There is plenty on this subject published, for example this is from Litespeed explaining speed wobbles. I fixed mine by putting a new fork on with a slightly different rake but the solutions seem to be different for everyone.
FRONT END WOBBLE-AKA SHIMMY
When it comes to high-speed down-hills, nothing can frighten a rider more than instability in the front end of the bicycle. Many cyclists have experienced this at one time or another. So what’s the deal? The issue of front-end wobble or shimmy is rarely as simple as a worn or poorly adjusted component. This kind of problem is quite common with bicycles, so I thought I’d take a few minutes to go through the most common causes of shimmy and also debunk some of the assumptions about the problem.
First, you can make ANY bike shimmy if you do the right (or wrong) things. Even a perfectly aligned, perfectly adjusted, very stiff bicycle will shimmy under the right conditions. Of course, your bike is likely none of those things, but we will get to that. Second, in the majority of cases, there is more than one thing contributing to a shimmy. Usually, two or three things will combine to start the cascade of events. But ultimately, all shimmy results from structural oscillation.
The natural structural frequency of the bicycle varies depending on the factors listed below. When they combine to reach the range of a shimmy (about 5-10 cps), the wobble can be started and sustained if the structural frequency is not changed. Note that I indicated that a shimmy CAN be started. Even when a bicycle is capable of sustaining a shimmy, it still requires an impetuous. This can be a quick shift in weight, bar movement, road surface irregularity, etc. Shimmy is rarely a “stability” problem per-se, but a structural problem. The use of the term structure in this case describes, not a structural part of the bicycle, but the entire structure including the bicycle and the rider.
Longitudinal shear is the tendency of the structure to twist (or shear) along the directional axis of the bicycle. The stiffer the builder makes the frame in longitudinal shear, the more difficult it will be to initiate a shimmy because of the higher structural frequency. This twisting (or shear) in a cyclical fashion is really what shimmy is.
Let’s go through the list of the most common contributors to bike shimmy in order of likelihood and effect:
1)The rider
2)Speed
3)Frame set Materials
4)Frame set Design
5)Road Conditions
6)Component Problems
7)Tracking and Alignment
The rider:
Although I feel a bit sheepish pointing this out, I must tell you that the most common cause of shimmy actually has to do with what the rider is doing at the time that shimmy happens. I don’t mean that the rider is messing up, but that a rider’s position or riding style can be the cause of, or contribute to, a shimmy. If a rider’s weight is too far forward on the bike, and/or the rider’s reach from saddle to bars is too short (too vertical arms), a frame will have much more force applying torsional stress on the front triangle. This is why descending or sliding forward can cause a shimmy, and sliding back in the saddle has a positive effect. The other way a rider can contribute to shimmy is when they “tighten up.” Having tension in your arm and shoulder muscles is the worst thing you can do to contribute to shimmy as it lowers the structural frequency of the structure and reduces the natural dampening characteristics of the human body. At speed, tightening up your upper body can be a natural thing to do, but it is also likely to set up a shimmy. Relaxation is likely to make that shimmy go away.
Speed:
As bicycle speed increases, so will the gyroscopic effect on the wheels. This (one would think) would have a stabilizing effect on the bike, and it does in one aspect of stability, but it is the wrong kind of effect in relation to shimmy. Gyroscopic effect helps keep the bike going in one direction and maintain a consistent attitude, but it also lowers the frequency of the entire structure, bringing the bicycle closer or to the point where a shimmy is possible.
Frame set Materials:
The materials with which a frame set is constructed can affect the structural frequency of the bicycle in two ways. First, they determine in large part how stiff the structure will be. The stiffer the structure, the higher its oscillation frequency. This is why a more flexible frame has more of a tendency to develop a shimmy. The other way that material choice can affect shimmy relates to the natural dampening characteristics of the frame material itself. Titanium and steel are the most resilient of current frame materials, while aluminum is somewhat less so and most of the composites are downright dampening. The major effect that these dampening materials can play in shimmy is not that they can really lower structural frequency, but that composites and to a lesser degree aluminum can help dampen developing structural oscillations by absorbing energy. The materials used to fabricate the fork affect the frequency of the frame too, but it appears that stiffness per-se is rarely a problem. The choice of fork material can effect the dampening of oscillations or their tendency to develop just as they can with the frame itself.
Last in the materials area is impending failure. As a tube begins to fail, the frequency of the frame set drops drastically. Indeed, impending failure has the most powerful effect on frame shear modulous and thus shimmy potential of all the structural causes.
Frame set Design:
Again, this category of shimmy contributors is quite varied, so stick with me. We’ll start with front-end geometry. The inherit stability of the front end of a bicycle is determined largely by trail. Trail can only affect stability in a seriously negative way at higher speeds where low trail (too much rake) can contribute to shimmy. Over compensating with high trail though is not the ideal solution as it makes for inconsistent handling and can cause other difficulties at lower speeds.
