Bicycle Mechanics - My bike pulls right to my wits end.

About 6 weeks ago I was riding my ’86 Trek 310 ,

http://i254.photobucket.com/albums/hh117/auchencrow/TREK%20310/TREK310.jpg

and got hit by a car at about 40-50 MPH, as it crossed into my lane coming around a curve. – (It was sort of a side swipe really, but the car still had to be towed because I blew out the left front tire, tore the outside mirror off and pushed the fender back into the driver’s door jamming it shut). My bike fared pretty well – the front wheel zipped by the car without ever making contact, and the only apparent damage to the bike was a twisted handle bar, and a pedal that was torn out of the left crank. . . Or so I thought . . .

After the accident, I brought the bike to my LBS for an estimate. The mechanic verified that the frame was not bent using the frame gage, and since the Insurance was paying, I decided to have them effect the repairs, and do a general tune up.

When I got the bike back, it looked good as new, but when I got to riding it, I found that the bike mysteriously pulled to the right.

I went to the bike forums immediately for an answer, and found a thread that described a case where a pair of fork blades had not been bent back, but had been displaced laterally about ½ inch, causing the bike to pull right.
http://www.bikeforums.net/showthread.php?t=79546

I went into the garage and using a bicycle plumb bob (made from string and a broken presta valve) I saw that my bike ALSO had both fork blades laterally displaced ½” to the left.

I then made a crude alignment gage from a 2 by 6 and two U-bolts, establishing a centerline with a known, good fork, and my laser level.
Next, I gradually bent each of the forks to perfectly align +/- 50 MM to centerline, using a small square, and a section of PVC pipe, (See pics)

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Forkgage.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Forkgage004.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Forkgage002.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Forkgage005.jpg


I reassembled the bike and tested it. No improvement!
NEXT -
I verified that the front and rear wheels where true and centered between the stays.
I verified that the front and rear axles were seated in the dropouts, and the QR’s secured.
I verified that the headset bearing cages were lubed, and not installed upside down.
I replaced the old caged bearings with 2 new bearing cages ( I am all out of loose ball bearings)
I verified that the headset was properly adjusted, without excess play or binding.
I added an additional spacer in the head set, just to be sure the headset nut was not bearing down on the threaded tube.

STILL no improvement! The bike pulls right.

I don’t know what else to do. Is there something I am over looking? Any suggestions are appreciated.

I don't understand how a bike shop would not find the bent fork. From the pictures the fork is severly bent and probably not repairable. Now I'm wondering if the frame is bent. Instead of working on it yourself you need to get another bike shop to check it out and file a claim with the insurance company.

Al

If the steerer tube is bent, even slightly, the bike can pull strongly. If the fork blades were bent, then it could be the tube is also bent.

I don't understand how a bike shop would not find the bent fork. From the pictures the fork is severly bent and probably not repairable. Now I'm wondering if the frame is bent. Instead of working on it yourself you need to get another bike shop to check it out and file a claim with the insurance company.

Al

Hi Al1943 -
The fork was bent almost exactly 1/2", which sure is a lot - but surprisingly on the built up bike, it was really hard to see - not obvious at all as it is when a fork is pushed rearward. - (in fact without the aid of a plumb bob, I doubt that you would not ever notice. So in defense of my local bike shop, I can see why they missed it. In any case, both the forks arms are now perfectly aligned (at least as well as my known good fork, taken from a Miyata 710).

BTW I too suspected the frame but had it measured at the LBS, before committing to repairing it in the 1st place. I did collect $200 from the drivers' Insurance company, but basically that just covered what we knew about, not this latent problem.

first of all, congrats on surviving the crash and being victorious over the car.

secondly, you appear to be a better mechanic than most LBSes (which you probably confirmed for yourself).

lastly, have someone else ride your bike and see if they get the same pull. maybe they can help see what the issue might be. it could be that your bike always pulled left and now it's fixed! or, maybe you suffered a slight injury you aren't really aware of...

If the steerer tube is bent, even slightly, the bike can pull strongly. If the fork blades were bent, then it could be the tube is also bent.

Hi Mos6502 -

I can guarantee that the forks are aligned +/- 50mm to the steerer tube's nominal centerline, but if the steerer tube is bent slightly, I coudl overlook it. - How could I measure it? - And by bent, do you mean curved along the length? or abrubtly bent where it joins the fork crown?

Do you suppose spacing the axle downward in one of the 2 fork dropouts would test for this condition? - Or maybe I should swap the threaded fork assembly with another bike to test it ? - I have a Bianchi that is about the same size - not at all sure it would fit though .

first of all, congrats on surviving the crash and being victorious over the car.


Thanks illwafer! - but it's always a Pyrrhic victory when you win an argument with a car, so DON"T go trying that yourself!


... maybe you suffered a slight injury you aren't really aware of...

illwafer - I'd liketo subscribe to that theory, except I have a bunch of other road bikes in my garage that DON'T do it, even when I ride them hard with no hands.

