An electronic threading dial?

[Arquibaldo]

Just a crazy idea, if I may:

I bought a very nice old lathe, but it came without a threading dial - I think it never had one...

This got me thinking: could an electronic threading dial be built? Has anyone ever heard of it, or
could give me some pointers?

The lathe is manual, and will continue so. It is easy enough for me to add encoders to the spindle
and the lead-screw, so that a microcontroller can readily compute the gear ratio and the relative
position of the two. Now where do I go from here? What would I have to compute and display to tell
me when to engage the half-nuts?

May be, if this is doable, it could be even better than a mechanical threading dial, allowing me
to disengage and reengage correctly the half-nuts in situations where the mechanical dial is
useless: metric threads and uncommon pitchs...

Nelson

[serriadh]

It sounds like it would be doing a fair chunk of the work an electronically controlled leadscrew would be doing anyway. There's no reason that a stepper-driven, spindle-synchronised threading system couldn't use the existing halfnut and leadscrew, and would make a nice flexible power feed into the bargain. Oh, and there's be no more gear changing, or noisy gears at all for that matter.

It is easy enough for me to add encoders to the spindle and the lead-screw, so that a microcontroller can readily compute the gear ratio and the relative position of the two.

Careful, there. You'll need the spindle to rotate enough for the encoder on your leadscrew to advance to its next gap and send a signal to your controller. Depending on the accuracy and precision of your leadscrew encoder, you'll probably want more than one pulse (ideally you'll want more than one complete rotation) but either way there's a finite length of time required to compute the thread pitch before you can start displaying useful thread dial information.

May be, if this is doable, it could be even better than a mechanical threading dial, allowing me to disengage and reengage correctly the half-nuts in situations where the mechanical dial is useless: metric threads and uncommon pitchs...

Take a look at the explanation of how thread dials work here... scroll down do the green text by Gary Giddings.

If you're feeling keen, you can read my (quite possibly incorrect!) summary of how a thread dial should work.

The thread dial indicator basically tells you when the thread you are cutting and the thread of your leadscrew match up: ie, the cutting tool is perfectly aligned with the thread you are cutting, and the halfnut is perfectly aligned with the leadscrew thread. If you engage the halfnut at this point, you'll carry on cutting the thread where you left off.

You can't engage the halfnut at just any point though, because the threads in the halfnut and the threads in the leadscrew only engage one per revolution of the leadscrew (assuming a single start leadscrew). Your ETD would have an index pulse at this position, I imagine. Mechanical thread dials have a marker where the two are aligned.

I'm assuming that you're using an imperial (or English, if you like) lathe here, so it doesn't have a metric leadscrew.

For threads with a whole number of TPI:

When cutting thread pitches that are a multiple of your leadscrew's pitch you can engage the halfnut once per leadscrew revolution, on that index pulse.

When the thread TPI and leadscrew TPI have a Lowest Common Denominator greater than 1, you should engage the half nut every LCD revolutions. For a 12TPI thread on an 8TPI leadscrew, the LCD is 2, so you can only engage the halfnut every other revolution. This is because two eigths is a whole number of twelfths (3).

If the LCD is 1, you can only engage the halfnut every (leadscrew pitch) revolutions... ie, for an 9TPI thread on an 8TPI leadscrew you can only engage once every 8 revolutions. This is because eight eighths is a whole number of ninths (9).

For cutting other threads:

For thread pitches which which are not whole numbers of TPI (say M6, which is 25.4TPI), you need to be a little more clever. You need to find the distance into which each thread pitch will go a whole number of times. In this case, 5" which is 127 threads at 25.4TPI and 40 threads at 8TPI. So, every 40 revolutions of the leadscrew, the thread, tool, halfnut and leadscrew will be perfectly aligned, and you may engage!

This is why when you're cutting oddball pitches the halfnut is generally left engaged, and the lathe is reversed instead, keeping the tool and leadscrew and halfnut in perfect sync instead of waiting a minute or two between each pass whilst counting large numbers of revolutions!

Note: this is the third revision of the 'other threads' bit, so there's still no guarantee I got it right! It now makes more sense to me, but I'm working this stuff out from scratch rather than repeating information by someone else who might actually know what they're talking about.

[Arquibaldo]

Dear serriadh:

Thanks a lot for your great contribution, particularly your summary and the explanation by Gary Giddings.
I understand now much better how a threading dial works, and can go ahead with my project. As a minimum,
I can now replicate in electronics the function of a standard, mechanical threading dial, which only has
to "know" the position of the lead-screw... By the way, I found out that the reason this lathe has no
threading dial might be that it uses a very odd leadscrew: 5 TPI instead of the usual 8 or 4 TPI...

I also found out that this idea has already been proposed more than 7 years ago, at the
homeshopmachinist bbs, thread #4146 - there was even a guy saying that he was developing it to sell
as a commercial product...

About electronically controlled leadscrews: yes, that would be the ultimate goal. I have been interested
in then for many years, starting with a very strange contraption from the days before affordable steppers
and microcontrollers, until more recently, the work of John Dammeyer at the E-LeadScrew yahoo group.
However, this is a much more serious project, possibly involving extensive modifications to the machine
to couple the stepper (which has to be quite powerful) to the lead-screw and disabling the quick-change
gearbox... sometime in the future I'll get bold enough to try this, but on and older beat-up lathe I also
have - the new acquisition is too well-made to encourage any radical modification...

Gear changing and noisy gears: yes, they are a pain but, call me nuts, I admire the mechanical ingenuity
of the gear train in a traditional, well made lathe: the set of gears (mine has 15, from 20 to 127 teeth),
the banjo, the QCG, allowing thousands of different combinations - really a mechanical masterpiece...
(you can see I am old...)

I'll report here on any progress I might achieve.

Nelson

[serriadh]

I'm currently using an old lathe that was slightly mistreated by its previous owner, and as a result I can't do any work with the leadscrew at present. I removed the gear train and banjo so as to reduce running noise and decrease the number of bits capable of eating me.

The final gear in the train, the one that was fixed to the leadscrew, was held on with a pin and a collar, and once they were out of the way it just left a length of plain shaft that would be fine to mount a motor coupling to. If your lathe has been mounted to a bench or cabinet or some other solid surface, a motor mounting bracket can be assembled that also mounts to the surface without having to drill new holes into the lathe framework. Indeed, were it not for the awful condition of the rest of the lathe, that's exactly what I would have done!

The process would be largely uninvasive and easily reversible... in fact it could potentially be less destructive than an ETD, which would need two encoders fastened to the lathe (instead of just one for the ELS, on the spindle) and a switch of some kind fastened to the half nut engaging system, so it can be turned off and on when the nut is engaged and disengaged respectively.