Best way to prevent bolts coming loose?

I think that is true for a pedal wrench. For tightening the bearing caps, on the other hand, I could probably over tighten the bolts with the short end of the tool in my hand. So, we need to be careful not to over tighten those ones.

I am somewhat OCD about checking the tightness of stuff. After gathering enough experiences over and under tightening parts of my unicycles, I’ve come up with a set of rules for tightening stuff. Basically they go as follows:

Check the tightness of things often when they are new/replaced. Err on the side of under-tightening things, take a quick ride up and down the street, maybe hop a bit, then check the tightness. If things loosened up during your short ride, the screws and bolts will turn with much less force. Try to remember how much tightening force you used last time. If you need to re-tighten, use a little bit more force this time. If something has not obviously come loose, don’t use more force to tighten it; just leave it alone. If you use this method, things will eventually stop coming loose with such frequency, and you will have avoided, hopefully, over-tightening the parts. Over time, you’ll develop a sense of how tight to make various bolts.

That sums it up for me. The problem with this design is that it combines two main functions into one bolt:

  1. Preventing the bearings from moving in any direction (requires low torque)
  2. Preventing the bolts themselves from coming loose (requires high torque if friction-based, which causes bearing to deform).

Bicycle stems have a star-nut-and-bolt to exert light force on the steerer tube bearings, which prevents play. Once adjusted, the assembly is locked in place by the stem’s steerer tube clamp bolts which can be tightened to a much higher torque and won’t easily become loose. A solution relying on this principle would work well but would be more expensive to machine. Personally I like Nylock and I don’t like threaded housings because if you mess up the thread in some cases (e.g. mad4one) you can’t just put a nut there and you will need a new frame.

Putting shims between the bearing housings is not ideal because the spacing is not infinitely adjustable, unlike the star-nut-and-bolt system, but it’s a better than no shims at all.

Agree that the bearing cups have a horrible design. On a new uni frame I do: Insert a lose bearing (no wheel) and thighten the cups by using different shims. As I have a single bearing, I can turn it by hand (finger), so it is easy to adjust the proper amount of shims (fix the bearing, but do not deform it, it is easy to recongnice when turning by finger)). Shims are available from 0,1mm and bigger. Once shims for both sides are fixed; I insert the wheel and can thighten the bolts with a proper amount of torque.

If you strip a thread in a frame, it may be possible to remedy the problem with a Helicoil insert or similar. I’ve had that done on motorbikes and cars in the past.

That might work on a Nimbus or KH frame. But definatly not on a m4o frame as there is simply not enough material.

In my case finding the correct thickness of spacer washers was a trial and error process. I went though a bin of washers of the correct diameter and found that all were not the same thickness. Using a micrometer I found a range of thicknesses from 0.042" to 0.072" thick. Then I mixed and matched different thicknesses to get the ideal spacing. Different thicknesses of shim stock would also work but the more pieces used would be more difficult to keep track of the correct location when the wheel is removed. Ideally a single spacer washer for each bolt would be best and that could be machined given some time.

In a perfect world the whole bearing holder would be precision milled so that no spacer washers were required when the bolts were torqued to the correct amount.


In a “perfect world,” the two bearing caps would be mounted to the frame, tightened to a specified torque, and then align-honed (both sides at the same time). This is what they do with automobile engine blocks (main bearing journals) to ensure straightness. Then remove the caps and install the wheel. That way, not only are the individual bearings correctly tightened, but it also accounts for minor differences in leg length and skew, which would tend to bind the bearings, even if the holders were correctly sized.
In the real world, this is way too expensive a process to use with something as inexpensive and low-production volume as a unicycle frame.
So it is what it is. Jim’s method is probably about as good as it’s going to get. Personally, I just tighten the caps until it “feels” right, and that’s that. At about 6 years so far I haven’t had any problems. (But I do use blue Loctite. I think it helps.)
Just my opinion! Your mileage may vary! Good luck! :slight_smile:

Thanks for all the replies. I had a really hard time getting the spacing right on my 19" because trial and error meant I had to remove and re-install the wheel for every attempt which was very time-consuming. Kunstrasen’s idea of using a spare bearing for testing shim combinations which is very clever, so I ordered a cheap 22x42x12 bearing for the equivalent of less than 2 USD just for that purpose. I also ordered a bunch of 0.1, 0.3 and 0.5 mm shims. After a bit of experimentation 0.4 mm seemed about right.

I added added Nyloc nuts on the top side and got rid of the silly spring washers. One of the nuts was really tricky to tighten because it is surrounded by the brake rotor, brake caliper and mount, spokes and axle. I had to squeeze a spanner in and literally turn the nut one degree at a time:

Other than to save a few grams, what’s the benefit of having the threads cut directly into the frame? My Tealiang uni uses stainless steel Nyloc nuts which conveniently drop into a matching hexagonal socket that prevents the nut from spinning—no spanners needed there. It never comes loose, it’s cheap and easy to cheap to replace should the threads strip:

Personally I use bond-it A43 threadlocker and reapply periodically after I’ve taken the bolts in and out a few times.