I would not be at all surprised if the steer tube and head tube are bent.

Nice job with the photos.

Al

I would not be at all surprised if the steer tube and head tube are bent.


- Do you know a way to measure it? ?
If not I'll try the fork from my Bianchi. (It would look kind of goofy but I would be ecstatic just to get to root cause - I'll know how to fix it then).


Nice job with the photos.

Thanks Al1943 !

Since the shop missed the forks being offset it would not surprise me if their checking of the frame is highly suspect as well.

You did an excellent job of making up a jig for measuring and straightening the forks. This displays that you have a pretty good handle on the basics and how to set up and measure angles and extend lines via squares and plumb lines to check this stuff. I'd go ahead and do the same sort of treatment on your frame looking for a possible twist in the main triangle or a similar offset in the rear stays.

I was going to say that if the crown top is flat just use a straight edge along the tube to find out. Obviously that's not going to work here. I guess you could hang a plumb from a bolt in the brake hole, then level the crown out upside down and see if the tube centers on the plumb.

It's probably be easier to just go with a new fork.

I think there are several more possible issues. To run straight, the bike needs to have the wheels both in the same plane, and the main triangle and the center of gravity in that plane. The BB axis needs to be parallel to both of the wheel axles. Some alignment issues could be a subtle result of the crash, the result of your years of use, or have been latent in the frame from the factory. Trek did not always align everything perfectly back in the day.

1. The seat tube might not be straight, or perpendicular to the BB axis. The best way to see this is on an alignment table, but if you can see with a straightedge that the seat tube is warped, this problem exists.

2. The head tube might be twisted out of plane with the seat tube or the BB axis. You may be able to see this if you can sight the head and seat tubes from the front or rear. It might be important to strip the bike to some degree to be able to see better.

3. There could be a distortion in the rear triangles that has the rear wheel laterally out of plane with the main triangle.

4. A rear triangle distortion could cause the rear wheel to lie in a plane that is twisted with respect to the BB axis. My Trek 610 had this problem, and my aligner milled inside the dropouts to ensure the rear axle would lie parallel to the BB axis.

It could also be that your fork is still not right, despite your careful approach.

The most effective way to resolve all of this at once is to have the entire frame aligned at once. I had it done by a nearly local expert framebuilder, Ron Boi in Kenilworth, Illinois. His prices are on his website, http://rrbcyclesusa.com/. Framebuilders have to align hand-made frames before delivery, so they know the methods in a way your LBS might not. Besides, I think an alignment table is necessary to see some of the problems and know when the frame is right.

I had my Trek 610 aligned by Ron and may send him some other frames. I also had my Masi aligned, and have been amazed by how much improvement both bikes showed.

... I guess you could hang a plumb from a bolt in the brake hole, then level the crown out upside down and see if the tube centers on the plumb.


Thanks Mos6502 -That sounds like a good method -I'll try it


It's probably be easier to just go with a new fork.

I don't want to go there, unless (a) I first verify the steer tube is bent and (b) find a suitable replacement Chromalloy 27" fork with the proper tube length, CHEAP.

Since the shop missed the forks being offset it would not surprise me if their checking of the frame is highly suspect as well.

You did an excellent job of making up a jig for measuring and straightening the forks. This displays that you have a pretty good handle on the basics and how to set up and measure angles and extend lines via squares and plumb lines to check this stuff. I'd go ahead and do the same sort of treatment on your frame looking for a possible twist in the main triangle or a similar offset in the rear stays.

Thanks for your vote of confidence in me, but I'd venture the experienced bike mechanic at the LBS still has the edge on me, and I'm pretty certain what he told me about the frame is accurate. But if my fork swapping experiment does not pan out, I may probe the frame thing deeper as you say.

It seems that you may be able to make use of a rotary laser level to do some of the measurements you need, especially if you wish to check the frame. Just a thought. You would need a good one to be accurate. The cheap ones are only good to 1/4 inch over 20 feet- not so good.

Hi Road fan -
and thank you for your rather comprehensive post: I am going to try to rule out the fork steerer tube before proceeding with a deep dive into frame alignment (beyond what the LBS did) - but I think your little treatise gives me an undertanding at least, of what I'm looking for.

Not sure I'm ready to take the TREK out to to Kenilworth Il just yet - but I think I will give them a call to ask about the frame alignment option, if it comes down to it... I don't have a lot of money to spend, but this was a really good bike, and I feel like I "owe" it something, because I was not killed or permanently injured on it. (I know it's silly of me to personalize a bicycle. but I'm probably not alone in doing that).

It seems that you may be able to make use of a rotary laser level to do some of the measurements ...

Great minds think alike!

