correcting ( or compensating for) a frame twist.

I've just stripped my modified 1965 Atlas frame back to bare metal to repaint, repair old accident damage, and check alignment.
Level checks across the top rails behind the steering head (where the fuel tank sits) confirmed the centres of the front engine mounts are aligned to these and the steering head bearings. I used this as my main datum.

An alloy rod through the front engine mounts was another datum to measure levels and various dimensions front to rear like swingarm, rear shocks, engine and axle centres. Shocks position, engine centre and sprocket alignment was spot on in both axis and on the frame centreline.

Sighting back along the upside down frame revealed old crash damage, a twisted swingarm tilting the wheel along with corresponding wear to the tyre. I realigned the new swingarm pivot and axle to each other with the axle slots the same distance from the pivot centre for accurate chain adjustment

A dry fit with rear wheel, swingarm and shocks fitted to frame confirmed everything was now in alignment.

This particular frame wasn't too bad considering the mileages and accidents over the years. No cracks. Frame tubes appear very good and the metal still thick judging by the tap test and looking up through the drain holes. 

Despite the anomalies some of which appear normal as Dave Degens suggested,  I never noticed any adverse handling or odd behaviour at any speed, solo or two up fully loaded. Great frame.

hello

i had a featherbed framed bike it was all over the place, measurement wise. the tls brake front wheel was over to the o/s by 5mm. with 7 inch forks. i rectified that with a commando oval type slider, on the brake side.  this gave it a central position , well i assumed it would.

i then found the front wheel spindle bent which i have had 3 second hand replacements in the past all bent in the same place. the pinch bolt side .i thought it was Nortons way of levelling things up. you could move the wheel from central position over 5mm either way just bye rotating the front wheel spindle. anyway i renewed the spindle and that cured that one.

then onto the rear wheel. it was offset to the o/s by 4mm or so. thinking it was the wheel spacers . i renewed those, still the same. i levelled the bike with spirit level and plumb bobs . i eventually found it was a twist in the swinging arm and not a side to side adjustment. so i replaced the swing arm, that cured that one.

i then marked up a centre line running through the frame. from the front wheel spindle through to rear wheel spindle centre, frame spacer, plus headstock centre ect.  then i used one of those laser levels, that shows a laser line. i run it down the centre of the marks on the frame from the front to rear, all lined up as they should.  when i built the bike it bike went down the road ok.  you could loosen your grip on the handle bars and it road straight as a die. if you did that before it was anyones guess where you would end up.

barry

These original frames were made in jigs with clearances accounting for variations in tube OD, lug ID and a host of other problems the frame makers (Norton??) had to deal with, so if you think of the basic geometry of a Featherbed frame, the critical alignment dimension from steering head to swing-arm pivot could be several mm out in x, y and z.  Oh yes, and they were hand-brazed.

The question is, did this really make a difference? 

Today, I could take my never-crashed 99 Slimline frame and get it measured on a CMM, and they would likely tell me it's scrap, but clearly that would be a 2023 answer to 1960 manufacturing.

That's a good excuse to align it yourself, 0.23” off centre is not right from a supposedly professional company when you can get it way better without fancy jigs.

My 1965 Atlas frame is steel MIG or arc welded, no brazed joints found. Gas or arc by me for various mods done in 1976.

Final checks show under 1mm misalignment between steering head and all the centres mentioned in my earlier post. Steering head is also 90° to frame, swingarm and engine datums. Everything is level with each other after reworking the swingarm and pivot.

Some original frames may be much more accurate than you think despite all the variables Russ mentions. Most bent bits I found were from my own crashes but nothing like Barry's shocker. Bulldozer involved?

It's lucky for us that all our roads are as flat as a billiards table so we regularly ride hands off for miles...??