Front end shimmy, trail and weight distribution

[Vincentsavary]

I went touring last week and my bike did something it never did before: front end shimmy. The bike without any load feels a tad twitchy (but not even enough to be considered an issue) and putting the slightest load at the front usually fixes the “issue” and makes it very stable and it traces really well. Because of this, I tend to load it up more at the front than the rear when going on tour. But last week, for the first time I was a bit heavier at the rear than the front, and the shimmy appeared. Nothing crazy, but enough to be annoying. To be clear, by “shimmy” I mean the steerer having a hard time keeping straight, and doing slight but fast left-and-right after, say, hitting a bump or riding one-handed. On the second day I shifted a little bit of the weight at the front and it helped, but didn’t fix the issue. Now I’m wondering what is causing this? Could it be because the frame has a relatively low trail for a loaded bikepacking bike? I would think that the center of gravity in relation to the wheelbase might have something to do with it too? To me, instinctively it makes sense that rear load affects steering, but I just never really thought much about it and can’t really visualize how that works. I consider it a minor inconvenience, I won’t do anything serious about it, but I’m just curious to understand the phenomenon. Any insight would be appreciated!

Here are the numbers I assume are useful in this situation:

HTA 72deg
Rake 50
Trail 63
Tire size 650b x 2.3”

STA 73
Chainstay 440
Front center 627
Wheelbase 1070

[hampco]

What's the frame material and, if steel, any idea what the tubing is? Under-sized top tubes can cause or add to shimmy, imhe.

[Vincentsavary]

It is indeed steel, fully Columbus Cromor, which I would think is adequate but I'm also not a framebuilder

[Mark Kelly]

Have you tried pumping the front tyre up to a higher pressure?

[hampco]

What Mark said - and is the headset spinning freely, not brinnelled? This was interesting: https://en.wikipedia.org/wiki/False_brinelling

Cromor is pretty beefy but if it's a 1" top tube that could be the culprit.

[Vincentsavary]

I checked the headset, no loose and spins freely, and it's a White Industrie headset so I would assume they use good quality bearing that wouldnt cause false brinelling (but interesting article indeed). I also checked the front hub and brake, and everything was fine. Front tire pressure wasnt particularly high but still higher than I normally run. And top tube is 1-1/8".

I could be missing something but I'm pretty sure the shimmy doesnt come from a mechanical issue, mostly because it never did this before and doesnt do it when un loaded.

[e-RICHIE]

A crown in the road, or any small object unseen, can cause a momentary lapse in the bicycle's innate desire to continue in a straight line. This can be unnerving if (IF) you don't expect it when it happens. Magnify it out depending on the severity of the conditions and or speed and - volià - a small shimmy sets up.

This isn't a design issue.

[Vincentsavary]

I've road this bike in every condition, from pavement to singletrack, without ever experiencing this type of shimmy. That's why I'm wondering if the weight distribution might have to do with it, since I was mostly riding on pretty smooth Vermont pavement. I'm not saying the design is inadequate, but I was thinking it might have some undesired behaviour when loaded a certain way.

[e-RICHIE]

maybe a side-view photograph of the bicycle would help talking points.

[Rody]

Nothing wrong with the design numbers.

Too much weight over the front wheel can cause the contact patch of the tire to engage farther forward with the pavement, causing a "shopping cart" type of shimmy.

It's important to remember that a bicycle is designed for the rider weight to be positioned between the ground contacts points for optimum performance. When you begin loading up weight directly over the front or rear axle, it changes behavior.

You found the endpoint of additional weight the bike can take over the front contact patch...back it off now to where you experienced predictable performance results.

Need to carry more? Talk to a builder about a bike specifically designed for your anticipated use and loads.

Best wishes,

Rody

[Vincentsavary]

Nothing wrong with the design numbers.

Too much weight over the front wheel can cause the contact patch of the tire to engage farther forward with the pavement, causing a "shopping cart" type of shimmy.

It's important to remember that a bicycle is designed for the rider weight to be positioned between the ground contacts points for optimum performance. When you begin loading up weight directly over the front or rear axle, it changes behavior.

You found the endpoint of additional weight the bike can take over the front contact patch...back it off now to where you experienced predictable performance results.

Need to carry more? Talk to a builder about a bike specifically designed for your anticipated use and loads.

Best wishes,

Rody

I'm not quite sure I understand what you mean... maybe my phrasing wasnt clear but the problem only appeared when I put more weight at the rear than usual (and less at the front than usual). But I guess the idea is the same, I just need to better divide the load and there shouldnt be a shimmy?

[Andrew R Stewart]

If shimmy was truly understood no one would design a frame that would, or load a bike so it could. It took 4 tries, (over about 30 years...) for me to end up with a bike that didn't shimmy when carrying a self contained touring load. Larger diameter main frame tubing and steerer were the biggies IMO. Andy

[Rody]

I'm not quite sure I understand what you mean... maybe my phrasing wasnt clear but the problem only appeared when I put more weight at the rear than usual (and less at the front than usual). But I guess the idea is the same, I just need to better divide the load and there shouldnt be a shimmy?

Front or rear, the principal is the same. You are correct, divide the load within your performance parameters.

As for the frame, folks have loaded toured on K-mart bikes to the most expensive customs and made their destination. The untold story is the performance and pleasure they experienced along the way.

