Universal thread cutting

Ive seen earlier posts on members having brain crumble over metricating a myford - kit costs £190+ just for the combi slides, as for the gears and new yoke etc plus all that work to select gears etc - you know the drill

So I wondered given the marvels of CNC etc is there a means of reading the chuck speed/position and then translating it to a servo motor that drives the lead screw. So no messing with intermediate gears

go onto to a mars landing - if you have VFD then its a short skip for it to become a servo motor driven chuck - then you can do circular index drilling, and maybe make a hobbing attachment to shape gear blanks of your choice

Its getting a bit Elon Musk here?

Well its what we do in lockdown

btw this was inspired by my recent pass thru the day of my mispent youth on a Ward 7A capstan and training up to be an auto setter on those Herbert monsters - anyone remember

Yes. It has been done many times in the past.

John Stevenson and Brian Thompson's gear hobber was an early demonstration of the principle.

Look up 'electronic lead screw (ELS)'. This is an open source project.

On YouTube, the latest incarnation is by Clough42 (which has also been discussed on here in the past).

A VFD is not a servo.

For any indexing on a servo-driven (i.e. position-controlled) spindle, an integral brake is somewhat crucial.

Wow some serious brainz have gone into this topic

Not so Musky at all. However one wonders whether the old ML7 lady is right for this, given the potential expense and time

Quick google didnt show up any ready made cnc screw machines, maybe not looking in the right place

but my knowledge of cnc world is very limited

If you want a lathe to cut imperial/metric threads at the touch of a button look no further than the ELS by Clough42, many of us have gone down this route. Can't comment on your Mars landing reference, but CNC will do more or less anything you want if you have the knowledge & funds.

Tony

OMG there 37 vids, but many thanx for the tip, always good to lean new stuff

will post my stuff on a dirty cheat going metric on compound slides shortly, waiting for some parts

Go for it, the vast majority of the electronics will carry across to another lathe if you ever change. I built a Joe Noci version (described in MEW), wouldn't be without it.

Numbers like 21, 63, and 127, will depending on the error you can tolerate give you change wheel set ups to do the job with little cost or trouble. Not sure where a VFD would be much help ? You need DC stepper motors. Noel.

This is one of the reasons my ML-7R-cum-Super 7 will never have a gearbox. For metric threads and multi-start worms etc., for my two-pennyworth, the change-wheels provide for greater flexibility and outweigh the convenience of a gearbox any day of the week.

Interesting point, Chris.

Perhaps it's too easy to think the more complicated the solution the more efficient the result; or that the solution even needs to be complicated; and I have seen others and been caught out myself by this tendency.

I suppose the main advantage of a gearbox is the ability to switch from screw-cutting to plain turning and back rapidly - I have a gearbox for my ML7 but am yet to fit it - but the chart in the change-wheel cover does give the wheels for several standard metric threads just from the ordinary change-wheel set. My Harrison L5 lathe manual gives similar combinations, but I think I have a 127T wheel for it anyway, and it has a narrow-range 3-speed shaft gearbox that extends the change-wheel sums.

It's probably worth looking at the intended work. Brian may wish to make long, very highly accurate threads so does need very close matching.

For most of us though I suspect most of our screw-cutting is over short lengths, less than 10 turns, and to ordinary accuracy standards; not long threads and still less, high-precision lead-screws. For those two, rare, instances it is usually simpler and not very expensive to copy industry and base the components on stock studding / lead-screw rod and nuts.

Consequently I spent a few lunch-breaks at work calculating by spread-sheet change-wheel combinations for odd-numbered and metric pitches on a simpler lathe than a Myford, with a 1/8" lead-screw and smaller change-wheel range; and was surprised how close many came within sensible tolerances. Having done that, all I then need is to print a copy for ready use! No need to search for prime-numbered wheels (though I did buy a 63T wheel), electronics,stepper-motors and the like.

By all means build an NC lathe as its own project but I am wary of going Long Way where the Short Way reaches the same point just as properly, with less overall cost and effort.

I have been there - making the thing far too complicated for its own good, then spotting how I could have done it and wondering why I threw so much extra time, steel and electricity at it. It's a matter of balance, but efficiency and quality count, not complexity!

If only short lengths are screwcut why not use 33 and 34 tooth wheels on the output after the tumbler, I thought this was pretty standard practice?

