Larger diameter top tube than Seat Tube

[Micah Lee Vestal]

I've been looking into this. One thread I found on another forum suggested that there are two ways to ovalize the tubing if you have a larger TT than ST. The first was to miter first and then ovalize. The second was to ovalize and then miter.

Which is better or more "correct" and what is the best method for ovalizing (i.e., what tools and techniques do you suggest)?

[echelon_john]

bench vise. strong coffee. mark axis. big squeeze. then miter.

if you want to get fancy, you could use wedges in the bench vise to try to taper the compression a bit.

[fortyfour]

Personally, I ovalize, trim the excess off (which also releases tension at the very end of the tube after ovalization) and then miter. If you miter then ovalize, depending on how much you need to deform the tube, you will most likely need to clean that up with a second miter as ovalizing the tube will deform the miter.

The top tube on the frame I'm currently working on has a 1.25" top tube and 1.125" seat tube. The ovalized bit is very subtle:

[sonny]

ask mickey
it's alu, but he is familiar

[Micah Lee Vestal]

Thanks! I really appreciate the answers.

One more question. Once I ovalize the tube, how do I make a template for the miter (since I don't have a mill and such, I will have to miter with a file)? Do I just use the same calculations as if the tube wasn't ovalized? For instance, if I am joining a 31.8 TT to a 28.6 ST, do I just use those measurements for my miter template? It seems like it should work properly, but I want to make sure I'm thinking about this the right way. (btw, I'm planning to use the mitering tool on Nova's website)

[echelon_john]

One trick that I've used in this situation is to put a tube scrap of the same size as the tube you're mitering to use as a guide for a cut line. You won't get perfect this way, so you'll want to start long, but it's workable. Another option would be to use a lug to mark scribe lines (again, leaving them a little long) before you squeeze the tube; the top and bottom won't change by very much, so if you cut/file to those scribe lines, you'll be in the ballpark, and will have to go a little deeper on the sides.

Since I do virtually everything by hand, I keep lots of lugs around to use as templates even if I'm going to squeeze tubes. They give me a starting point, and I work down to actual miter size/length.

I'll say that this is also a case--especially top tubes--where having your vise setup close to your jig setup is really helpful. Fitting tubes up by hand require lots of back and forth, and you don't want to be walking back and forth across the room between swipes with the file.

Finally, I find it really helpful to have a sharpie handy to mark the area that needs metal taken away; I do this repeatedly throughout the process. Again, the point is to make it so that when you go from jig to vise to file, you're actually working on the correct area. Even when you've down it a lot, moving and reorienting the workpiece can make this hard to remember/easy to screw up.

[Mark Kelly]

A normal tube junction is the intersection of two cylinders which is easy to calculate and plot. When you "ovalise" you make one tube an elliptic cylinder.

An elliptical cylinder is like an ordinary cylinder sliced obliquely into many many layers and then these layers restacked; imagine cutting lots of thin diagonal slices off a baguette and then stacking them vertically. If the baguette was round, the stack will form an elliptic cylinder.

This suggests a solution to the problem: First calculate the equivalent angle of the diagonal slices, which will be the the arccosine of the ratio between the minor and major axes of the ellipse. Say you have squished a 34.9mm tube down so that the short axis is now 31.8mm. The long axis will be approximately 38mm. The angle is thus arctan (31.8/38) ~= 33 degrees.

The intersection of the two tubes will be equivalent to the intersection of the actual seat tube and a notional 31.8mm tube offset by the required angle, so if your sta is 73 degrees the intersection will be the equivalent of a 31.8mm tube at 40 degrees (= 73 - 33). Of course if you use one of the mitre calculators to print a paper template that template will be for the notional equivalent cylinder, not for your elliptic cylinder.

I think you could get around this by angling the template to the tube by the offset angle when aligned to the major axis: fix a thin rigid support to the centre axis of the template, place the front edge on the tube at the major axis, lift the back edge until the template is at the correct angle to the tube and then carefully wrap to fit. I will try this today to see how it goes.

[Mark Kelly]

Since I posted the above I've been thinking about the much more difficult problem of a tube "ovalised" in the opposite direction (eg a flattened top tube on a cross bike) and I think I have an approximate solution.

It relies on the concept of an equivalent angled cylinder as per the above but since the offset angle and the tube juction angle are on orthogonal planes, they can no longer be summed. On the other hand the offset angle acts like a linear offset to one side of the junction axis. In this case the offset is equivalent to the sine of the offset angle multiplied by the radius of the tube to which the "ovalised' tube is joined. This will hold for points outside the axial offset only, for points inside this offset an approximation to a cylinder of diameter equal to the major axis of the ellipse should be adequate.

You will thus need to print three templates and cut and paste them together. I'm sure this is as clear as mud without illustrations so at some stage I'll try to get something sorted.

[Micah Lee Vestal]

A normal tube junction is the intersection of two cylinders which is easy to calculate and plot. When you "ovalise" you make one tube an elliptic cylinder.

An elliptical cylinder is like an ordinary cylinder sliced obliquely into many many layers and then these layers restacked; imagine cutting lots of thin diagonal slices off a baguette and then stacking them vertically. If the baguette was round, the stack will form an elliptic cylinder.

This suggests a solution to the problem: First calculate the equivalent angle of the diagonal slices, which will be the the arccosine of the ratio between the minor and major axes of the ellipse. Say you have squished a 34.9mm tube down so that the short axis is now 31.8mm. The long axis will be approximately 38mm. The angle is thus arctan (31.8/38) ~= 33 degrees.

