Are quick releases safe?

If you're the kind of person who frequents internet bike forums (and if you're reading this we guess that you probably are) you may be aware of something of a kerfuffle surrounding front wheel quick releases and disc brakes.

Here's the story so far. James Annan is a keen tandem rider. He got himself a custom fork built for a road tandem equipped with disc brakes. And first time out he and his stoker found themselves lying on the ground with the fork ends crunched against the Tarmac and the front wheel lying some distance away.

Looking into the causes of the incident, he concluded that braking forces had pulled the front wheel out of the dropouts. Sounds unlikely? That's pretty much what a lot of people thought when Annan described his experiences and hypothesis on Usenet. Lots of people said it couldn't possibly happen, not with a properly-fastened quick release. And lots of people looked at the pictures of his tandem fork and wondered what on earth such a spindly thing was doing with a disc brake on it in the first place.

Conventional wisdom says that a quick release skewer doesn't have to work all that hard on a bicycle. The weight of the frame and rider sits on the axles, so all the QR has to do is to support the weight of any wheels that may find themselves off the ground. But, says Annan, conventional wisdom doesn't know about disc brakes...

When a bike brake locks, it becomes effectively fixed at the brake pads. There's a reaction force at the contact patch of the front tyre equal to the force produced by the weight of the bike and rider. That force attempts to pull the bottom of the front wheel backwards, pivoting it around the brake pads. Even if the brake isn't locked, the force exists, just less of it. But of course the front wheel is also held at the axle. With a rim brake up near the top of the fork, the force on the axle is pretty much horizontal and not all that large as the brake's a long way from the axle.

But a disc brake calliper is low down, near to the axle and often level with it. Which results in an approximately downwards force acting at ninety degrees to a line running between the hub axle and the centre of the brake pads, a force that is very close indeed in direction to the open end of the fork dropouts...

But it'd have to be a pretty big force to pop the front wheel out, right? For that to happen it'd have to be a force big enough to not only exceed the retention ability of the front quick release skewer but also big enough to lift the whole weight of the bike and rider upwards and off the front wheel. James Annan did some sums. And guess what? It can be that big.

Worst-case scenario

In many ways, his tandem was a worst-case scenario. For a start, the fork dropouts were at the worst possible angle, and had no retention lips on the. And the four-pot brake caliper he was using was sat low down, about level with and behind the axle. His front skewer had a smooth face under the lever end, rather than serrations. He was riding on the road with good grip so the front wheel would lock rather than slide. And he was on a tandem, a big, heavy bike with two people on it. Tandems are also long with a low centre of gravity, allowing you to decelerate a lot faster than a solo bike without the back end lifting. More weight means a greater reaction force, which means a bigger force on the front axle, in this case big enough to pop the whole shooting match off the top of the wheel.

At this point Usenet turned its attention to the fork itself, and concluded that it was badly designed. It had thin, flexy legs, vertical dropouts, no retention lips... So Annan looked at regular suspension forks on solo bikes like the ones most of us use, with a slightly forward-angled dropout. And did the sums again. He showed that a fairly normal-weight rider braking sufficiently hard could again generate a net downward force at the front axle big enough to comfortably exceed the ISO standard for quick release clamping force.

So why doesn't it happen all the time? Well, for a start most quick releases themselves comfortably exceed the ISO standard. And there needs to be a certain amount of weight being decelerated about as fast as it's possible to decelerate. And most forks have retention tabs on the dropouts to stop an incorrectly-fastened front wheel falling out.

So far, so hypothetical. But reports started to come in (several to us) of front wheel-related incidents, either axles shifting in dropouts, wheels becoming loose or indeed coming off altogether.

Back in the comfy world of conventional wisdom, front wheels can only come out if the front quick release isn't properly secured. It's not an uncommon problem with novice riders. But many of the people reporting certainly weren't novices, having years of riding experience and certainly knowing how to operate a quick release skewer.

Annan carried out some more investigations, and found many instances of front wheels either coming out or becoming very loose but being spotted before they fell out. We asked BM members about it, and there were a fair few with similar experiences. Again, generally these incidents would have been put down to user error, but we're still talking about experienced riders here. Notably, significant numbers of people report having to tighten their QRs to the extent that they can't open them by hand to keep their wheels in place...

But how could wheels fall out past the retention tabs? Up to this point there's really no disputing Annan's calculations. Cleverer people than us have looked at them and raised no objections. You can plug in lower rider weights and larger rotors and get smaller forces coming out, and skewer tests have shown that the potential pull-out force is much higher than the ISO standard, but it's not all that massive an amount in either direction. But the retention tabs ought to stop this happening. With a retention tab in place, it's not enough to exert sufficient force to make a properly-closed quick release slip in the dropouts. It's actually got to come undone.

This is where the second part of Annan's hypothesis comes in. Engineering theory shows that repeated transverse loadings on threaded fasteners - like the nut end of a quick release skewer - can cause them to become loose. Additionally, heating and cooling cycles can have a similar effect due to the threaded parts expanding and contracting. The first part of the hypothesis suggests that sizable transverse loadings exist, and there's certainly a fair bit of heating going on down at the bottom of a disc-equipped fork. Not to mention the twisting that can occur under braking and over bumps.

So, Annan says, what happens is this. Due to the repetitive loads placed on the axle by braking, the QR skewer gradually unwinds itself. Eventually it unwinds sufficiently that the downwards force from heavy braking is sufficient to pop one end over the retention lips - and remember that the brake calliper is only on one leg and thus applies more force to that side of the axle. With one end out, it takes very little for the other end to come adrift too, and down you go. It's a controversial hypothesis, and given the nature of the forums in which it's being discussed opinion is hugely polarised. (Various discussions and feedback are linked from Annan's original page).