Could you kindly expand a little on how prime-numbered gears (63t not being one of them) might be of any use to someone?

If we start at 20t and stop at 127t, the primes are:

23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97, 101, 103, 107, 109, 113, 127

127t is used not because it is prime, but because it is fundamental to the definition of the inch as 25.4mm.

Of all the others in the list above, which might one even consider searchng for?

I can think of only two, 23t, for NPT threads and 113t which might be useful for the 355/113 approximation of pi.

I think the point about prime numbered gears is that they are unusual but in some situations they can be used to give good approximations of the ratio of spindle to leadscrew to cut required threads. The 127 will give an exact conversion between mm and inch threads but would a 126.77 ratio for example be good enough over a short distance. I'm not suggesting 126.77 is an actual value anyone will come up with but spreadsheets do kick out lots of odd numbers when they are employed for these calculations of possible combinations.

Martin C

To go back to the original post...

• Yes, the ELS solution deals with the thread cutting problem and works with both metric and imperial whatever the lathe "nationality".
• I feel duty bound to mention that CNC using Mach 3 will do exactly the same things plus a lot more at probably lower cost and with less hassle. People say that M3 can't do threading - well it can on a Myford...
  
• However it won't control a VFD as a rotary servo and a VFD can't index rotation with an induction motor. You could replace the induction motor with a stepper but that's a lot of hassle. And anything using a belt or gears to drive the spindle will have poor rigidity.
• You can use a CNC lathe to shape gear cutting tools...
  basically by profile turning. This one is the circular approximation to involute, but it could easily be true involute or cycloidal. But I can't think of a way that a CNC lathe could do better at hobbing a gear, though a CNC mill with rotary axis can (Neil Wyatt amongst others have discussed that here I think).

The value of 127 is indeed tied to the definition of metric and imperial conversion, but the fact that it is prime is no coincidence. It's the result of decomposing the conversion value into its prime factors.

There's no need to hob gears on a CNC mill. Using a single cutter is simpler (no need to make a hob) and possibly more accurate, albeit slower:

And the finished pinion, made to test the fit with an internal gear, which was the purpose of the exercise:

Andrew

Indeed. I was just making the point that if you do want to hob a gear (e.g. to get a better tooth form) then it's possible. I didn't add but should have done that you can use a single-point tool to do so.

I don't agree with that. One disadvantage of using a single cutter is that one needs a 3D CAD model. But that model can be as accurate as needed. A hob forms the tooth shape via a series of facets. A single ballnose cutter also creates a series of facets, tending towards scallops as the cutter tends to perpendicularity. It is easier to control the facets/scallops in CAM than with a hob. A disadvantage of hobbing is a tendency to undercut gears with fewer teeth, the exact number of teeth being dependent upon pressure angle. With a single cutter the CAD model can easily be adapted to eliminate undercutting.

Getting back to the OP, while electronic leadscrews are versatile they're not on my list of things I want or need. My lathe will cut a wide range of TPI and (approximately) metric threads and DP/module worms. So far I've only needed one thread I couldn't screwcut - a 1 tpi lead 2 start worm. The coarsest thread my lathe will cut is 2 tpi. Although I subsequently dumped the two start arrangement by then I'd already set up the CNC mill for cutting the thread, so I stuck with that arrangement:

Andrew

I should correct myself guys, when I said VFD wrt ELS it seemed to cause a veer off topic, I meant that when you went to modern electronics control you could jump through to stepper motors with inherent servo motor control, not using an induction motor for the purpose - perish the thought. I must get into that vid series by Clough42 - will it bring me up to speed

Can I ask what the ELS users do about returning the carriage for th enext cut when doing say a metric thread on an imperial leadscrew equiped machine? Do you still have to leave the half nuts engaged and then use the driven screw for a fast rewind after retracting the tool.

JasonB, Yes you have to keep the half nuts engaged at all times if cutting imperial threads on a metric lathe or vice versa.

Tony

If you set a lathe up for CNC with a ball screw you will probably have a metric pitch one nowadays and it is always engaged whatever the lathe is doing. You tell it where zero is on the Z axis either by manually zeroing at some useful point such as a workpiece face or by a home switch if you are doing repeated parts over a long time. The CNC controller then keeps track of where it is (or should be if there are no missed steps) and returns to the start point for the next pass once the tool is retracted.

Martin C