Thanks for your vote of confidence in me, but I'd venture the experienced bike mechanic at the LBS still has the edge on me, and I'm pretty certain what he told me about the frame is accurate. But if my fork swapping experiment does not pan out, I may probe the frame thing deeper as you say.From your post and pictures, I'm pretty sure your fork is fairly straight now. Here's a Park Tool - Frame Alignment Gauge (http://www.parktool.com/products/detail.asp?cat=48&item=***-2) that I suspected the shop used on your bike. While it tests to see whether the dropouts are in-plane with the seat-tube, it does nothing to test the alignment of the head-tube with the seat-tube.

In a lot of non-head-on crashes, especially those that result in the fork being bent sideways, the headtube ends up twisted as well. While its center may still be in-plane with the seat-tube, the top and bottom are twisted to opposite sides. That is, the seat-tube and headtube are no longer parallel laterally. What this does is put the front contact-patch of a straight fork off to the side. One way to fix this is to actually bend the fork in the opposite direction by 1/2".

Well, I tried to substitute the fork from my Bianchi, but found that the fork bearing races were incompatible with my Trek's head tube bearing races.

I did check out the alignment of the steerer tube on my Treks' fork though, and with the laser and my rigged up fork alignment fixture (see pic), I confirmed that the tube is aligned within about 1/16" of the centerline running between the two forks.
(I feel confident that the fork is not reason why the bike pulls right.)
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Rearstays2.jpg

I also did my best see if the frame was bent. The best method I came up with was simply laying an aluminum channel flat against the frame's triangle, and measuring the distance to rear drop outs on one side, and then the other. (see pic) There was less than 1/8" disparity from right to left.

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Rearstays1.jpg

I think though, I just succeeded in verifying what the LBS had measured (probably with the very Park gage cited by DannoXYZ. . . )

Based on what I’ve eliminated thus far, I think both Roadfan and DannoXYZ may be on to something: i.e., The head tube and seat tube might not be in the same vertical plane. - I suppose it wouldn’t take much twist, projected out to where the rubber meets the road to cause big problems, but it's hard to fix what you can't see. (Maybe if I get a long 1" bar to insert through the head tube, I would be able to see it better?)

Bending the forks another 1/2 inch in the opposite direction to compensate might help some, but I do have to say that bending the first 1/2 inch back to centerline did not ameliorate the pulling condition much if at all.

I tried to call Road Fan's frame builder today to diiscuss it, but I got no answer at the phone listed on the website. - But the problem does seem to beg an expert with the right equipment: It's getting a little bit beyond me.

Thanks everyone for your insights.

I don't understand how a bike shop would not find the bent fork. From the pictures the fork is severly bent and probably not repairable. Now I'm wondering if the frame is bent. Instead of working on it yourself you need to get another bike shop to check it out and file a claim with the insurance company.

Al

Looks like he's doing a fine job on his own.

I wonder if you could fix the frame with level on the seat tube, then set the level on the head tube to compare. If the 2 fixed cups in the head tube are the same diameter, then you can set the level on them. You may not be able to notice a small problem, but a glaring one may show up.

Have you checked to see if one of the fork blades hasn't been bent back. place a hub in the forks and use your laser to look at the fork from the side. aim the laser straight down the axle of the hub and see if the fork crown is in alignment.

Looks like he's doing a fine job on his own.


True, my concern is that the insurance company would not pay off after the owner had worked on the bike, but would more likely take the word of an uninterested third party bike shop.
I agree that the OP is doing a very fine job of analyzing the problems.

Have you checked to see if one of the fork blades hasn't been bent back. place a hub in the forks and use your laser to look at the fork from the side. aim the laser straight down the axle of the hub and see if the fork crown is in alignment.

Yup- I've checked it a few ways, and both blades are absolutely aligned in the F/A direction relative to the fork crown.

Well, I say notify the insurance company that there is additional damage that was not visible during the initial inspection and that you are due a equivalent functional replacement. Do it quickly or it will be too late. I would be flustered after all this effort and would want to just end the whole thing. But this may be a fun challenge for yourself and better than getting a new frame.

Your photo of the fork and laser-beam seems to show that the fork-blades are bent off to one side... Was this picture taken before or after you bent it back? Or it could just be parallax error due to the camera being off to one side.


I think though, I just succeeded in verifying what the LBS had measured (probably with the very Park gage cited by DannoXYZ. . . )

Based on what I’ve eliminated thus far, I think both Roadfan and DannoXYZ may be on to something: i.e., The head tube and seat tube might not be in the same vertical plane. - I suppose it wouldn’t take much twist, projected out to where the rubber meets the road to cause big problems, but it's hard to fix what you can't see. (Maybe if I get a long 1" bar to insert through the head tube, I would be able to see it better?) Yes, you've done the same test as the bike-shop with your gauge. What we need to do is compare alignment on three points. The front dropouts to the bottom-bracket shell to the rear-dropouts. We've confirmed somewhat that the rear dropouts are close to aligned to the same plane as the BB, now we just need to do the front.

