Before I start, just to answer Spynappels question. I use standard butane/propane mix for all my small torches, and standard propane for the large burner.
Now onto this massive post for such an easy little job. This is definitely a 'takes longer to show than do'.
First off, I need to make a new drive nut for the tacho. So the correct sized drill (7/16") for the 1/2" BSF thread was put into the end of the bar to the depth that was needed.

This is a bad pic because I couldn't get the camera to focus on such a small part. What it is, is an undercutting tool, and this was fed into the bottom of the hole and a small area was relieved. This is normally done if you are single pointing an internal thread, to allow an area for the tip to run into. But in this case, Stew had loaned me the correct sized tap, and I wacked a thread in with that.

With the thread cut, I machined a small 'washer' face and give the thread a good lead in chamfer.

In this shot, you can see the undercut I mentioned.

And the 'nut' fitted perfectly onto the thread on the engine.

For the nut and drive to be made perfectly concentric, I required a true running thread, the same as the one on the engine, to be made.
In this instance, I mounted up some mild steel, and got to work. This mandrel will not be removed from the chuck until all machining operations are completed on the nut.

It is well over a year since I have done any single point threading, and this is the first time this lathe has done any at all. Plus, because I am using a metric lathe to cut an imperial thread, I will not be able to disengage the leadscrew, but reverse the lathe back to the start point each time. So this is a bit of a learning curve for me as well.
I am going to be doing it 'by the book' for my first attempt.
First off, turn the blank down to the required OD of the thread, in this case 1/2".

Then a small spigot was turned on the end that is the same diameter as the root of the thread, 0.420" (the thread depth is 0.040" for 1/2" BSF), then the spigot was shortened slightly so that it would not interfere with the nut being screwed onto the thread.

A recess for the tool to run into was made to the same depth as the root diameter.
That is the metal prepared for cutting the thread, now for the machine.

Unlike American (unified) or metric threads, BSF uses 55 degrees as a thread form instead of 60 degrees. I need half of that figure, so 27.5 degrees.
I put my topslide (compound) in line with the crosslide, and then swung it back by 27.5 degs. You remove this figure from 90 degs, NOT add to 0. So 62.5 on a normal 90 deg scale. If you had a 60 deg angled thread, then you would set your topslide to 60 degs (90 - 30= 60).
On my machine, I have put extra index marks around the crosslide top to allow me to do this indexing a lot easier. If you have a 90-0-90 or 0-360 scale, then you will be able to do this on your normal index mark.

With the topslide now set at the correct angle, then the toolpost was slackened off and the tool tip set using a threading gauge (slightly different to an American 'fishtail' gauge, which doesn't have the BS angles on it). Once set, the toolpost was locked up solid, and the angle rechecked.
As you can most probably gather, I am using the offset angle method to do this threading, not the direct feed in method.
I am playing about with my lathe here, and if I can do it this way, which is that hardest way, then I will have no trouble using other methods.

First off, I touched the tool onto the thread top OD using the crosslide dial.

The crosslide dial is then zeroed. Thru the whole threading operation, this dial will be either wound back to retract the tool, or after retracting, it will be wound forwards to this zero setting. All cuts will be put on using the topslide.

With the crosslide at zero, the tool was fed forwards using the topslide until the tool touched on the spigot for the thread root diameter. The topslide dial was then zeroed.
So if I now only use the topslide for feed, when the crosslide dial is at zero and the topslide dial is also at zero, the thread should be at the correct depth.

The machine change gears and dials were set for what the chart told me to do.

To show better what I am doing, I blued up the bit of steel.
With the crosslide at zero, I brought the topslide in until it just touched on.

At that setting, I engaged the half nuts and took a cut. The half nuts will now stay permanently engaged until the full thread is completed.
This 'swipe' was checked with my thread gauge to make sure things were OK.
It was now a case of retract crosslide, reverse the machine to wind things back to the start, forwards on the crosslide to zero, put a bit more cut on with the topslide and take another cut, then repeat from the beginning of this sentence.
When you start to get close to the zero on the topslide, you start checking with the nut for fit.
What I do, when very close, I do no more topslide cuts, just minute cuts with the crosslide. This cleans up the threads. When very, very close, I put no more cut on at all. Just let the tool go down the thread a couple of times at that setting. You will find that this removes minute amounts from the thread and really cleans them up.

Thread cut.

Nut fits just fine.
I did in fact have a die for this thread, kindly loaned to me by Stew, and I was going to use it to finish this thread off. But then I thought, lets see if this lathe can cut an almost true thread, and it looks like it does. Confidence in cutting an accurate imperial thread on this metric lathe has just gone up by 1000%.

The little root depth spigot was turned off and the thread was given a nice chamfer on the end. I now have a perfectly true running thread to cut the nut on.

Oh! s£RF^&**(
I cut a bit too close to the nut end, and the end fell off. That is just me being too confident and in a rush to get the job finished.

But as usual, I had a setup spare in my back pocket, and it was duly finished off to obtain exactly what I wanted.

The engine couldn't be checked out until this nut was made.
Unfortunately, when it was fitted, and the cover was put on, I stuck a transfer punch down the hole that the other bit feeds into and gave it a tap, turned the engine thru 180 degs, and then another tap.
As you can see, the pop marks should have been in the middle of the nut. This shows that the welded on bit isn't in line with the centre of the crank.

So it looks like that was the cause of the original failure of the spade on the end of the gear shaft.
Now that I know what the problem is, the cure is a fairly simple, but a largish set up on the mill, and a little more turning.
A small 'just' job has turned into a major fix.
But that is what friends are for.
Bogs