SLIGHTLY IRONICALLY, THE FIRST MOTORCYCLE I EVER OWNED HAD A RIGID FRAME AND A PAIR OF GIRDER FORKS WHEN IT CAME MY WAY.

Not too long after it’d been transformed into a runner, not because I was a mechanical prodigy as a child, but because my Uncle Greg had a spare engine in his coal shed. It got modified with a pair of telescopic forks, and a swinging arm rear, and my dad did pretty much all of that as my mum wasn’t too happy about the idea of my pre-teen self wielding a hacksaw, and he wasn’t too happy about the idea of me messing with his welding gear, and using up gas and rods without producing any actual welds. Still, it did leave me with a bit of a fascination for tubular steel motorcycle frames.

Having eventually figured out the welding part, the biggest obstacle I could see to frame building was the bending tube part. To be clear about that, not that tube bending was the only obstacle, but that it played into everything else that you needed to do. For instance, what’s the point of having a jig that accurately positions the steering head, engine, and mount for the spindle for the rear wheel or swinging arm, if you can’t accurately bend the tube to connect the parts? It’s relatively easy to hardtail a steel tube frame – you don’t need a specialised jig as the swinging arm provides a reference point, and simply bolting the wheel into the axle plates, and tacking them

to the swingarm, will have pretty much solved your jigging problems there (Fig.1). Approached sensibly, adding a stretched hardtail, single top tube, and even a little bit of rake to a standard steel tube frame isn’t all that troubling to do (Fig.2) – it’s selfjigging for the most part (unless you start with a bent frame); none of the tubes need more that one bend in them and, as long as both sides end up looking the same, the angle of the bends’s not that critical. Apart from the fact that they can be hard work to use, the Record/Hilmor-type conduit benders I’ve been using lately are perfectly adequate for the job (Fig.3).

However, once you start building complete frames, things rapidly get more complicated. Things like widetyre Harley Big Twin frames (Fig.4) not only need much more accuracy in the degree of bend, but the plane of the bend needs to be measured, too. The emphasis on the word ‘needs’ in that last sentence will vary quite considerably because steel tube is quite a forgiving material to work with, and minor mistakes can often be tweaked out with a little experience. The obvious thing about that is that by the time you have a reasonable amount of experience, you’re probably not making big mistakes… or at least you’d hope not. When things get over-complicated then it’s a good idea to make a dummy, like the one in Fig.4 and, if necessary, cut and weld that to get it to fit before using it as a pattern to bend a real one. Even then, though, once you get past using pieces of tube with just one bend in them, it becomes necessary to have some means of measuring the plane of the bends you’re making. Consider the simplest scenario I can think of right now; a pair of drag bars, which’re nothing more than a length of tube with a couple of bends in it. The first problem is that the bends need to be identical, and evenly spaced about the middle of the tube. Given that the bender has an angle stop, then making identical bends is easy, and measuring from the ends of the tube to make a couple of marks to align with the former (Fig.5), so the identical bends are identically

spaced, isn’t rocket surgery either. The handlebars are still going to turn out unusable if the bends don’t lie on the same plane, and what that means is that if you placed the hypothetical handlebars on a flat surface, they wouldn’t rock around. If you place them on a flat surface, and the middle section, and one end, lie on the surface, but the other end sticks up in the air, then the bends aren’t in the same plane, and the handlebars are going to feel weird. When the tube’s bent, it’s going to follow the plane of the former it’s being bent on, so it’d be possible to place a level on the former, and record the deviation from the vertical, make the first bend, then set up for the second bend, and use a sheet of material placed on the first bend to make sure that it was aligned with the plane of the former, before making the second bend.

That doesn’t seem like too much trouble as long as the bends’re all in the same plane, but what happens when you have to make a right and left-hand version of the same thing, and the planes of the bends are rotated by 11 degrees to each other? Do you need to add the three degrees that the former is off the vertical, or take it away? Personally, I prefer to level the bender, and simply flip an adjustable bubble around to measure each tube with the same tool on the same setting, which avoids stupid arithmetic errors, and keeps the error in whatever you’re using to measure the angle of the plane from accumulating on you.

This, at long last, brings me to the CX500 mono-shock seat rails that I was bending last time. Having made the bend in a length of tube, I didn’t need to worry about the plane of the bend because there was just the one. I cut the tube overly long, bent it, and then drew around it on a piece of masking paper taped to a board (Fig.6), and flipped it over on the outline, and cut the ‘leg’ that overlapped the outline down so the two legs were that same size. The

resulting ‘U’ shape wasn’t doing it aesthetically (Fig.7), so I decided to put a bend in the ‘legs’.

Now if you read what went before, you’ll see there’s a problem in aligning the bends in the legs, not only with each other, but with the bend at the back of the ‘U’, too. A quick check showed that the top face of the bender frame wasn’t too far off level, but placing the level on the former showed it was a mile out (Fig.8). Rather than get too involved, I tried using a G-clamp to squeeze the former up against the frame of the bender, and that got it nearly vertical, and a 3mm shim under the ends of the legs got it spot on, according to the level.

After eyeballing where I wanted the bend to be, I used the reference bend to mark the bend line (Fig.9) to position the tube in the former, and then measured and marked the other leg to match.

That let me position the tube in the bender, and use a flat piece of material across the bend at the end of the ‘U’, and a level, to make sure the bend was at 90 degrees to the plane of the former (Fig.10). I used a piece of quarter-inch rod that I’d bent to the angle I was looking for (Fig.11) to eye up the bend and, once I’d bent the first leg, I lined up the bend mark on the second leg, set the loop up so it was level, and again used the quarter-inch rod to eye the angle up. Eyeing the bend is a bit of an acquired skill, but I’m planning on modifying the bender to make it a bit more suitable for this kind of work, and bend repeatability is on the list of things I want to address when do that, but in the meantime the bender does have a stop.

With the loop propped back into position on the motorcycle, I was happy with the way it looked but, before I can weld it on, I want to make the seat rails, and that means sorting out the rear mount for the tank, which I don’t have the rubbers for yet.