Most bolts to mount accessories and minor parts to bikes are M5. A good example are bottle cage bolts. Components are often attached by M5 and sometimes even M6 bolts (eg long bike in quill stems, threadless stem preload).

Even the worst property class of an M4 bolt offers tensile strengths of at least 3.5kN, that is way in excess of what is needed for a bottle cage or a stem bolt.

In soft materials ripping off inner threads is a concern and may occur well before the bolts fail. However, it is not clear at all if wider bores are stronger. The minimum thread engagement at least is smaller in narrower inner threads.

In case of rivnuts, and braced on threads, smaller diameters might cause less stress concentration in bike frame tubes (see mattnz's comment).

What are the reasons that speak for large bolts?

  • 3
    I wouldn’t call M5 large, I’m not entirely sure what your question is ? Are you saying M5 is large as in 5mm ? I suspect they are used as it’s a common size. If I lost a bottle cage bolt I’d find at least 3 or 4 M5 or M6 bolts floating around in my kitchen draw. Anything smaller than M4 though I’d have to probably go out and buy some
    – Dan K
    Commented Jan 6, 2021 at 12:25
  • 1
    I guess it’s just for safety/wear margin? In case of bottle cage screws you are correct that they are completely over-engineered, demonstrated by the fact that aluminium versions are available. Not sure what the reasoning behind this de-facto standard is.
    – Michael
    Commented Jan 6, 2021 at 15:11
  • 1
    @Michael Nylon bottle cage bolts are available too.
    – MaplePanda
    Commented Jan 6, 2021 at 16:21
  • 3
    Your premise that "In case of rivnuts, and braced on threads, smaller diameters should cause less stress concentration in bike frame tubes." might be flawed (reference required.). When radiusing corners to stop cracks, bigger will better distribute the stress. Obviously this need to be balanced with larger holes making the tube weaker.
    – mattnz
    Commented Jan 7, 2021 at 1:27
  • 1
    @mattnz stress raisers at smaller radii vs slightly smaller fraction of the tube circumference is quite a relevent point. That smaller holes are better cannot be taken for granted.
    – gschenk
    Commented Jan 7, 2021 at 9:25

5 Answers 5


M5 is used far on far heavier luggage too, like rear racks. You don't want too many sizes to keep in your spare parts (I carry a couple of M5 and a couple of M6 screws on the bike in my toolkit, plus nuts, and have used them).

Standardisation is useful even when you've got more room for spares and don't care about the weight - the vast majority of the screws in my optics lab are M4 or M6, though I've used everything from M1 to M10 there when necessary, and not just in my own designs.This holds at manufacturers too - they don't want a massive parts inventory, even of cheap parts. Similarly on bikes, other sizes are used if necessary:there are M3 and perhaps smaller screws in gear assemblies; I think I've seen M2.5 holding covers on housings, but axles are big.

We should assume they're screwed into aluminium, where fatigue on the mating thread is an issue, and, we should also assume they're taken in and out a few times, torqued quite hard each time, as you would fitting a luggage rack. A challenge for you is to try stripping a range of threads in aluminium, all with the same depth engaged - perhaps drill and tap into 3mm aluminium plate. Maybe a thought experiment would do, though I have the tools to try it here. With a few mm of engagement, M3 will strip with some effort, M4 possibly, M5 unlikely and M6 only with a cheater bar. But each time they're done up the aluminium gets weaker. That will soon make up for the thicker plate you'd have in reality - as would using a screw that's a little short.

One way to get round the limit of threads stripping is to specify a maximum torque - and that's what we do on bike stuff. Typical engagement calculations don't treat this as a variable. Engagement length is reduced with bolt size, but the number of turns remains roughly constant. Note that the usual calculations for minimum engagement length are based on the screw breaking before the thread strips. That's only one failure mode and in many applications there simply isn't that much material. This table starts at M6 but expects over 12mm into aluminium alloy, 6mm into steel. You'd be lucky to get that much on disc brake mounts. At that point the larger threads should allow more holding force in the same depth, up to a point. The deeper threads of larger sizes mean far more material which will deform less.

Another factor, common on steel into aluminium, is the threads seizing (grease helps avoid this). Then you want a big enough hex socket head to avoid stripping the thread. A big head means a big thread. Bike bolts, especially for accessories, are often button head, which normally uses a smaller hex key than cap head (disc brake attachment bolts are often cap head). on bikes, special button head screws with bigger sockets than you'd expect from other fields are common. It's quite easy to round out the head in a normal M4 button head, far, far harder in a typical M5 on a bike.

Another reason is tolerancing. Going from M5 to M4 wouldn't be too much of an issue, but going to M3 would. Those bottle cage screws don't line up perfectly between bikes; there's a reason cages have slots. You'd need unusual wide heads to give as much freedom. Washers would provide a bit, but not as much and add hassle. This is also very true when parts are being modified in a workshop, where M5-M6 threads are at a sweet spot: easy to cut the thread, not too easy to break the tap, and taps are cheap. This would apply in traditional frame building as well.

