Thread outside diameter doesn’t need to be nominal size you can go .305/.310″ with no problem, also is your die sharp and a decent quality make and are you using a tailstock die holder?

Tony

I recently had issues screwcutting a 2″ long 5/16″ x 32 ME thread on 303 stainless steel for slide valve rods. First rod screwcut fine but the second rod was a disaster. In the end I bought a new die and mounted it in a sliding die holder in the tailstock. The die cut a lovely thread under power at about 40rpm. Given the stainless steel I used Rocol for lubrication. Before threading I machined the stock to 5/16″ diameter with no concious effort to be undersize. The die came from Avon Engineering Supplies:

https://www.avontapdie.co.uk/me-5-16×32-split-die

The valve rods and slide valves assembled:

Julie

On 22 October 2025 at 07:01 JasonB Said:

This becomes more critical as the pitch becomes smaller and/or diameter increases as you will end up with a very weak thread as it will not be full depth.

As an example this is a section through a 1/2″ x 40 TPI ME thread.

Nominal dia on the left, reduced by 0.1 x p in the middle which is hardly any loss of depth. But on the right we have Dave’s  (SOD) reduction of 2.5% off nominal.

I’ll leave the  reader to work out the consequences of applying the suggested 2.5% of Nominal reduction to some of the constant pitch thread, anyone for M40x1.0 😯

The situation can be made even worse if you start using larger tapping size holes for the mating female thread as the combined lower engagement percentages can give very little engagement.

On 22 October 2025 at 07:01 JasonB Said:

Do be very careful if taking Dave’s advice to reduce Nominal by 2-2.5% as that does not take into account thread pitch so is a bit Duff. Although his suggested dia is correct it bears no relation to the percentages he mentions.

…

To be pedantic, the advice is Tubal Cain’s not mine!

I submit the approximation is a reasonable Rule of Thumb, and Jason is fussing over a minority case!    Though adjusting diameter by percentage becomes dubious for extremes like 40TPI, it works well enough on common thread sizes, metric and imperial.  Calculating from pitch is more work, especially with Imperial threads.

The following pitch example is ⁵⁄₁₆” X 32ME.

• Imperial threads are specified by TPI, making it necessary to first calculate the pitch.   Not difficult when you know how: 1/TPI (pitch in inches).  1/32tpi = 0.03125″
• Calculate diameter reduction from pitch × 0.1  Here 0.03125″ × 0.1 = 0.003125.   Practically, 3 thou
• Convert the imperial fraction size into the decimal equivalent:  5×16″ = 0.3125″
• Therefore, target diameter is 0.3125″ – 0.003″ = 0.3095″

The pitch arithmetic suggests reducing to 0.3095, whereas 2.5% rule of thumb is 0.30469″ or 0.30625″ at 2%.  If tweaked to 1% when cutting fine threads, % reduction suggests  0.309375″, which is the same as Dormer.

Practical note – when guesstimating the reduction, the start diameter should be as specified for the thread.  Imperial and Metric stock is often ‘nominal’, so measure it first!  That ⁵⁄₁₆” rod might be 8mm metric.  Or vice versa.

A practical factor favours percentage reduction slightly.  Cutting tends to raise burrs and squidge the thread so they’re a shade bigger than expected.  This is probably why John got a good result from 2.5%, even though as Jason says the advice is a bit duff for 32tpi.  Oversize shows up mostly in malleable alloys like mild-steel and the softer Brasses, and is most obvious when making tight fitting threads for maximum strength.   Whilst the maths is good, you can’t ignore the real world!

What the workshop is for makes a difference.  I’m into experimental work with occasional repair, not models.  I rarely make close tolerance threads because home-made cut threads are significantly weaker than rolled screws.  When strength matters, I buy in. My main application is bespoke fasteners, where the thread can be down to 60% engagement without causing problems.     Others take far more trouble, routinely adjusting split dies to get more accurate threads.  Possibly unnecessarily.

Always good to ask, “how good does it need to be?”   My training emphasised that mindlessly pursuing ‘quality’ is an egregious sin.  Professional engineering is about meeting requirements at minimum cost.  The specification and what’s ‘fit for purpose’ and ‘value for money’ are defined in advance. The requirement determines where a product sits between too cheap and high-end, not the builder.  Part of this avoiding ‘paralysis by analysis’, including not wasting time on unnecessarily laborious calculations,   An irony!  We Model Engineers are more likely to commit the opposite error, which is avoiding maths at all costs!  It’s true: I have to overcome my own laziness and stupidity.

However, Model Engineering is a hobby, and we might pursue making the best of all possible objects, time and cost irrelevant.  If the goal is producing a Gold Medal winning exhibition model, then the builder must work to a consistently high standard.  I think most of us have lower requirements, and need not fuss.  But the choice is personal! It’s fun get jobs ‘spot on’ and learning how to excel when needed.  Next time I cut a 40TPI thread, I shall remember Jason!  Next time I cut M10, I won’t!

Dave

Slight truncation of the top of a male thread doesn’t weaken it. Similar to tapping hole being bigger than root diameter for female thread. Making it generally undersize so it is a sloppy fit does weaken it.

I agree with Jason here. A model engineer should be able to do this based on pitch. If you’re threading rebar with standard fastener threads you’re forgiven for only considering the diameter, but I think most people here hold themselves to a slightly higher standard.

But if I’m allowed to pick a nit: Neither metric nor Unified have rounded crests. IIRC it’s from Withworth, though I’m sure other threads from the same era do the same.

That is true. A slight truncation has little effect on strength and none on fit.