1. re-install the fork on the bike, adjust the bearings for smooth rotation with no play

2. make an alignment table from a flat board about 5-ft long, draw a straight line down the exact centre.

3. one one end, attach a tall V-block about 70mm high to hold a spare rear-axle. Screw some nuts/cones onto the axle to position the axle centered on the board (130 or 135mm apart depending upon your frame's OLD).

4. around where the BB-shell lays, screw down some right-angle brackets facing each other 68mm apart. The BB-shell should fall right between the brackets if the rear-dropouts are aligned.

5. the fork will be free, you can make another V-block with axle to ensure it's rotated 90-degrees to centreline. Then measure on this axle where each fork-tip lands.

... this may be a fun challenge for yourself and better than getting a new frame.

True - and the other thing is that I have so personalized the bike, that I want to help it "recover" from its injuries :)

... Or it could just be parallax error due to the camera being off to one side.

Exactly.


...What we need to do is compare alignment on three points. The front dropouts to the bottom-bracket shell to the rear-dropouts. We've confirmed somewhat that the rear dropouts are close to aligned to the same plane as the BB, now we just need to do the front.....

Thanks Danno, you have given me another way to go just as I was running out of ideas!
- A

Frame alignment isn't only about the frame and fork, but must be such that both wheels lie in the same plane, with neither vertical nor horizontal twist. It's the position of the wheels that determines tracking.

If you can find a 5' straight edge, you can do a fairy accurate tracking test. First check that the rear wheel is dished correctly and that both wheels are properly mounted without tires. Be sure that the rear wheel is centered between the chainstays. Lay the straight edge along both wheels to make 4 point contact with the rims about 4-6" from the ground. If the wheels track correctly you'll get solid 4 point contact with the front wheel straight. Slide the edge up and down, it should stay in perfect 4 point contact.

If you can find a second straight edge, narrower enough (or appropriately notched) to clear the forks, do the same test both at the tops and bottom of the wheels at the same time looking for 8 point contact. This is the acid test, but harder to do because of the fork and seatstay so you might have to pass on it.

Sometimes you can get the 4 point contact only by canting the rear wheel slightly between the chainstays. If that works and is still fairly close to the center, note the position and build the bike that way and try riding to see if it tracks better.

If you can't establish solid 4 point contact the frame and fork are mis-aligned. It could be rear triangle, head tube twist, or a bent fork. Note that the fork must not only be centered side to side, but the axle position must be equally forward or the centerline (rake), and the dropouts equally high so as to hold the wheel vertical and not twisted. Similar requirements apply to the rear triangle.

BTW- An entire set of gauges can be made with shorter straight edges, which take the plane of the rims, and a laser pointer to extend the line back to front. This is how automotive tracking is checked.

You need to find a shop or frame maker with a set of frame alignment gages.

Putting your location in your profile might make it possible for someone to recommend one.

Well, I say notify the insurance company that there is additional damage that was not visible during the initial inspection and that you are due a equivalent functional replacement. Do it quickly or it will be too late. I would be flustered after all this effort and would want to just end the whole thing. But this may be a fun challenge for yourself and better than getting a new frame.

+1

If they give you crap...(oh wait..its an insurance company..not 'IF, but 'WHEN they give you crap), just print out this thread (with the pics of course). Drop it in mail to the agent. Shouldn't have any problems then.

FBinNY’s method to check 4-point contact with the wheel rims to validate frame alignment is a good trick to know – but in this case, I already made a homemade frame alignment gage, along the lines that DannoXYZ had suggested: The point was to project a line defined by the centers of the rear axle and the bottom bracket shell. The projected line ought to dissect the front forks down the middle. . .

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/FrameAlignmentgage1.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/FrameAlignmentgage3.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/FrameAlignmentgage2.jpg

- But it missed the middle by ¾” ( 20 mm) - see pic below. So the frame is bent: The head tube is twisted counterclockwise just slightly, as viewed from the front of the bike. The other part of the problem is that the bottom bracket shell is displaced to the right
(This makes sense, since the vehicle made whacked the top of the left crank before ripping out the left pedal.)

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Off3quarters.jpg

Now my dilemma is how to straighten the frame: I would like to fix it myself – so I suppose could peddle East on I-75 North, (but even my Insurance Company would balk at that).

Does anyone know a better method? – or if worse come to worse, the name of a good frame builder within reasonable distance of the Motor City?

So the BB-shell is off-centre to the right and the fork-tips end up on the left? Wow, that's actually two different bends in the frame...

Recentre rear-triangle in-plane with seat-tube and BB. This will put your BB-shell back in-line with your alignment board. The previous alignment test with the straight-bar and caliper to each drop-out was imprecise because it assumes the head-tube and seat-tube are in alignment with each other.