The intersection of the two tubes will be equivalent to the intersection of the actual seat tube and a notional 31.8mm tube offset by the required angle, so if your sta is 73 degrees the intersection will be the equivalent of a 31.8mm tube at 40 degrees (= 73 - 33). Of course if you use one of the mitre calculators to print a paper template that template will be for the notional equivalent cylinder, not for your elliptic cylinder.

I think you could get around this by angling the template to the tube by the offset angle when aligned to the major axis: fix a thin rigid support to the centre axis of the template, place the front edge on the tube at the major axis, lift the back edge until the template is at the correct angle to the tube and then carefully wrap to fit. I will try this today to see how it goes.

Since I posted the above I've been thinking about the much more difficult problem of a tube "ovalised" in the opposite direction (eg a flattened top tube on a cross bike) and I think I have an approximate solution.

It relies on the concept of an equivalent angled cylinder as per the above but since the offset angle and the tube juction angle are on orthogonal planes, they can no longer be summed. On the other hand the offset angle acts like a linear offset to one side of the junction axis. In this case the offset is equivalent to the sine of the offset angle multiplied by the radius of the tube to which the "ovalised' tube is joined. This will hold for points outside the axial offset only, for points inside this offset an approximation to a cylinder of diameter equal to the major axis of the ellipse should be adequate.

You will thus need to print three templates and cut and paste them together. I'm sure this is as clear as mud without illustrations so at some stage I'll try to get something sorted.

Thanks Mark! I will have to think about this for a minute before I fully understand but I appreciate the response (I never took geometry so some of the terms are foreign to me). I really look forward to hearing about your test. I don't have a tube set yet so I can't really test anything yet myself.

[Mark Kelly]

Warning: error in equation.

Should read arcCOS (31.8 / 38) ~= 33 degrees.

Damn stupid 10 minute limit.

[Micah Lee Vestal]

I understand the reason that the tube mitering templates wont work now. It's easy to see what happens by just rolling up a sheet of paper into a cylindrical shape and then squishing it into an elliptical cylinder. If you were to draw a miter template on the paper before squishing, the angle of the miter would become steeper when you squished it (not sure if I worded this in a way that makes any sense).

[Mark Kelly]

OK, I've tried this out and it kinda works but it's so error prone that I think it's basically useless.

[Micah Lee Vestal]

Sounds like I will have to do it by hand like echelon suggested! Somebody much smarter than I am should create a tube mitering program that can create templates for oval tubing :)

[Chauncey Matthews]

You're handicapping yourself by relying on templates. You won't need them when you've made enough practice joints.

[Micah Lee Vestal]

You're handicapping yourself by relying on templates. You won't need them when you've made enough practice joints.

Don't get me wrong, I want to be able to do it by hand. How do get the angles right?

[Pi Guy]

Don't get me wrong, I want to be able to do it by hand. How do get the angles right?

A decent protractor and piece of scrap tubing will go a long way.

[RyanHeller]

Micah,

Sorry I missed your question you posted about this in my build thread. The way I did it was to ovalize the TT to the ST diameter and then draw my miter. To do that I printed off a tube miter template that was intended for the ST diameter (approx. 29.3mm). I then laid this out on my now ovalized 31.8mm TT in the correct phase with my HT miter. Now the template obviously won't fit all the way around because it was printed for a 29.3mm tube. Disregard this and draw out the miter, we will fix it next. The miter at the end where the two edges of the paper template meet is now slightly too large, while the other side of the miter is slightly too small. Fix this by redrawing the miter with a marker to fix the size issue. This will get you close enough to the miter the tube and measure with a proctractor and machinists square to make sure everything is correct. Always miter less than you need and constantly check and recheck your fit. I hope this helps.

[Micah Lee Vestal]

Micah,

Sorry I missed your question you posted about this in my build thread. The way I did it was to ovalize the TT to the ST diameter and then draw my miter. To do that I printed off a tube miter template that was intended for the ST diameter (approx. 29.3mm). I then laid this out on my now ovalized 31.8mm TT in the correct phase with my HT miter. Now the template obviously won't fit all the way around because it was printed for a 29.3mm tube. Disregard this and draw out the miter, we will fix it next. The miter at the end where the two edges of the paper template meet is now slightly too large, while the other side of the miter is slightly too small. Fix this by redrawing the miter with a marker to fix the size issue. This will get you close enough to the miter the tube and measure with a proctractor and machinists square to make sure everything is correct. Always miter less than you need and constantly check and recheck your fit. I hope this helps.

Thanks Ryan, I really appreciate it! I think I finally have a good idea of how this will be accomplished.

[feldybikes]

One trick with miter then ovalize: miter to the diameter of the larger tube, then squish till it mates up to the small tube. (e.g. 28.6 ST and 31.8 TT: miter at the same "depth" as the 28.6 ST, but use a diameter of 31.8 for the miter) No need to deal with multiple templates -- though that is clever. You might need to clean up a bit if you don't ovalize on the proper axis, but it'll be really close.

[Eric Estlund]

I'm not sure if I'm reading this correctly- are you trying to miter an oval tube to fit another oval tube? If so, stop. Save it for later when you feel better about filing miters repeatably and without mental stress.

If you are filing an oval tube (say a top tube) to fit a round tube (say a seat tube) then don't worry about the oval tubes cross section. The tube being mitered to (the round one) needs a round hole, regardless of the shape of the intersecting pipe. Figure out your miter depth to miter depth length and file away.

At the end of the day, bike tubes are just bits of metal. Feel free to bang on em, bend em, squish them, file them, cut away and generally have at in an effort to get what you want. But before you drive yourself to nuts, build a few standard combo bikes and get the process worked out. Then go mental.