pin would shear at the first sign of an overload, but the plastic gear would tolerate an overload long enough to move the object while still stripping in a stalled situation like for example crushing someone.

      It has a weaker plastic gear instead of a shear pin to make the mechanism tolerant of transient loads while still having a definite force limit.

      1. Except it didn’t strip; it broke. I’ve been around the kind of people who do this sort of stuff enough, and been on the receiving end too much. If you can cut cost while also ensuring people need to buy overpriced service parts and maybe even labor, a lot of companies like that a whole lot. If cynicism isn’t enough proof, just consider that it’s their intended failure point but you can’t buy just the gear…you need to buy the whole gearbox. Manufacturers pull it all the time. $0.20 bushing worn out on your Toyota shifter? You can only get a new one with a new $200 shifter assembly! Headlight motor gear stripped on your Chevy? Buy a new motor, because that’s the only way they sell that plastic gear. Child lock button on your dodge corroded (and designed to fail in the locked mode)? Buy a whole new driver door control panel. I won’t mention things done in my departments for NDA reasons, but to a lot of corporate scumbags it’s the height of engineering. Sure, people like us get pissed off and go cast gears or solder a bypass or machine the end of the shifter to fit in a different bushing, but most people just buy the extra parts they don’t need.

        1. My response was strictly to the shear pin vs weak gear argument and I think it’s pretty valid. I didn’t watch the video becuase I’m lucky enough to be able to just machine a new gear.

          However I cannot argue with you about how shitty design to fail parts or parts that have been “value engineered” to the point where they fail for no reason are. I’ve been on the short end of that stick myself…2010 Camaro SS had a knocking sound from the transmission, just had first kid and didn’t have time to investigate so I went to the dealership. The torque to yield driveshaft bolts had loosened up…$450 and 6 months later same issue; I replaced them with 12.9 bolts, torqued them below yield and backed them with distorted thread nuts. The real fix cost me $40.

          Point being the specifically in automotive, recurring service is a money maker so there is a horrible mixture of poor design, designed to fail and pure price point driven “value engineering”. The worst aspect as you mentioned is that you can’t just by the one broken piece…they sell most things as an assembly because it’s more profitable to replace the whole thing rather than fix the broken part.

          Another example…I’m a machinist and spend most of my time fixing machines in my shop. My current project is a $15K air compressor that wouldn’t build over 50psi. I discovered that a screw deep inside the machine had backed out, it is the only screw that the service manual doesn’t call out a thread locker for. Once it backed out, a small piston traveled out of its bore and sheared off an oring, which caused the malfunction. Now instead of making the piston removable from the outside like every part of the machine, this critical part was speced to not have thread locker and requires removing a 200lb steel casting for service. On top of that the factory orings are only available in a $5K kit…a few measurements later and $100 with of parts it’s fixed.

          If you want to see stuff specifically engineered to be expensive to repair, look at industrial equipment…you’ll wish you just needed a $200 shifter or door switch cluster.

          1. Having transmission fluid and filter changed on a 2007 Ford Expedition. Around then is when Ford decided automatic transmissions no longer needed dipstick tubes. There’s still a dipstick but it’s a stubby bit of plastic under a screw on cap.

            To refill after dropping the pan to change the filter, there’s a fill tube and funnel kit. Then the vehicle has to be put on a lift (preferably twin post or drive on rack) or parked over a pit, to check the fluid level, which must be done with it warmed up. The stubby dipstick is conveniently located really close to one of the catalytic converters.

            There *were* aftermarket dipstick tubes available but now I can’t find one for any Ford F150 or Expedition earlier than 2010. Ford has of course changed things down on the transmission. Even worse is the aftermarket tubes for Mustangs are available pretty cheap but the truck one which are less than twice as long cost over twice as much.

            Adding up the price of an aftermarket tube, the filter and fluid, I called the local Ford dealer and they told me “Around $250.” Can’t do *this* job myself for less, and since the transmission doesn’t leak yet there’s no point spending more on a job I don’t plan on having to do ever again on this vehicle. Plus there’s the whole being unable to find an aftermarket tube for a 2007 F-150 or Expedition.

        2. One thing you’re forgetting is that labor costs at most mechanics are very high. Most mechanics charge in my area charge $80/hr – more at a dealership. If a part costs $0.20 but you have to spend three hours taking stuff apart and putting it back together, that’s $240.

          Of course, this is very frustrating for the home mechanic. Three hours on a weekend is nothing to save $200. Most people aren’t home mechanics though, so the system isn’t really optimized for them.

