Bob.Murphy wrote:taps... are a bit of a blunt instrument.
I mentioned different classes of fit with taps because I didn't know your background. This is a case of the "right" way of doing something being more complex than the much easier "possibly-OK" way and I didn't want to be a person who insists it must
be done the right way in case you weren't in a position to do it that way.
I have indexable threading bars (and even Whitworth-form inserts) with shanks down to 1/8"-ID, but even with the right bar and insert the problem in making this bespoke nut is getting the size right. By 'right', I mean having the desired tight, but not too tight, fit on a threaded rod that is buried in the lathe. If the rod were separate you could sneak up on it by testing the fit after every pass that removed another thou. But, you can't do that without disassembling your lathe, and if you disassembled your lathe you couldn't make the nut.
In any case, the only way you'll know if you succeeded is after you've made the nut and disassembled the lathe to test its fit on the shaft. Given this, something to think about is buying a tap that would give the tightest fit (i.e. having the class of fit giving the smallest ID), making the nut, disassembling the lathe, and testing it. If it gives an acceptable fit, great. However, if the nut you just made is too tight it now is a no-go gauge, and if too loose you can use it to judge how much smaller it needed to be. Then reassemble the lathe and make the next nut with an indexable threading bar using the information obtained from the first nut.
Nevhunter's split-nut suggestion would be a good one to pursue, but only if the cavities surrounding the two nuts are large enough to fit the required clamping mechanism.