1. Use the BB-bracket as centre. I would replace the right-angle brackets with beefier units about 5mm thick with gussets to ensure the right-angle. Then drill two tight holes for the BB-spindle. Slide them over the BB-spindle with their legs facing outwards and bolt down tightly. Now you've got an origin/zero point for reference.

2. Re-check the rear-dropout alignment by measuring their relationship to centreline. With precision V-block and axle, you can measure the dropout distance to the middle of the axle. I bet the rear-dropouts are off by little. You can re-verify using a 2nd test using a string. Clamp one end to the outside of the left dropout and run it forward and around the center of the head-tube and back around to the outside of the right dropout. Clamp down QR. Measure the distance between each string-run to the seat-tube. The difference between the two sides is 1/2 the misalignment of the dropouts. Use Sheldon Brown's 2x4 method to bend the rear-triangle laterally to center in the plane of the main triangle.

3. Align head-tube with seat-tube. To undo the counterclockwise twist, you need to clamp the seat-tube securely to a bench. Make V-blocks roughly the length of the seat-tube and clamp/bolt it down it to a VERY sturdy bench with the head-tube sticking out (BB on left, seat-post on right). The V-blocks should be tall enough that the rear-triangle is above the bench; you don't want anything touching the rear that you've just fixed. Then clamp the down & top-tubes near the seat-tube with V-blocks as well. This fixes the seat-tube in place and lets you twist the head-tube.

Then get a 5-8ft section of water-pipe that barely fits into the head-tube. Shim it up so fits very tightly. Fit the pipe into the head-tube from the bottom up so that it's goes all the way through and flush to the top of the head-tube. Make sure you have it shimmed evenly and tightly all the way through. Then grab the other end of pipe and lift up, you can even use a hydraulic floor-jack pushing up on a 2x4. This twists the head-tube in the opposite direction as the original impact.

I'm not sure how you would verify how much bending and twisting is necessary though. Maybe the laser can be used somehow. Or you can use a camera with big-zoom from across the room. Position the camera so you can see the edge of the seat-tube and head-tube. Move the camera so that those two lines are very close. Then verify that they are parallel.

So the BB-shell is off-centre to the right and the fork-tips end up on the left? Wow, that's actually two different bends in the frame...

Recentre rear-triangle in-plane with seat-tube and BB. This will put your BB-shell back in-line with your alignment board. The previous alignment test with the straight-bar and caliper to each drop-out was imprecise because it assumes the head-tube and seat-tube are in alignment with each other.

...Re-check the rear-dropout alignment by measuring their relationship to centreline. ... measure the dropout distance to the middle of the axle....re-verify using a 2nd test using a string.... run it forward and around the center of the head-tube and ...Measure the distance between each string-run to the seat-tube...Use Sheldon Brown's 2x4 method to bend the rear-triangle ...

...Align head-tube with seat-tube. ... clamp the seat-tube securely to a bench. Make V-blocks ... clamp/bolt it down it to a VERY sturdy bench ...V-blocks should be tall enough that the rear-triangle is above the bench...get a 5-8ft section of water-pipe that barely fits into the head-tube... twist the head-tube in the opposite direction as the original impact. ...



Thanks DannoXYZ - I am familiar with Sheldon's 2x4 method for the rear stays. Relative to the head tube twist - I think I get the picture (below) , but I lack the fixture details and the means to make them. I will have to improvise then... Maybe I can build a wooden truss to hug the subject frame tubes, and anchor it to a wall to react against. - Even at that, I'll need a miter saw (I was looking for one on CL anyway). . . and I'll need a few days to show any progress.

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/Frametwister.jpg

One of the problems with aligning based on the BB shell, is that it's possible for a correctly aligned frame to have a slightly canted BB, especially based on the external BB shell references. Try using the BB face as a reference, but understand that a very tiny error in fixturing (or in the actual build) can create the illusion of large mis-alignments throughout the frame.

Assuming you have access to a truly flat table, (harder than you think) you can make a shaft that will take the line of the head tube, and cradle it in V-blocks, then shim the rear triangle to the correspondingly correct height so the frame is suspended on a plane parallel to the table. This will allow accurate measurements of seat tube alignment (aka head twist). Mount a 10mm shaft perpendicular into a base. Use this to check for squareness of the axle position (height of the dropouts). Don't forget that if there's any curvature or twist in the table's surface, it'll be magnified in any readings taken from it.

Only when you have a true picture of where the errors actually are, and the least bending that will establish alignment should you make any adjustments. Otherwise, you'll be doing more than necessary and possibly introducing new and more complex errors.

Note, if the head tube axis is exactly perpendicular to the rear wheel axis, (on plane with the wheel) with the seat tube on the center plane (though not necessarily in plane) the bike will track correctly with a straight fork. I make this point because squaring the head tube to the seat tube takes major brute force, and then if the rear triangle must also be adjusted to the seat tube, that'll be more work than simply verufying the tracking.

