Bent frame quizz

[GregT]

If the steering stem is not vertical (or to be accurate, in the same plane as the rear wheel) then the contact patch is offset and the trail is biased to one side of the theoretical centerline of the bike...

Result....with the bars apparently centralised, you'll ride around in circles....

As an aside. a couple of years ago I was going to the track and got passed by an MV. I noticed the rider was sitting on it a bit oddly so dropped back and got a good look from directly behind. From directly behind the wheels formed an X.....It was nearly new too.
What caught my eye was the amount of body english he was having to use to get it to go in a straight line.

[John Kat]

GregT is right about the trail.
It's the trail that makes our bikes go straight as long as the frame is not bent
If the frame is twisted i.e. when the wheels are no longer in the same plane (and appear as an X from the rear) then the front wheel will try to follow the intersection point of the steering stem axis with the ground.
If the wheel is offset to the left of the frame's centerline, the intersection point is also offset to the left and...the front wheel will try to go left and...the bike will follow making a left turn
My first conclusion after measuring the front wheel offset was identical to Agemax as when you initiate a turn, you countersteer so that the CG makes you go in the other direction.
This is off course true but not in a steady state condition.
My last question was: how come you need so much correction eitheir by "body english" or by pulling on the bars?
I believe the answer is again related to the trail.
The 10 mm offset must not be considered in conjunction with the wheelbase of the bike but in conjunction with the trail itself.
So given a trail of 110 mm an offset of 10 mm represents an angle of over 5°at the handlebar.
No wonder one can feel it...
So in summary, if your front wheel is offset to the left of the bike's centerline your bike will pull to the left

[John Kat]

surely if the wheel is still vertical relevant to the rear wheel, and pointing straight relevant to the rear wheel, then the bike should still travel in a straight line.
even though it is 10mm to one side, they are both pointing straight ahead.

also known as crabbing, in the car world

You are definitely right.
If both wheels are in parallel planes the bike will travel in a straight line but...the rear wheel will create it's own trace.
The handlebar will be slightly turned with respect to the frame centerline by approximately 10/1500 radians or .38° only.
I forgot to add that this time the steering will have to go to the right!
This is another proof that if you feel that you need to pull a lot on the handlebars to go straight it's because your frame is twisted in an X pattern.
Offsetting the rear wheel to catch the front offset will NOT change the problem...
Skewing the rear wheel in the swingarm will not help eitheir...
The front wheel follows the intersection point of the steering stem with the ground and the bike follows!
That's it.

[Nessism]

How do you know the wheel is sitting "straight" in the swingarm. You can not wheel alignment based on the hash marks on the swingarm.

What happens when you use the string method to align the rear wheel parallel to the front? What do the hash marks on the swingarm tell you? Typically, they may be slightly off which is fine. The string tells the real story.

[tatu]

Years ago when I was a mech for Suzuki I remember being told that there had been a run of GS1000s (not necessarily only 'S' models) with bad swinging arms that were replaced under warranty. I never saw or had to do any, just heard the story. My first 'S used to wear the front tyre unevenly, more wear on one side than the other, I never did get to the bottom of it. In those days the bike used to get ridden much harder than these days

[John Kat]

How do you know the wheel is sitting "straight" in the swingarm. You can not wheel alignment based on the hash marks on the swingarm.

What happens when you use the string method to align the rear wheel parallel to the front? What do the hash marks on the swingarm tell you? Typically, they may be slightly off which is fine. The string tells the real story.

In fact having the rear wheel perfectly aligned is of secondary importance as I believe I have demonstrated below.
On top of that if you align the rear wheel to a front wheel that is offset because the frame is twisted just adds misery to misery
I garantee it will not get the bike to steer straight when the frame is bent.
Guess how I know that?

[John Kat]

Years ago when I was a mech for Suzuki I remember being told that there had been a run of GS1000s (not necessarily only 'S' models) with bad swinging arms that were replaced under warranty. I never saw or had to do any, just heard the story. My first 'S used to wear the front tyre unevenly, more wear on one side than the other, I never did get to the bottom of it. In those days the bike used to get ridden much harder than these days

I heard some similar stories over here.
In fact when "Moto Journal" tested the GS 1000 ST in 1980 they stated that the bike had a handling problem.
They thought it came from the front fork?
I really believe there was an issue with the frames.
When you think about it , 10 mm of deviation at the tire level means a deviation of the steering stem of approximately 10/1000 radians or .57 degrees of arc.
In other words it means that the stem top bearing is offset by 2 mm with respect to the bottom bearing.
Not much for sure but a great difference in roadholding...and clearly a manufacturing challenge

[posplayr]

How do you know the wheel is sitting "straight" in the swingarm. You can not wheel alignment based on the hash marks on the swingarm.