Rust makes much o this worse, but it holds for stainless fasteners too.

  • 3
    @gschenk In various jobs I've used a lot of steel metric fasteners into aluminium, and occasionally overtightened things that weren't working at normal tightness. I've never stripped M6 or M5 (but haven't used M5 as much), though I have bent their Allen keys. M4 I have, but only recall cases where the engagement was insufficient or the thread was already damaged. M3 a few times. Heads round off far more often, normally trying to undo things. I'll make an edit about engagement length, that's too much to fit in here
    – Chris H
    Commented Jan 7, 2021 at 7:01
  • 1
    I'd like to add that bike parts or tools aren't always the best quality. With larger screws you can get away with softer metal and larger tolerances, but tightening the 2.5mm screws on a cheap Floyd Rose copy is a bit unnerving...
    – ojs
    Commented Jan 7, 2021 at 13:44
  • 1
    They'll be some 4-40 and 1/4-20 in the lab just to mess you up.
    – D Duck
    Commented Jan 8, 2021 at 16:14
  • 1
    @DDuck worse, both 1/4-20 UNC and 1/4-20 Whitworth. Despite the sign I made for the lab door (which felt rather cruel when I had visitors from Imperial College, bringing kit with them)
    – Chris H
    Commented Jan 8, 2021 at 16:30
  • 1
    @ChrisH And you'll have a CP03 like part from MicroBench with a missing & unobtanium M2.2 grub screw
    – D Duck
    Commented Jan 8, 2021 at 18:11

Not everything is designed at its theoretical maximum strength. There are always practical reasons for any engineering choice, including "mundane" things, like the fact that I can buy M5 bolts anywhere in the world, and I have a wrench to fit M5 bolts on my compact multi-tool. And the fact that most other bikes also use M5 bolts for their water-bottle cages (i.e. compatibility). Also, larger bolts are more resistant to stripping the threads or the head, and they are easier to find when I drop them in the grass.

Practicality is not a bad thing. The engineering starting point is usually compatibility and practicality, and we deviate from that only if there is a really compelling reason to do so. Saving a fraction of a gram is not a compelling reason in most applications. Even re-sizing all the bolts on a whole bike would only amount to a few grams of weight.

If there is a need to remove every un-needed gram, titanium and aluminum bolts are readily available. I have some aluminum M5 bolts that weigh about 1 gram compared to 4 grams for steel. So by replacing your water bottle cage bolts with aluminum, you can save 6 grams. Please note that you will save more weight by not painting the frame...paint can weigh as much as 30g. Aluminum bolts are readily available on eBay, etc.

  • The weight saved by using lighter bolts is soon undone when you carry 3 or 4 gels and a cereal bar. Needless to say the heaviest item on my bike is the water bottle as a litre of water is equivalent to a Kg.
    – Dan K
    Commented Jan 6, 2021 at 16:40
  • 7
    The heaviest on my bike is me. Around factor 9. A proper hair cut (Pantani style) might also save more than the light bolts. SCNR.
    – Ralf
    Commented Jan 6, 2021 at 17:03
  • Fun fact about paint: I think that the Cyclingtips Nerd Alert podcast mentioned that lighter paint colors tend to weigh more. Unfortunately, I forget which episode this was.
    – Weiwen Ng
    Commented Jan 6, 2021 at 17:12
  • Well, using something oversized is bad engineering. My construction teacher always said in civil engineering you fail for under sizing a bolt (bridge collapses) while in mechanical engineering (automotive) you fail for oversizing bolts. That may or may not save weight or costs is just not optimised or close to perfection.
    – gschenk
    Commented Jan 6, 2021 at 23:24
  • 4
    Your information about the strength of the 5mm bolt comes from laboratory conditions. Literally, parameters like tensile strength are measured under controlled lab conditions. If you have ever participated in a test like notch-strength testing, which simulate real failures, you will know the even getting repeatable results in the laboratory can be difficult. In the real world we have to deal with corrosion and wear, vibration, under and over tightening, thread tolerances... I think these bolts are not really as oversized as you are assuming based on a glance at some numbers on a paper. Commented Jan 7, 2021 at 6:10

Most of it comes down to fatigue. Even M2 bolts will hold a bar to a stem long enough to make it around the block, but when you need to withstand pothole after pothole the failure method you are designing for becomes fatigue.

All bike components need to pass ISO, CPSC, EN..etc standards in order to be sold in various markets. The testing prescribes various combinations of fatigue tests- for a bar and stem interface a combination of 200,000 fully reversing cycles of an out of phase load, followed by 200,000 cycles of an in phase load. Bolts are usually the first thing to go in this test, with stems and faceplates usually outlasting bolts by at least 3 times. Ie. typically you would see a stem make it through 3 test approvals and a bolt only last one approval.