As I said, get a clear 3 dimensional total picture before you pretzel this frame even more.

One of the problems with aligning based on the BB shell, is that it's possible for a correctly aligned frame to have a slightly canted BB, especially based on the external BB shell references. Try using the BB face as a reference, but understand that a very tiny error in fixturing (or in the actual build) can create the illusion of large mis-alignments throughout the frame.

Assuming you have access to a truly flat table, (harder than you think) you can make a shaft that will take the line of the head tube, and cradle it in V-blocks, then shim the rear triangle to the correspondingly correct height so the frame is suspended on a plane parallel to the table. This will allow accurate measurements of seat tube alignment (aka head twist). Mount a 10mm shaft perpendicular into a base. Use this to check for squareness of the axle position (height of the dropouts). Don't forget that if there's any curvature or twist in the table's surface, it'll be magnified in any readings taken from it.

Only when you have a true picture of where the errors actually are, and the least bending that will establish alignment should you make any adjustments. Otherwise, you'll be doing more than necessary and possibly introducing new and more complex errors.

Note, if the head tube axis is exactly perpendicular to the rear wheel axis, (on plane with the wheel) with the seat tube on the center plane (though not necessarily in plane) the bike will track correctly with a straight fork. I make this point because squaring the head tube to the seat tube takes major brute force, and then if the rear triangle must also be adjusted to the seat tube, that'll be more work than simply verufying the tracking.

As I said, get a clear 3 dimensional total picture before you pretzel this frame even more.

+1 I was going to post something along these lines, but you said it well already.

One of the problems with aligning based on the BB shell, is that ... a very tiny error in fixturing (or in the actual build) can create the illusion of large mis-alignments throughout the frame....

Hi FBinNY -
You are right in saying that any slight error on the BB face would project out to a large error, but I think DannoXYZ was directing me to center the "REAR" triangle (stays) relative to the centerline of the BB, (not the whole frame). The beefy right angle blocks against the faces of the BB were just to ensure it doesn't walk when I reef on the stays with the 2x4's, per Sheldon's method. (At least that is the way I understood it.)



Assuming you have access to a truly flat table.....

Stop there! - For sure I don't have any table that is even remotely flat enough for these purposes. (I do not even have any V-blocks, etc. I just have a minimum of basic tools, and not too much experience.


....Only when you have a true picture of where the errors actually are, and the least bending that will establish alignment should you make any adjustments. Otherwise, you'll be doing more than necessary and possibly introducing new and more complex errors....

There is that risk - But I think I have a reasonable chance of establishing the vertical axis for the seat tube and head tube. (and I am guessing it accounts for at least half of my problem).
- I hope the Sheldon 2x4 method will take care of the remaining misalignment, but if - as you suggest, the axis running through the drop outs is not perpendicular to the seat tube, I'm SCREWED.
( I think you phrased it more gently: "that'll be more work than simply verifying the tracking."

PS - I'm waiting to get a call back on a miter saw. It's the minimum I will need it to rig a fixture to hold the bike (as well as the occasional household task).

Sorry, my earlier post wasn't as much to give you a how to, but a why not to, explaining the risks of using the BB shell and seat tube as the points of reference and to keep you from getting in over your head, or unnecessarily damaging your frame.

I strongly doubt that there's any head tube twist resulting from the accident. If the head and seat tubes were OK with respect to each other beforehand, the accident wouldn't have had nearly enough force to change that without causing far more damage than seems to be the case.

Here's a fairly simple way to check. Fixture your bike in a horizontal plane. Use a bubble level set on the seat tube to get that exactly level. Now check the head tube. While the frame is there, mount the rear wheel without a tire, lay your channel across the rim vertically (parallel with the seat tube) and see if it too lays in plane. BTW, if the head and seat tube aren't both level at the same time (head tube twisted) relevel the frame by the head tube when checking the rear wheel.

A similar method can be used to check that the front wheel is true in the fork. Cut the other end of your homemade fork fixture so the crown overhangs and mount it in a vise. Clamp the fork, mount the fixture in a bench vise using the bubble level to get the steerer perfectly level. Mount your tireless front wheel and with the channel and level across the rim perpendicular to the steerer, and rotate the fork until the wheel is level. Now see if it's also level parallel to the steerer. If it isn't it's because the wheel axle isn't perpendicular to the steerer (one dropout is higher than the other). You correct this by tweaking the curve of one of the blades, or if it's only a tiny amount, filing the inside of a dropout.

Hi FNinNY -
I think the bubble level approach you describe is doable for me, and will at least IMPROVE my prospects for establishing whether the seat and head tube are truly in the same vertical axis, as well as determining if the dropouts (front and rear) are perpendicular to that axis.

I do believe there is a good possibility that the force of the accident was enough to induce soem misalignment though. (It was enough to bend a forged left pedal crank).