        3. So what you’re saying is, car makers and others should make everything out of titanium so it will last forever? Good luck affording the vehicle in the first place, let alone power windows… But let me guess, you only purchase used vehicles without a warranty – So how are you going to get a used vehicle if the first owner can’t afford one either? Or simply never sells it, because it was made out of titanium and will last forever, negating the need to buy a new vehicle altogether…
          Last I checked, automobile companies are in business to make a profit, and in making a profit they also employ people. If they made cars that last forever, they’d quickly go out of business and no one would have a job.
          I bet, or actually know, its not as simple as you think it is, and there is not one automobile maker around that purposely designs things to stiff their customers…. It’s too competitive of a market for that to work for one thing. Also, I’ve never heard of a “value engineer” and neither has AvE’s dumb ass… They don’t exist. There is NO ONE who goes to engineering school to learn the art of making things fail so they can stiff customers, there are no departments at any automakers full of “value engineers” who’s job it is to rob you… If any company did that, they would go under in a week. Too many other companies vying for your money to do stupid stuff like that. No, Mr. AvE just likes to say things like that because it sells subscribe clicks to people that already believe such non-sense.

          Now take off that tin foil hat and get back to those oil changes….

          1. No, he’s simply saying that “built in obsolescence” isn’t a myth, it’s a fact.
            And I’ll add that it’s contributing to waste, we dont recycle fully so we throw away items that 99% of the parts are still fine because the 1% part failed exactly as it was designed to do.

            If we required a 10yr warranty on things, they would last more like 20 years.
            But hey, how would the company make money if they can’t sell you the same item over and over every ten years ?
            I guess I must be a communist for dreaming that we might value longevity over profit.

            I work in manufacturing and product design / engineering. So I see how this is done on an almost daily basis.
            If the warranty is 3 years, no point specing a component with a MTBF of 10 years – depending on the market for the product and the customer expectation.

            But I’ll even sell you the same product with 3yr or a 5yr warranty. Only I’ll up the price of the 5yr product to cover early life returns. Exact same components, market only expects 3yr warranty but you are hapy to pay extra.

            It’s not that you set out with a goal to make it “fail early” it’s failing right on time. It could last a lot longer if the BOM cost went up. But again we are talking a few cents, which over a production run is your mortgage.

            WIth cars it’s quite simple. The person buying it new is your customer.
            The person buying it from an approved dealer is your customer.
            After those 5-6 years, the next person is buying outside your structure. They are no longer your customer so who cares what breaks for them.
            That’s how 90% of your target business works, so again, dont concern about the 10% which is disproportional cost to make profit from.

            Value engineering. Try searching for it on google and get yourself a free education.

          2. ” There is NO ONE who goes to engineering school to learn the art of making things fail so they can stiff customers”

            They may not learn the art of making things fail specifically so that they can stiff customers but almost all engineers do learn the art of making things fail. Its called Failure mode and effects analysis (fmea) and it is quite critical in engineering as knowing how your mechanism is going to fail makes it easier to make a better product for a lower cost.

          3. First of all I said that I’m a machinist, not a mechanic. I have a porter for that. …perhaps you should take a reading comprehension course.

            Secondly, titanium alloys are poor choices for most things outside of their aerospace/military uses…they are prone to strain harding and galling that would require unreasonable mantinance in most applications.

            Thrid if you really think “value engineering” (VE from this point forward) doesn’t exist you are either very naive, completely ignorant, ot just out of your mind. While I will admit that you do are correct in saying you don’t go to school to learn VE, it comes later in the soul crushing world of manufacturing.

            It goes like this, you are tasked with designing a bracket and linkage mechanism (similar to the shift assembly above)…you design an elegant mechanism that will be reliable and serviceable. You submit your work and a bean counter responds:

            1. The alloy you have specified is too expensive. This is a not safety critical part so a 10:1 margin is not needed, redesign for standard low carbon steel.

            2. Your design would require machining from a solid billet in a 5 axis CNC machine and is too expensive. Redesign for a near net casting and minimal 3 axis machining.

            3. There are too many expensive bearings in the assembly. Redesign for minimal polymer bushings.

            4. The tolerances specified are very tight and too expensive. Redesign for looser tolerances.

            If you refuse, you lose your job…but you have bills and don’t care for starving to death. So you comply and redesign your assembly to meet a price point…i.e. your design has just been VALUE ENGINEERED. In addition, instead of expensive but readily available bearings the design is now full of custom polymer bushings because they were $0.03 cheaper per unit.

            At this point a higher level bean counter says “all these individual parts are too expensive to see individually” and finally we come full circle…the design is now prone to failure because it was designed to a price point and is unservicable because it’s full of custom parts. The decision is made to sell it only as a unit, because it’s faster to replace that way and makes more money in the end.


      2. Good analysis – as with every design one needs to match the solution to the problem, but would have been nice to have a spare gear if that was the “safety” so the equipment was not down while a new part was ordered – but a metal gear probably defeats the safety.

    2. My instinct is that they wanted the cheapest BOM possible while pretending to have some modicum of safety features. Expecting a plastic gear to strip at some specific uncomfortable-non-fatal load sounds like a pretty uncontrolled process while a shear pin will deliver reliable results every time. I’m sure a discussion was had and the conclusion was that a plastic gear gives some safety /and/ it’s cheaper than a metal gear. Win all around for the MBA folks.