Believe it or not, most bolts do not last through an approval cycle in a stem handlebar interface test. Even the highest quality bolts fail in this test - rolled roots, american sourced. So manufacturers make the assumption that the bolts will be replaced in the lifetime of the component. Or at least they have to, because the tests are generally overkill.

  • 1
    Did the last part of the answer get cut off? It abruptly ends. Awesome first answer though!!!
    – MaplePanda
    Commented Jan 7, 2021 at 18:18
  • 1
    Welcome to Stackexchange - I've taken the liberty of merging your comment with your answer. You can use edit to change your own answer (can even edit other's Q&As once you have enough rep) Good first answer too. You can browse the tour for background on how SE is organised.
    – Criggie
    Commented Jan 7, 2021 at 18:40

Apart from the practical advantages of having a low number of different bolt sizes, and the tensile strength that you cite in your question, there is also the rather underestimated fact that bike parts need to withstand brutal vibrations.

If you have a road bike of 10kg weight and 7bar in the tires, the accelerations of the frame due to road roughness are quite severe. Significantly worse than those that a car wheel is subject to. And bike parts need to withstand this for a long time.

That said, I have often cursed that fact that luggage racks are typically only attached with M4 bolts. M5 is the bare minimum that is needed to provide enough robustness, and I would go for M6 at this point if I could. I have simply killed too many luggage rack screws already...

  • My hybrid and tourer both use M5 for the rack. I drilled and retapped one bike from smaller to M5
    – Chris H
    Commented Jan 7, 2021 at 16:43
  • Dynamic loads is a good point.
    – MaplePanda
    Commented Jan 7, 2021 at 18:17
  • 1
    @MaplePanda Not just dynamic loads, rapidly changing loads. That's what's killing aluminum parts in the long run, aluminum takes a bit of damage everytime that it flexes back and forth. Aluminum frames can only avoid breaking by providing enough material for extreme stiffness and/or suspension. Commented Jan 7, 2021 at 19:54

Tensile strength isn't the only thing that matters. Even an M1 bolt would have sufficient tensile strength to prevent a bottle cage from being ripped off the frame. However, consider the size of the head, and the contact area with the bottle cage. Now consider the pressures this would generate as the side-to-side vibrations during a ride generate a torsion that attempts to pull the head through the soft carbon fiber composite, aluminium, titanium, or even plastic bottle cage.

Even without doing the math and looking up the relevant material properties, it should be intuitively obvious that such a small screw won't shear off, or rip out of it's threads, but rather its head will pull through the bottle cage.

Furthermore, such a tiny screw head would require a high degree of precision in placement of the bolts and the corresponding holes on the bottle cage.

While this could certainly be solved with the appropriate engineering, I'm not sure many people would appreciate the need to pay for an engineered, precision machined bottle cage. There'd be hardly any advantage, and numerous disadvantages:

  • an additional tool must be carried for repairs
  • the screw wouldn't be interchangeable with other screws on the bike
  • such a tiny screw dropped in the dirt would likely be lost
  • 1
    While I can follow your for quite small bolt sizes, are you certain this is the reason, for example, not to use M4 in bottle cages? Standard* head and washer of a socket M4 are 7 mm and 9 mm, respectively, while for M5 it's 8.5 mm and 10 mm. Since the slot for the M5 bolt needs to be wider as well engagement area with the part fixed to the frame would be the same. Your following arguments also apply to cases for unreasonably small screws. However, above M3 they seem to lose weight. *ISO4762, ISO7089
    – gschenk
    Commented Jan 8, 2021 at 17:29
  • 1
    @gschenk Your question specifically calls out tensile strength, implying that is the most critical failure parameter. I mean to point out it's not. As for making the bolts a little smaller, sure, you can do that. On just about anything except orbital spacecraft, there's some margin you can trade if you want to save every microgram above all else. But that won't change the fundamental observation that it's not the tensile strength of the bolt that dictates the design.
    – Phil Frost
    Commented Jan 8, 2021 at 18:36
  • @gschenk In other words, if your question is "why are bottle cage bolts M5 instead of M4 or M3?" what you are asking is "why aren't bicycles meticulously engineered with razor-thin margins and absolutely minimum weight above all other concerns?" The answer should be obvious.
    – Phil Frost
    Commented Jan 8, 2021 at 18:40
  • Well, there are more bolts than just bottle cages that seem oversized based on the premises of the initial question. And that not just by razor thin margins (remember 3.5 kN for a bolt). Your answer does indeed answer why not tiny screws are used, however, limitations mentioned in the other answers may prevent such tiny screws in the first place. I wrote my comment above to provide feedback, for improvement, not fundamental criticism.
    – gschenk
    Commented Jan 8, 2021 at 19:05
  • 1
    @gschenk Again you are treating the tensile strength of the bolt as the only thing that matters. Tensile failure of the bolt is not the critical parameter: you must consider other things that could fail that would result in the bottle cage detaching or the frame being damaged, as well as the numerous practical considerations already enumerated.
    – Phil Frost
    Commented Jan 8, 2021 at 20:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.