Thank you for the tips -your experience is showing, and I appreciate it.
.. I am going to Lansing MI for the day, I hope to dive back into measuring Tuesday given your advice.

This is my set up with 3 wooden v-blocks (made with my new Miter Saw). I used a quarter ream of scrap paper to level them, with an old Stanley level positioned across the frame tubes three ways: (1) seat and top tube (as shown), (2) seat and down tube, (3) down tube and top tube. The level was calibrated to a second level so I have faith that the level was accurate (or both levels are off the exact same direction!)

I laid a threaded rod across the bare front rim.
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6200.jpg

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6201.jpg

When I rotated the front wheel the rod intercepts the bottom bracket about mid way. This is a hint that the wheel is shifted left relative to the BB (given that the rod is laying on TOP of the rim, not passing through its center).
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6191.jpg

I noted that the REAR wheel is about in the same horizontal plane as my level. (I used the top edge of the Stanley level to position my laser in the horizontal plane).
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6199.jpg

When I rotated the Front wheel so that the threaded rod was vertical (in bike position), I saw that it was off by more than an inch (~28mm) as measured from top to bottom. - This is very consistent in both magnitude and direction with what I saw on the other alignment fixture, which I had made per DannoXYZ's instruction.
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6192.jpg

BTW - With the frame leveled in this manner and the fork set perpendicular to it, the twist was evident to the naked eye.
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6194.jpg

=> I would conclude that the bottom of the seat tube was pushed hard right, by merit of the vehicle hitting the left crank.
=> It evidences itself as twist in the head tube that projects out to roughly one-inch of vertical misalignment in the front wheel, from top-to-bottom.
=> The REAR wheel alignment looks reasonably OK in the vertical plane, so one would conclude that the rear triangle moved WITH the seat tube (more or less, given my ability to measure.)

. . .My next challenge is how to Un-twist the frame.

=> I would conclude that the bottom of the seat tube was pushed hard right, by merit of the vehicle hitting the left crank.
=> It evidences itself as twist in the head tube that projects out to roughly one-inch of vertical misalignment in the front wheel, from top-to-bottom.
=> The REAR wheel alignment looks reasonably OK in the vertical plane, so one would conclude that the rear triangle moved WITH the seat tube (more or less, given my ability to measure.)

. . .My next challenge is how to Un-twist the frame.VERY good analysis and test!!! What happened during the impact was that the tyre's contact patches on the ground stuck them in place. When the car hit the bike at the crank/BB area, this bent the centre of the frame to the right. The fork-blades, being an untriangulated cantilever arm ended up being bent to the left the most. This also twisted the head-tube counter-clockwise as you've determined. To a lesser degree, the rear-triangle was also bent to the left. Or perhaps the direction of the impact was such that the rear-wheel was lightly loaded and most of the force went into the BB-shell and front-end.

With your new alignment jig, you have some actual numbers you can measure. Use some geometry calculations given the displacement of the indicators you've set up to determine how much the head-tube is twisted relative to the seat-tube.

I think you're on the right track. The drawing of your clamped frame looks good. You may want to use a big bolt, washers and internal-spacer to clamp the BB securely to the bench-top. Then support the entire length of the seat-tube with V-blocks. The end of the pipe above the head-tube you may want to clamp to the bench as well. Then use a hydraulic floor-jack to push up on the free end of the pipe coming out of the bottom of the head-tube.

You can shine the laser through the head-tube (or any light actually) and project it onto a wall of known distance away. This can be used as a gauge to tell when you've bent back the head-tube by the required number of degrees.


It's actually possible to align the just two wheels so they are parallel in the same plane, or in parallel planes with a bent-up frame in between (BB not centered in these planes). The bike will still track straight, but will lean to one side to place the C.O.G. in between the two contact-patches on the ground.

VERY good analysis and test!!!

Thank you!


... Or perhaps the direction of the impact was such that the rear-wheel was lightly loaded and most of the force went into the BB-shell and front-end..

Very likely, considering all the forward momentum.


.... The drawing of your clamped frame looks good. You may want to use a big bolt, washers and internal-spacer to clamp the BB securely to the bench-top. Then support the entire length of the seat-tube with V-blocks. The end of the pipe above the head-tube you may want to clamp to the bench as well. Then use a hydraulic floor-jack to push up on the free end of the pipe coming out of the bottom of the head-tube. .

This will stretch my fabrication skills to the limit -(and my wife won't let me start on it until I have her new PC configured) but I hope to have something constructed in a couple/few days.



You can shine the laser through the head-tube (or any light actually) and project it onto a wall of known distance away. This can be used as a gauge to tell when you've bent back the head-tube by the required number of degrees.

Super idea.


...
It's actually possible to align the just two wheels so they are parallel in the same plane, or in parallel planes with a bent-up frame in between (BB not centered in these planes). The bike will still track straight, but will lean to one side to place the C.O.G. in between the two contact-patches on the ground.