Few would argue, though, that threaded fasteners can come undone. And at the end of the day the QR skewer is just another threaded fastener. There's no particular reason why it should be immune, although good skewers have hard, serrated faces to prevent unscrewing.

Risk assessment

Which leaves us with two questions. How much of a problem actually is it, and what can be done about it? Well, we're pretty convinced that an approximately vertical dropout and rear-mounted brake calliper is a poor design which leaves the quick release and retention lips having to do things they were never intended to do. The quick release skewer was invented a long time ago, certainly a long time before discs were invented. The whole dropout/skewer system has gradually evolved, and its presence on modern bikes is really just a convenient hangover from bikes passim. A subtle redesign of the fork would remove the possibility entirely - putting the brake calliper on the front of the fork leg would tend to push the front axle upwards rather than downwards, and indeed early disc brakes from Hope and Sachs had a front mounting. It wasn't until the development of a standard mount that it moved round the back.

These aren't things that we can do ourselves, though. It's something that the manufacturers would need to address. And thus far the manufacturers haven't had much to say on the subject, although it seems implausible that none of them are aware of it - we've asked a fair few ourselves... Trek have said they're investigating, although it's the legal department that's involved. Which is not Trek saying, "Here's a problem that we need to fix" but "Is there a problem here that's likely to lead to legal action?" Pace put out an open letter saying, essentially, "It can't happen". And that's about it. We know that others are looking into it, even if they haven't said so publicly. Several industry bodies, including the CTC and Australian Competition and Consumer Commission, are looking at the issue. Of course, it's all a rather interesting legal paradox. Say that a manufacturer decides that this is a big problem and redesigns something to make the problem go away. The new design is safer, but by redesigning it they're effectively saying that the old design wasn't as safe as it could be and leaving themselves open to legal action. This isn't unique to this one issue - it's a wonder any safety improvements ever get made to anything...

So what should we do? Do we need to do anything?

Let's look for some perspective. There've been a lot of reports of front wheel related incidents since this issue came to light. Whether that means that there are any more incidents than in previous years it's pretty much impossible to say. Just as previously such reports would be dismissed as operator error, so now we get the distinct impression that all QR-loosening incidents are tending to be attributed to disc brakes. We think there's a bit of over-reporting going on - just because something other than operator error could cause skewers to come loose doesn't mean that operator error has gone away. And indeed casual observation of riders on the trail suggests that there's still a healthy number of people not doing the things up properly. But of course it cuts both ways. The existence of operator error doesn't mean that there's no other mechanism by which a wheel could fall out.

It's worth remembering that there are a lot of variables here. Rotor size, rider weight, QR clamping force, available grip... Some investigations of skewer clamping suggest that good skewers are capable of resisting a pull-out force of some 3-4,000N which is well in excess of what most will experience, particularly if you're a light rider on a light bike. But again, there's a lot of variability - some skewers aren't much good at all. Some forks have extremely deep retention lips that require the nut to be undone several full turns to clear easily, which makes it look fairly implausible that they could loosen sufficiently to be forced over the lips before the rider notices. But again, not all.

To be honest, we think the debate over the unscrewing issue is fairly irrelevant. What's pretty clear is that having the open end of the dropouts roughly in line with the forces on the axle under braking isn't terribly clever. If the skewer breaks or comes undone for whatever reason, having things arranged so the axle is driven downwards under braking isn't what we'd call fail-safe.

But still, the fact remains that the vast majority of people have never had a problem of this sort. We're certainly not going to suggest that everyone immediately junks their QR forks for 20mm through-axles, although if you're very heavy and brake very hard you might want to consider it. But you should be aware of the possibility, and we've got a few recommendations to minimise the chances of this happening to you while the boffins look into it.

First, look at your hubs and skewers. The faces of the hub locknuts and of the skewers should be well serrated and preferably steel. Serrations dig in to the dropout surface which helps to prevent unscrewing and also increases the skewer's resistance to slip. There are lots of hubs and skewers out there with completely smooth surfaces - not good.

The skewers themselves need to be capable of exerting the maximum possible clamping force. Avoid titanium skewers - the shafts can stretch considerably more easily than steel, which is clearly not ideal. Skewers with internal cams like Shimano, Campagnolo and Mavic appear to work better than those with external cams, but either way make sure that the cam is well lubricated. The effectiveness of external cams in particular is hugely reduced if they don't run smoothly.

Don't overtighten skewers. If they're shifting it's tempting just to give the nut another turn and stand on the lever, but doing so stresses the skewer and makes it more likely to break. In most cases, you need to have the lever open at ninety degrees, then tighten the nut until there's a little resistance. At this point you should be able to push the lever fully home. If it closes very easily, you can tighten the nut a little further. If it doesn't close easily (hurting your bare hand is a bad sign) then loosen the nut a little.

Check your skewers often. Generally we stop pretty often out on rides, for gates and snacks and map reading and regrouping. Use the opportunity to quickly check that all is snug.

Having the caliper further from the axle is good from the point of view of axle loading, which implies that bigger discs are better. Unfortunately bigger discs put bigger loads on the bottom of the fork leg so you may be trading one problem for another.

Where from here?

If any engineer sat down with a clean sheet of paper, they wouldn't come up with a design like the one we have for holding wheels into disc-equipped bikes. The forces of tradition and the demands of backwards compatibility are strong. But generally, it works as long as the various components are satisfactory. There's really no rational objection to coming up with something failsafe, though, and we wouldn't be at all surprised to see some fiddling with dropout angles and calliper position in the near future. Clamp-up through-axles obviously make the whole thing a non-issue, but they're currently aimed at the DH market and are fairly beefy as a result. Perhaps someone'll come up with a lighter, smaller, XC/enduro version?

For now, check those skewers (like you should check all the other fasteners on your bike)...