What happens when you use the string method to align the rear wheel parallel to the front? What do the hash marks on the swingarm tell you? Typically, they may be slightly off which is fine. The string tells the real story.

The string method doesn't generally measure for the front and rear wheel being out of plane (i.e. the X being talked about). Short of having a large flat surface so you can measure the rims from the vertical, not obvious how to get a good measure of what amounts to frame twist.

I agree that you should not rely on the hash marks as well.

When I did my string measurements I assumed by frame was straight so I was just measured off of the tires. With a 170/60 the edge of the tire pretty much follows the rim. I measured between the front and back of the rotors on both sides but only at one level. The string goes well below the axle and comes across the disk also below the axle (IIRC)

I suppose though you could probably get a pretty good idea of frame twist if you measured both above the axle and below the axle with as much space between as you can gain. Assuming you have two true wheels, four sets of strings would measure the magnitude of the "X".

[John Kat]

I suppose though you could probably get a pretty good idea of frame twist if you measured both above the axle and below the axle with as much space between as you can gain. Assuming you have two true wheels, four sets of strings would measure the magnitude of the "X".

I agree with the principle of the method.
The issue however is that we are trying to detect an "X" factor of .57° of arc
I see an easier and more accurate way to confirm if the frame is twisted
You perform the string or laser measurement as I described previously with the "centered" rear wheel as a reference and you note the offsets on eitheir side of the front wheel.
You then take a second set of measurements with the fork legs moved up as high as feasible in the triple trees.
If the calculated offset angle is the same in both cases you are now sure the frame is twisted
PS: I read all the comments about using the rear wheel as a reference but don't forget that the investigation started because the bike was pulling to one side.
As I believe I demonstrated the bike pulls towards the side of the offset in a direction and an order of magnitude that is way beyond a misaligned rear wheel.

[posplayr]

I agree with the principle of the method.
The issue however is that we are trying to detect an "X" factor of .57° of arc
I see an easier and more accurate way to confirm if the frame is twisted
You perform the string or laser measurement as I described previously with the "centered" rear wheel as a reference and you note the offsets on eitheir side of the front wheel.
You then take a second set of measurements with the fork legs moved up as high as feasible in the triple trees.
If the calculated offset angle is the same in both cases you are now sure the frame is twisted
PS: I read all the comments about using the rear wheel as a reference but don't forget that the investigation started because the bike was pulling to one side.
As I believe I demonstrated the bike pulls towards the side of the offset in a direction and an order of magnitude that is way beyond a misaligned rear wheel.

If the engine was out, I would do all measurement together. If not the fork compression method might be good enough (probably with fork springs out).

You still trying to accomplish the same thing by measure the out of plane offset between front and rear wheel which requires 4 measurements.

I'm not sure where you get 0.57 deg. If you figure a 1mm tolerance at the tire (110/80-18 for example), I get 677 mm diameter (26.7") which is 338 mm radius. 1mm tolerance out of 388mm is 2.9 mrad or .17 deg. Are you assuming a 3 mm tolerance for error?

If you can get a 1 ft spread on the upper and lower part of the rim, then you can basically directly measure the offset at the contact patch. Unless you have 12" stroke on your forks your error tolerance will increase inversely with the reduced stroke relative to 13.3 ".

Bears repeating, this assumes some really straight wheels.

[John Kat]

I'm not sure where you get 0.57 deg. If you figure a 1mm tolerance at the tire (110/80-18 for example), I get 677 mm diameter (26.7") which is 338 mm radius. 1mm tolerance out of 388mm is 2.9 mrad or .17 deg. Are you assuming a 3 mm tolerance for error?

The .57° is the deviation of the fork stem versus the vertical plane.
As I measure 10 mm offset at the ground level and as the perpendicular length to the ground of the fork plus the wheel is around 1000 mm, the angle is approximately 10/1000 rads or .57°.
I haven't gone into an error calculation because my purpose was to understand the dynamics involved in a twisted frame