I see why it is advisable to clamp the top portion of the head tube.


I have to ponder this a little bit. I think I might build a box crate around the raw frame to trap the essential datums on the frame, and to provide a reaction surface for the bending bar. As long as I can preserve the line of sight for the light shining through the head tube onto a wall at a known distance (as you suggested), I will have a means to effect the required change in angle without grossly overshooting the mark.

Finally, we have arrived at the end of the story. . . Here ‘tis:

I constructed the bending fixture from 2x4’s, building it around the actual bike, and trapping the top and bottom of the set tube, as well as the top of the head tube. The bottom of the head tube had just one degree of freedom, to the right.

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6204.jpg
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6205.jpg

I taped on a $3 laser pointer that I bought at Office max. To avoid excessive scratching, I also taped up the metal brackets you see adjacent to the head tube, intended to obviate any loading deformation of the wood.

http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6206.jpg

Here the laser is tuned on. It was pointed at a starting reference point on my garage door.
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6207.jpg

After reefing on the frame with all I had, I managed to effect a 3” vertical displacement of the red laser dot at 10 feet. (This calculates to less than a 2 degree change in angle for the head tube, and was just a little less that the target number I had calculated from my little trigonometry exercise.).
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6208.jpg

I then disassembled the fixture and reassembled the bike, adjusted the DR’s and brakes, aired the tires and made sure the seat post was plenty greasy. (I will not countenance a stuck seat post!)
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6212.jpg

The result?

- Not all I had hoped for, but significantly better: The bike no longer pulls to the right, and it tracks straight - but I noticed that the Center of Gravity is off, because the whole bike leans to the left when riding no-hands. I surmise that the BB is a tad to the right of the plane of the wheels –(not surprising given the cause, and earlier cited as a possibility by Dannoxyz).
,
http://i254.photobucket.com/albums/hh117/auchencrow/Bike%20Forums/DSCF6213.jpg

OK – I couldn’t get it perfect, but at least I can ride it again. - Not a bad outcome for a near death experience for my old Trek-310 and me.

So thanks to all on the BF who contributed to this little repair effort, and especially DannoXYZ , and to FBinNY. I would not have gotten off the dime without your advice.

- A

Good job! It's wonderful to see classic bikes revived to stay on the road! :)

auchencrow, thank you for pointing me to this thread. I never noticed it before.

I have to admit I didn't follow everything you guys were saying. I'm impressed you got this far. How bad is the ride now?

Did you ever tell the bike shop what you learned and what you did? They may be interested.

If this ever happens to you again, would you do it all the same way?

I have two frames/forks to straighten now. I can't possible get into the sophistication you guys here show. I hope they are simpler. I've done a few, but nothing like yours.

That rig you buillt around your frame? I'm pretty sure I saw something like that in a book about the Spanish Inquisition... :D

An excellent thread all around. You may still not think you did anything special but with a little help from us and a lot of your own efforts you both analyzed the issues to a fine degree and formulated and carried out what you needed to effect the corrections. All in all I rate you as a formidable frame guy. Someone with the right jigs could have done this easier and quicker but with no higher a degree of thought and consideration.

This must be an extremely valuable bike 'cause if I were sideswiped at 50 mph I wouldn't be messing around trying to fix the bike. I'd simply have the driver's insurance company buy me a new ride and let them deal with the damaged goods. Time is money.

Auchencrow,

I know that it's late now ( just saw this thread), but Assenmachers in Swartz Creek does frame building and repair. Had them align the fork off my Ross Mt. Hood and was very happy with the service. Very customer oriented shop. Price guide is on their website.

http://www.assenmachers.com/index.htm

Walt

Oh my - I missed this thread the 1st time around. I admire the OP's determination but he went to MUCH more work than usually necessary to find the problem (and still did not fix it entirely). The fork has to be right in 5 ways:

1. Fork blades equidistant from fork column front center.
2. Fork blades equidistant forward from from fork column SIDE centerline.****May have been overlooked in the above case.
3. Fork dropouts parallel to each other.
4. Fork blade side centerline aligned with fork column side centerline (not always true.
5. Fork column straight.
Only the 1st 3 would tend to affect leaning.

In addition the head tube needs to be parallel to the seat tube, which could of course affect leaning/turning.

Most of this can be determined fairly well by eye and using two straightedges. I know because that's how I did it for many years. Of course to straighten correctly I used Park tools and guages, but any decent frame-knowledgable mechanic can tell you where the problem lies in about 15 minutes, except for the fork column being bent. It's sad you could not find a shop to checkit correctly.

I'm on lunch break or I might describe the whole process but in brief one sights along each side of properly dished and mounted wheels along the bikes track (wheel to wheel) and down toward the main frame tubes, and then uses a straightedge at the fork crown and dropouts to check for the blades being equal. Dropouts parallels can only be checked easily with a guage of soem sort.