Earlier in this thread I mentioned the usefulness of a 38t gear in converting imperial lathes to metric screwcutting use. Since then we have seen a superb treatise on cutting BA threads on the small Cowells lathe. And a herculean task it has been.
I didn't make my case very clear (as is often the case with me) so with your forbearance I will try to clean it up.
_____________________________________________________________________________
My Exe lathe has a permanently fitted 18t gear on the spindle and drives an 12tpi leadscrew through user selected 18 DP gears. Obviously an 18t gear on the leadscrew will give a copy of the leadscrew pitch at 12 tpi.
The formula for any pitch is simple - chosen pitch x 3, divide by 2 gives the teeth on the driven gear (for simple trains of 2 idlers)
e.g. 12 x 3/2 = 18t
26 x 3/2 = 39t
40 x 3/2 = 60t - or 30t on leadscrew with 36/18 double and one idler
To cut 25.4 tpi (1mm pitch) needs a leadscrew gear with 38.1 teeth. Search as ye may, nobody makes one. So try 38 and it comes out as 0.9973753280839895013123 - ok, that's the last big number
but just to demonstrate, if my lathe was capable of cutting a thread length of 1 metre it would have 997.37 threads on it. More likely would be 99.7 over a 100mm (4") length. Like Dave said, you'll never see it.
Other metric pitches are handled by compounding the gear train and most standard M series are possible.
_____________________________________________________________________
Now my imaginary metric lathe needing to cut imperial threads. It has a 1mm leadscrew and I can choose which gears I use on the spindle. So I put a 38t on there. I chose 40 tpi as my target thread and found a 60t gear on the leadscrew would do just fine giving 40.105 tpi - the error is the same as the previous example but in the other direction.
Error = either 381/380 or 380/381 depending on application
So I'm rather tickled to find that if I had such a metric lathe in its 38t driver form the calculation for driven gear is the same as for my Exe lathe: desired pitch x 3 / 2 and a good range of regular imperial pitches can be obtained.
No spreadsheets were harmed in the making of this post, nor need be in applying it.
Ray