    1. I replaced Zamak gears on an Atlas lathe for a makerspace.

      Zamak is a horrible material for gears. It wears out really quickly, doesnt hold form, and breaks easily.

      It was used for economy’s sake. Its only advantage is taking compressive loads well, which is nice for gears.

      7000 series Aluminum if you must make gears from aluminum. Honestly, its easier for me to machine one from steel.

  2. Gears can be cut an a lathe, normally they are cut on gear cutting machine (very expensive).
    A fitter machinists skill is in being able to make a part when none is available. Could also be machined up from nylon stock or similar which is easy to work with. I have seen countless plastic gears etc. strip out and in my opinion it’s poor design.

  3. If you replace a plastic gear with a metal one, i would also very carefully check the electic safety compromises coming with that.
    Sometimes a plastic gear is used to electrically isolate the motor from the rest of the mechanism, in case of an electric fault in the motor.
    You may want to add a ground wire…

    1. Nylon plastic is self lubricating. Not that it literally oozes oil, but more that at the microscopic level there is oil trapped in the plastic that makes it slippery at the boundary layer. You couldn’t squeeze oil from the gear, for instance.

      1. Nylotron 6 is a dark blueish gray purple color molybdenum disulfide impregnated version of Nylon.

        I used to be a plastics and composites machinist, and I’ve machined so many custom plastics I’ve lost count.

        I did machine some 2 piece gears from natural nylon, which is very strong, but very brittle, for chicken processing purposes, but of all the plastics I would choose to make a gear out of, it would be either the bright green glass fiber filled nylon, or Nylotron 6. I did so many crane pulleys from Nylon 6 I lost count. It self lubricates, and is pretty strong.

        Acetal is too brittle, ABS and UHMW too squishy, though ABS is much less squishy, easy to distort all the same.

        1. We replaced steel sprockets in a machine we sold a number of with sprockets made from Nylatron GS. The steel sprockets couldn’t survive the wet salty fish processing environment. The only problem with nylon based plastics is their tendency to absorb water and swell. The filled nylons are better, but after 10-20 years of service, the nylon sprockets have grown enough that the chain no longer meshes as well as it should. The sprockets could go back in the cnc mill to be re-cut but the customer doesn’t care enough to bother. Instead, someone gets to play with a file in the off-season.

  4. I repaired a gear on an electric retracting step on a motorhome. The gear was mostly plastic, injection molded with some sheets of cheap steel laminated in the middle. It had a square hole in the middle with a raised hub on both sides.

    The plastic had cracked from the corners of the hub and the gear was warped. A bit of work with a press un-warped it and got the cracks mostly closed up. One hub was in the open inside the gearbox. The other side worked as a plain bearing inside a recess in the die cast housing.

    The ‘free’ side hub was a bit smaller for some reason. That led to my fix. I had a bit of brass pipe that happened to be a perfect fit over the smaller hub. I was able to press it on and force the cracks together. For the larger side I found a piece of steel pipe about right size and bored it then pressed it on. Then I turned the brass down to a slip fit in the recess inside the housing, flipping the gear upside down.

    After a good application of some grease and reassembly, the step worked pretty well, only needing an occasional kick to get it to fully retract. Had no problem extending. Much improved over being 100% not working.

    If the steel laminations had also been used in the hub sides rather than just in the main part with the teeth, the hub wouldn’t have cracked.

  5. Often a plastic gear is also used in the faster sections of a gear train to reduce noise – and of course cost. In the faster sections you have less torque, so you can get away with that. I don’t think its a safety thing – too uncontrollable. The safety has to be designed in by e.g. monitoring the current and reverse/stop if the drive is blocked,

  6. “They use plastic gears so they will fail eventually and you’ll have to buy a new one, its true because some funny chap on youtube said so, and I like his personality a lot, even if he has been proven wrong enough times, I will still believe him because he’s funny, and I like him and he confirms my tin foil conspiracy theories…”

    Metal gears are not always the best choice… Sometimes they do actually wear faster than plastic gears, and sometimes the cost doesn’t warrant the downsides of using them.

    I can assure you few if any companies that want to be in business for very long want to be known for poor MTBF….

    1. “Every company makes everything with top quality in mind, they never use cheaper materials as a result of cost benefit analysis. Value engineering is a myth. No company ever sacrificed product quality for profit”

      I can assure you. Many if not all companies that exist in reality, engineer their product to fit with market expectations.
      If you’re Rolex then it’s expected your over the top engineering goes into it and it’s priced accordingly.
      If you’re making toilet paper it only needs to work once. A few sheets lost at the roll due to production line issues is not going to be a product return issue.

      MTBF is entirely dependant on market expectations for your product segment.
      Engineer accordingly or else you’ll put the company out of business.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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