Andrew is correct, stronger, larger diameter tubing yields a more predictable ride for expedition bikes...Columbus Chromor does not fall into that category.

Run what ya brung and have fun!

Rody

[steve garro]

I've road this bike in every condition, from pavement to singletrack, without ever experiencing this type of shimmy. That's why I'm wondering if the weight distribution might have to do with it, since I was mostly riding on pretty smooth Vermont pavement. I'm not saying the design is inadequate, but I was thinking it might have some undesired behaviour when loaded a certain way.

It's just that you hit the sweet spot WRT it's vibration tunes to a certain resonance and that has to do with frame material and load - it's unlikely but also reasonable that the front to rear weight bias has something to do with it, or it may not, but this is what you are experiencing, I know it well from loaded touring, I had a Bontrager that was just awful with weight on it, you could get the front & rear to sway in two different planes if you shook the bars real hard.

- Garro.

[Mark Kelly]

It's just that you hit the sweet spot WRT it's vibration tunes to a certain resonance and that has to do with frame material and load - it's unlikely but also reasonable that the front to rear weight bias has something to do with it, or it may not,
- Garro.

To be clear, the tuning is of the loaded frame's resonant frequencies to the steering feedback frequency, the latter being largely determined by the rider.

Yes placement of the load matters as it affects the spring constant k in the resonance equation ω = SQRT (k/m). A longer spring of the same construction has a lower k.

[steve garro]

To be clear, the tuning is of the loaded frame's resonant frequencies to the steering feedback frequency, the latter being largely determined by the rider.

Yes placement of the load matters as it affects the spring constant k in the resonance equation ω = SQRT (k/m). A longer spring of the same construction has a lower k.

That's a lot of word salad to try to explain what many thousands of miles of loaded touring has taught me but OK.........I guess?


- Garro.

[Mark Kelly]

The point was that everybody so far in this thread has missed the fact that the rider is an essential part of this problem because they are part of the steering feedback loop. All the experience in the world won't help if there's a part missing from your conception of the problem.

You compared the problem to the Tacoma Narrows bridge failure which is both a furphy and irrelevant: the bridge did not fail due to simple mechanical resonance* nor is bike shimmy a simple resonance problem.



* It was probably aerostatic flutter https://en.wikipedia.org/wiki/Aeroelasticity#Flutter

[Andrew R Stewart]

Mark- Your point on the rider being an important factor is very much my experience. I have made well more than a few frames for myself and many have shimmied for me, some worse than others. Back when I was more frustrated by this I asked a few others to ride my various bikes WRT shimmy (especially when riding no handed). Almost no other rider could get the bikes to do what I call shimmy. I have also ridden more then a few bikes when working the LBS side of this stuff, looking for reported shimmy with more often than not, not experiencing the issue. So I very much believe that the rider is a big aspect.

The trouble I have is to then go that next step and be able to make a bike that will never experience shimmy for a given rider. The body is not an easily defined entity and goes through a wide range of change during a ride. If only mechanical dimensions are considered (stuff like femur length VS foot size) we miss where's the rider's mass is centered. Both of the aspects of flexibility and medical conditions need to be considered, hard to do with a number. IMO the majority of published data about fit and frame design has nearly no "anti-shimmy" discussion.

So if you have a magic formula you want to share... Andy

[Mark Kelly]

Andy

There is no magic formula.

If you have look at Andy Ruina and Jim Papadopolous's excellent work on steering dynamics Linearized dynamics equations for the balance and steer of a bicycle: a benchmark and review
(warning: very maths heavy) they show that a bike is necessarily unstable in some speed ranges.

As is well known, leaning causes steering and steering causes leaning. The paper also shows that there is a time lag associated with this; this seemingly minor point is actually crucial.

The instability of the bike causes oscillations around the neutral point which are controlled by the rider's steering input. We thus have an unstable system with a feedback loop and a time lag which is a classic problem in control theory: in the absence of sufficient damping there will frequencies at which the time lag causes a enough phase shift* that the feedback reinforces the instability instead of correcting it.

Altering any of the oscillation frequency, the feedback lag or the damping in the system will change this behaviour. Usually the most effective solution is to increase the oscillation frequency (increase k or reduce m in the word salad above) since damping is more effective when the frequency is higher. Increasing damping always helps: Cane Creek make a headset https://canecreek.com/product/viscoset/which incorporates viscous damping for this very reason.


* these are the same thing expressed diffferent ways: as an example a 0.1 second time lag is a 180 degree phase shift for an oscillation that occurs at a frequency of 5 Hz: 5 Hz = 5 cycles of 360 degrees = 1800 degrees per second so 0.1 seconds is 180 degrees.

[Andrew R Stewart]

Mark- I do understand pretty much all you say. I have followed this stuff for a few decades. I first met Jim P back in the early 1908s and have sat in on his seminars. Bicycling Science has been on my bookshelf since the first edition. Other early influences were Mullet's Mechanics (a column in a long ago English publication), Eisentraut's class and Bill Boston's neutral steering geometry (and the castor angle thing) work.

I really didn't think you had some mathematical or dimensional formula, but had hoped for some general guidelines that I didn't already know. I guess there's not much new under the sun:) Thanks for your replies. Andy