You’re right, I’m going about this totally the wrong way. Making a bunch of unfounded assumptions is leading to SISO (shit-in-shit-out) analysis…
When you put it this way it couldn’t be more obvious but this is genuinely one of the most useful pieces of engineering advice I think I’ve ever received.
That’s a very good point and something I will consider going forwards. I think I’ve been trying to look at the shaft in isolation without proper consideration of the surrounding components.
As tempting as it is to go with the “make it chunky and hope for the best” approach, I’m really trying to challenge myself here by learning why parts are specced the way they are by considering the forces they are subjected to. I always like to understand the reason behind things, such as deriving formulas and relationships, rather than taking knowledge at face value.
I’m going to revisit the load case again and see if I can find a good reference that gives me a somewhat realistic “ballpark figure”. From there I’m thinking it may be more efficient to whip up a low fidelity FE model of the wheel rim, hub, shaft, etc. instead of doing the hand calcs, so I can get a better understanding of how the parts interact.
I just want to say, your feedback is invaluable to me and I really appreciate the time youre taking to help me learn from my mistakes. I’ve seen your bikeboard builds and watched your videos on the analysis behind them and am very inspired!
I’m sorry to break it to you, but I fear a lot of those reasons sound like “we had too many customer complaints of bent 10mm axles so we moved to 12mm” of “we’ve used a design load of 10g for decades and it works fine”. Impact loads are inherently tricky to predict. But if you can do better I’d love to see it! Maybe there’s some skateboarders or BMXers that will let you stick a G meter on their rides for science.
Cool! The interactions between parts sounds like a good use of FEA.
Absolutely, you could try to match up the existing onewheels with anecdotes of what it takes to break them. I also have 8, 10, and 18mm mild steel bars on hand I can duct tape to my feet and jump on if you want data!
Tell me about it…I did my dissertation on predicting the dynamic response of 3D printed shock absorbers which involved taking into account viscoelastic behaviours and approximating drop shocks so somewhat similar to this in a sense.
If all else fails I’ll just slap a dynamic load factor on and call it a day but no harm in giving it a shot
After lots of umming, arring, and thinking I’ve decided to impulsively take the plunge and just order some parts. It’s easier to ask for forgiveness than it is permission, and it will be a hell of a lot easier to do the analysis with some concrete measurements and materials than it will be with assumptions.
I’m going to need to find a way to cut down the axle to a useful length. Any suggestions?
Also, I’m not 100% sure of the keyway dimensions on the axle and other components. I’m hoping its a 6mm square key but worst case scenario, theyre only cheap.
The bearings I have no clue how I’m going to mount in the rails yet. I’m thinking a machined pocket with a retaining plate bolted over the top but that seems rather crude. Again, I’m open to suggestions here.
I’ll probably need an assortment of spacers and shaft collars to make everything work nicely but that will be easier to work out once I’ve got some hardware to play with.
Addressing the elephant in the room here - the sprockets. I sized them based on “what’s the biggest one I can find on Ebay that fits a 25mm keyed axle and doesn’t cost an arm and a leg” then sanity checked it with the 3Dservisas Esk8 calculator site. I have a pair of Apex 8T ISO 06B moyor pinions, so with two motors and a 1:3 ratio I should have plenty of torque. Estimated top speed in the region of 30mph which is probably overkill. If I end up needing more torque I’m sure I’ll work something out later down the line.
My biggest concern at the moment is due to the offset inherent in go kart rims, there’s going to be somewhat of a large gap between the sidewall of the tyre and the side rails.
There are a few ways I could get around this:
Only use one sprocket and inset it just enough that the chain doesn’t rub against the tyre. This really is not ideal as I want to make this dual drive and may introduce balancing issues.
Make modifications to the wheel rim such that the hub can be mounted on the opposite side. I don’t like the idea of damaging the rim because I risk compromising it’s structural integrity. FYI It’s not pictured in my rough and ready CAD model but there are ribs on the other side that prevent hub mounting:
Add standoffs between the wheel rim and hub so that the hub can be mounted on the opposite side. This is probably the best option. A custom CNC machined spacer would be nice, but in a pinch three ESK8 axle spacers should do the trick.
Modify one of the sprockets to reduce the collar length. Wouldn’t gain much because the grub screw is close to the edge as it is:
So I think option 3 will be the route I go down.
Edit: This is what the spacing will look like with option 3:
Very little difference to be honest. I’m not sure if it’s worth the loss of stiffness of the hub/rim connection. And I might end up needing more clearance anyway for the chain so that idea is probably going out the window.
This is very clearly a fundamental flaw with using two sprockets but I’m a stubborn fucker and I will continue anyway. I’ll probably 3D print some chain guards anyway once the design is more mature to hide the gap and protect myself from the chain.
Motor mounts will look something like this where they mount to the rail and footpad. Exact dimensions TBC of course, lots of interference in current version (especially with the chain). Slightly concerned that only two of the motor mounting holes are being utilised so may instead design a 90 deg. angle bracket that mounts between all 4 motor holes and the footpad.