I would prefer something a little closer to the Nylok Tuflok mentioned above though as in my experience, this does a damn good job if the bolt had enough of it applied to start with.

I’ve never had a major problem with any bolts coming loose though. It happens occasionally, and I tighten things up either when it happens, or when I get home. If you don’t carry appropriate allen keys with you when you ride, you’re asking for trouble!

Bond-it A43, never heard of that stuff. But I think Henkel spent a lot on marketing, probably more than Loctite deserves.

Regarding the shims, I had to go from 0.4 mm to 0.3 mm. There was a clicking noise when twisting the saddle with my thighs, I figured it came from the bearing cups. After removing one 0.1 mm shim the noise went away, which means the bearing was not tight enough to prevent it sliding around in the housing. That’s precisely where the whole unicycle bearing housing design fails, because the force required to prevent the bearing moving is more than is required to deform the bearing.

Here’s my vision of a better design, where the detachable housing is built directly into the bearing. The frame-side housing would need to provide a small lip to engulf the bearing to prevent it sliding around, if that makes sense. AFAIK all unicycle frames currently available already feature this lip.

With this design no amount of torque would cause the bearing to deform, because there would be a gap (e.g. 0.2mm) between the top of the bearing and the frame. The threads could be built into the bearing too with bolts inserted from the top, but when disc brake mounts are present this is not possible which is one of the reasons manufacturers choose to put the bolts upside-down (the other reason presumably being to prevent gunk collecting in the hex socket).

Furthermore, with this design frame asymmetry issues (e.g. one leg longer than the other) can easily be corrected by adding shims between the top and bottom.

I wonder if reducing the contact points from the wheel assembly to the frame, to two small areas surrounding the bolts would cause issues.

It would cause issues if the bolt is the only thing preventing lateral slippage. My idea was that the frame would have a bit of a lip going all around the bearing to prevent it from moving laterally. Then the job of the bolts is simply to keep the bearing attached to the frame. This is exactly how current uni designs work.

I’d honestly suggest just riding for a decade, then deciding whether or not this is really an issue that needs to be solved.

I remain unconvinced.

I suspect it’s firstly so that you can’t lose the nuts, but it’s probably also much simpler to manufacture. Making a blind hexagonal hole in a piece of material cannot be easy. Drilling/milling circular holes and threading one of them is a piece of cake.

I’m pretty sure they don’t care if you lose the nuts, nuts are cheap and readily available. It’s almost definitely to do with the hexagonal hole being tricky to machine.

I don’t know if this design is mechanically superior to what we have now. The outer bearing race would have to be a bit thicker (as depicted in my model).

Another advantage of this design is that it eliminates the problem of stuck-in bearings, which can happen when bearings fit too tight in the housings. When you unscrew the bolts, the bearing comes right off under its own weight.

I’ve never experienced a stuck in bearing in a machined aluminium frame. There’s only ever slightly less than half of the bearing in contact with the housings, so the most you get is a bit of suction and the holding force of whatever grease you might have put in there.
I’ve seen it in stamped/formed steel ones but people buying those are buying cheap for a reason.

You would also lose Schlumpf hub compatibility.

Yep, that problem only happens with my cheapo Teliang 19" uni, which uses forged bearing housings (I think that’s how they get the hexagonal hole done too). Forging then machining would fix the problem, but may increase costs. Definitely not a problem with my KH, where the bearings drop right out effortlessly.

Re. Schlupf compatibility, that would indeed be an issue unless the hub itself comes pre-fitted with this sort of bearing, at least for KH unicycles.

My first two unicycles had the bearings pressed into circular holes in pieces of metal that were then attached to the bottom of the forks.

In those days, the circular pieces of metal had a single upwards-pointing prong which went up the inside of the fork leg. We called them “lollipop bearing holders” because of the shape.

If I have understood your idea correctly, it combines some of the supposed advantages of the lollipop bearing with fittings that more closely resemble what you find on a modern uni.

The machined bearing caps on my KH24, KH29, KH36 and my previous Nimbus 36 have simply never been a problem. My older Nimbus 20 and Nimbus 24 had cheap pressed steel caps and they weren’t a problem for normal riding either. My two Pashleys (20 and 26) had lollipops. The design of the 20 was very weak. The 26 worked perfectly well. I think you’re looking for a complex engineering solution to a simple maintenance problem. However if it results in a better, stronger, cheaper uni that’s quicker and easier to work on, I’m all in favour.