As I mentioned earlier, the center of gravity can contribute to shimmy, and here I will deal with a frame design’s effect on this. If a frame design positions the rider’s weight too far forward, or it leaves him with too little reach, this will have a major effect on lowering the oscillation frequency of the structure. As a custom builder, this is one of my biggest concerns when designing frames for individuals, but it should be just as big a concern to someone shopping for a new bike.
The head angle alone can also effect the structural frequency of the frame, but because this is rarely an issue for road bikes, it has only minimal effects on road frames and is pretty difficult to explain in 500 words or less, I think that I’ll leave it at that.
Even though this sounds like a materials issue, I will treat the choice of specific tubing and frame material as a design issue. As I mentioned earlier, the stiffer the builder makes the frame in longitudinal shear, the more difficult it will be to initiate a shimmy because of the higher structural frequency. Because tubing has its most efficient shear stiffness along its longitudinal axis, the most efficient way to stiffen the bike’s torsional shear is to use a stiffer top tube. A stiffer down tube helps, but not as much as the top tube does.
The weight of the entire structure also affects the oscillation frequency of the structure. To a point, the heavier bike/rider combination (all other things being equal) the more likely it is that the bike can develop a shimmy. As an example; heavily loaded touring bikes will increase their shimmy potential.
Road conditions:
Only in rare cases can road surface conditions actually be a primary cause of a shimmy. When you think about it that makes sense as road conditions can’t change the structural aspects of the bicycle. Keep in mind though that road conditions CAN initiate a shimmy that is waiting to happen. Rain groves, a bumpy surface, a small stone can all start a shimmy if the structure is in a potential shimmy condition.
**Component problems: **
Component choice, and especially wheel choice, has a significant effect on shimmy potential. I indicated that the gyroscopic force in a fast spinning wheel will lower the frequency of a bike. Well it works the same way with wheel weight and size. Any change in the inertia inherit in your chosen wheels will change the oscillation frequency of your bike. Because of this, smaller or lighter wheels will exhibit fewer tendencies to shimmy. 650-C wheels and well-built light tubulars are the most anti shimmy wheels while 700-C heavier looser wheels are the worst.
Because a tight headset increases the effective mass of the bicycle structure by partially including the fork and front wheel in the frame’s mass, it also increases the chance of shimmy. Another reason it causes the problem is that the tight headset increases the shear forces applied to the frame by not allowing the headset to absorb them. The tighter the headset, the more it affects shimmy potential. A slightly loose headset will negatively affect handling, but it will actually lower the tendency of the bike to shimmy.
In the same way that lighter wheels help with shimmy, so do lighter tires. Nuf said. Keep in mind though, higher tire pressures do not dampen potential shimmy as well as moderate tire pressure. There is NEVER a reason to run over 9 bar. in your road tires (even though I do to psyche myself up for a big race.)
Tracking and alignment:
Tracking (or wheel and frame alignment) while not a primary contributor to shimmy, can help initiate it. The way this works is that a bike that tracks poorly, for whatever reason, is almost always under shear stress. At those times when the shear stress is relieved momentarily the frame, by definition, will go through one shimmy cycle. This can then be the start of a full shimmy development.
Solutions:
Anything that will increase longitudinal shear modulus will decrease shimmy tendency as it will raise the structural frequency of the bicycle and rider. Use the above descriptions to use whatever changes you can to effect the way your bike works. This side of a stiffer frame or new fork, the other solutions in this area are pretty much limited to headset adjustment and wheel choice.
Think back to our discussion of center of gravity and act accordingly. This means that you should try to keep your weight back as far as is reasonable during descents. This also means that you should be sure that your bars are not too low or that you don’t have too short a reach.
LIGHTEN UP! What I mean here is that you should make a real effort to loosen up your upper body in situations where a shimmy might occur. Let the bike take care of itself. Of course, this can be quite difficult to do as nervousness can bring on tightness. Do your best.
Even though lowered speed will help with a shimmy, getting slowed down is just the thing that will make it worse. Putting on the brakes will shift your weight forward AND stiffen up your arms. Making a quick decision on how to deal with a shimmy in an emergency is difficult and each situation is different.
The final (and most effective) emergency tactic is to either touch the top tube with one of your knees, or to clamp the top tube with both your knees. This will dampen the shimmy cycle. It does not affect the structural frequency itself, but it does dampen the structure. The human body is a wonderful dampening device.
No Solution:
In rare cases, no specific cause can be found for a recurring shimmy. There was a case with one of our bikes where we tried essentially everything implied above and could not prevent our customer from getting a shimmy above about 32 miles an hour. Try as I might, I could not get his bike (with all of his parts) to shimmy at any speed between 0 mph and 54 mph. Although logic would indicate that the problem had something to do with him, not the bike, we never figured it out. The fact that he lives quite near Area 51 in the Southwest may be the most logical explanation.
The answer is out there …