Experiments in tube drawing

 We plan to make a new radiator to the original pattern for our 1916 fire engine. 

This is a slightly unusual design consisting of vertical stadium shaped tubes (3/4 inch by about 3/32) arranged in fives front-to-back and then with corrugated sheets between them to brace and add surface area. 

We pulled out a section of the original tube and it turned out to be about 0.2mm thick (36 swg?) and to our surprise it had clearly been folded out of brass sheet with a soldered seam. 

Asking a radiator specialist, this is apparently "lock seam tube" and not currently favoured. 

After a bit of CAD measurement it seems that the ideal tube to reproduce this would be 13.5 mm diameter, which isn't a readily-available size, especially not in 0,2mm wall thickness. 

However 17/32 x 1/32 tube is readily available from model shops as part of the K&S range of telescoping imperial tube. Readily available, but nit very cheap at £6 for a 12" length. 

I decided to start experimenting with this, and was slightly stumped on how to machine them until it occurred to me to split the die. I CNC milled them into some silver steel. 

I managed to hammer some tube far enough through the die to get started, then spent ages trying to think of ways to hold the end and to pull hard enough to draw it. I didn't want to go too far down the rabbit hole of making draw benches and specialised grippers until I knew whether it looked plausible and had some idea what forces were likely to be involved. 

Eventually I had the idea of just silver-soldering my swaged and crumpled bit of tube into a slot in the end of some threaded rod. 

Then I could use some bits of scrap tube, redundant parts etc to make a the Pugh Super Janky Drawbench™ 

As expected by several of us, the tube collapses into a figure of 8 shape. 

Which would work, but isn't what we were after. I think that this is possibly solvable with a floating mandrel, though. 

Anyway the main purpose of this test was to prove that we have the capability to make the tooling and to measure the drawing force required. 

Using a luggage scale on the end of a 19mm spanner I measured 3kg @ 21cm, or close enough to 6Nm. 

Using an online bolt calculator that suggest a tension force of 2.5kN to pull the tube through the die. 

So I will use that as the basis for the drawbench design.  This falls inside the normal working load of bicycle chain, so that will probably be the operating mechanism used, at least initially. 

According to the internet:

"for a standing person operating a rotating wheel or crank arm - the length of the arm (or radius of the wheel) should be approximately 0.4 m and maximum force should not exceed 130 N"

So that suggests a working radius at the sprocket of 20mm, which matches a 10 tooth sprocket. 

This is smaller than I would really like to be using, so there may be a need for a primary reduction.