Behind the motors will sit a fully enclosed waterproof 3D printed enclosure mounted to the rails and/or footpad containing the ESC and other electronics. Underneath the other footpad is where the battery will go. Yet to figure out how to route the wiring either side of the wheel. A channel inside the rail is out of the question unless the rail is made extremely thick, so it might have to be routed round the outside which is a little ugly…
I was worried I wouldn’t have the room to create a channel to route cables from one footpad to the other, but it should be possible to do something like this:
A few of the bits I’ve ordered arrived after boxing day. Seems that Santa was slighlty late but I’ll forgive him.
The axle, the hub, the bearings and the wheel rim. Which means I’m missing the sprockets but they’re on their way from China so I expect to be waiting at least another week for them.
In the meantime I’ve been playing around with the components.
There are a couple of issues with the hub/rim assembly. Firstly, the slight extrusion around the hub centre has a slightly larger outer diameter than the rim centre hole. You can’t see it very clearly in the images but take my word for it. The hub is offset by somewhere between 3-5mm along the length of the axle from the rim when assembled. I’ll have to revisit the CAD to see if this is going to be an issue but I don’t think it will be. Worst case scenario is that I’ll mount the hub to the other side of the rim with some standoffs which was one of the ideas I previously mentioned but scrapped due to the stiffness tradeoff. Alternatively I could take a hole saw to the wheel rim…but that is a last resort.
Secondly, in the process of playing around with how to mount the hub, I assembled some axle spacers I had lying around (I think they’re @Boardnamics from the caliber hangers possibly?) onto the M8 stud bolts in the hub. Due to a slight lip on one of the aformentioned bolts, one of the spacers had to be torqued down with a nut to push it in place. Total neanderthal move that…because now I can’t get the damn thing off. I’ll be replacing the stud bolts anyway because the threads are a bit fucked but I would like to save the spacer if possible. I’m thinking of taking a blowtorch to it but mechanical methods are preffered. Maybe a bearing puller will do the trick…
Today I tried my hand at cutting the axle to length. I was very nervous about this as I had no idea how hard the axle was going to be and without any appropriate power tools it was going to be tough. I hacked away at it for half an hour with a shitty little hacksaw and managed to cut off what I thought was 180mm of solid steel bar.
On the plus side, this should be fine. Although it won’t engage the whole way through the bearings on each rail, it should be fine for some prototyping and I have more than enough axle left over to try again.
It’s pretty clear that I can’t allow the extra hub offset on the left hand side without pushing the sprocket further out. In turn, that will make the entire build wider by about 10mm, give or take some. It doesn’t sound like much and that may be useful to give the chain some clearance at the expense of looking a bit uglier. It will also give me more room to adjust the motor positioning as they’re currently modelled as butted up to each other which would cause a lot of friction and heat generation.
Here I’ve also modelled in my bodged axle. I’ll definitely have to fix this as it will put a lot more stress on the bearings. But I’ll work with what I have for now for prototyping purposes. Putting the hub the other way round would give me plenty of room for the sprockets, make the assembly more symmetrical which may help with balancing, but from my trial fit it would severely impact the ability to access the valve on the inside of the rim that’s used to pump up the tyre.
So I think the plan moving forwards is to make the axle longer and therefore make the entire board slightly wider. Currently the CAD model measures in at 184mm wide. According to the image below, a OW GT measures in at approximately 240mm wide and a Pint X at approx. 200mm wide, so it will probably do the build a favour from an aesthetics and useability standpoint. I might beef up the rails a bit around the bearing housing too. I have avoided it because I know the thicker they are, the more expensive the plate stock to machine and therefore the more expensive the part, but it seems inevitable if this thing is going to hold up.
Also, a side note - I ramble a lot in my build journal posts. A lot of the time these are a way of me figuring things out as I type. I hope that by sharing my thought processes, fuck ups, bodge jobs and solutions I can help someone 5 years down the line stumbling across this post to build something of their own…
My two cents is that you should do whatever you need to get more clearance for your components. I’ve never done a scratch build where the clearance between moving parts IRL turned out to be less than in CAD, so i just add 5mm more than I think I need everywhere.
Yeah in particular I was looking at the chain clearance. I know your motor setup is symmetrical, but don’t underestimate the ability of chain tension/cornering forces/road bumps to squish your frame out of square. I’ve bitten myself on the ass with that one before
Also, I know I’m US based, but the shop I work in could’ve cut that metal to length in like 10 seconds. If there’s something you need in the future, feel free to reach out.
Definitely a concern. I have a lot more width to play with before things look disproportionate than I originally thought, so I will probably give myself enough clearance to also be able to upgrade the motors if necessary too.
Been using this project as an excuse to treat myself to a few nice tools. Nothing fancy, but picked up the following:
A benchtop vice (long overdue)
magentic soft jaws for the vice
a scribe
a machinist square
a 13mm extra deep socket for the hub nuts
a new set of files
a stud bolt extractor
a set of imperial to metric square drive socket adapters mainly for the stud bolt extractor
A few of those bits arrived today, including the stud bolt extractor. Despite not having a socket adapter for it I managed to use my willpower and sweaty hands to persuade the existing stud bolts out of the hub. I then cleaned out the threads with a tap. Good as new!
There’s yer problem. If you ordered the KEY stock, you would’ve gotten a square rod of steel. But since you ordered KEYWAY stock, you received an empty 6mm void in space that has to be installed on the rotary shaft to hold keys.
