What's new
What's new

A different way of doing metric threads on an inch lathe?

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
While dreaming about electronic leadscrew nonsense this popped into my head the other night. I had to wake up and write it down before I forgot it. I haven't built it yet, but it seems like it could work.

Curious to hear other's opinions.

So, instead of doing the 127/100 compound gearing or an ELS conversion, or some other trickery to convert the leadscrew ratio to what is needed for a metric pitch, what if we changed the ratio AFTER the half nut? What if we never touched the gearing at all.

This is my idea-

Assemble a toolholder for threading that has ways inline with the Z axis, allowing a half inch or so of side-side movement.

Pivot point is fixed to rear bedway, like a taper attachment bed clamp. Just a clamp with a stud sticking up.

Here's the magic- You take your metric pitch you want to cut and do some math. Say I want to cut 2mm pitch. 2mm pitch is 12.7 threads per inch. So I set the quick change for the nearest pitch- 13 TPI. Then I whip up a foot long or so lever that compensates for the pitch error. Lets say we attach the lever with a linkage rod (pair of heims) to the carriage 12" from the center of the pivot. 13/12.7= 1.023622 x 12" = 12.2834" is where we would attach the linkage rod that moves our second Z axis @ the toolholder.

You set the pivot point up so the arm is around 90 degrees and you're only threading an inch or two max 99% of the time so the ratio stays pretty close through it's arc.

For metric threads I think you could even use the half nut. The lathe thinks you're threading an inch pitch. The leadscrew's relationship to the carriage position is in perfect time with the thread dial. The ratio change happens AFTER the carriage movement.

To cut a different pitch you machine up a different lever arm and stamp into it what TPI to set the quick change for.

So to cut metric threads you would drop your aloris toolholder modified with a linear bearing truck on the toolpost and indicate it straight. You'd put the bed clamp on the ways. You'd install the appropriate lever from an assortment of levers you machined up (one lever could have multiple pivot points machined into it) and attach a couple link rods. Probably pop on and off in a few minutes if designed well enough.

Thoughts?
 

gbent

Diamond
Joined
Mar 14, 2005
Location
Kansas
Sometimes you will have to retard the tool bit, other times advance it. 2mm pitch on 13 tpi setting will require advancing the tool bit. Its darn sure thinking outside the box.
 

CalG

Diamond
Joined
Dec 30, 2008
Location
Vt USA
pantagraph!






eta

If anyone has an odd thread chasing mechanism for a Wade94 sized machine,.... give me a shout ;-)
 

winger

Stainless
Joined
Dec 28, 2005
Location
portage county, wisconsin
Would you still have to keep the half nuts engaged the whole time? I think so.

Just got done doing a m45 x 2.0 internal about 4" long. What a pain that was. The lathe has metric change lever at least so swapping gears.

Would have been a piece of cake on a cnc.

Dave
 

TGTool

Titanium
Joined
Sep 22, 2006
Location
Stillwater, Oklahoma
I've seen a setup something like that, I think on a New England Model Engineers display. It used an additional slide to provide the advance or retard. There was a bar that could be angled like a taper attachment somewhere between the ways with a slider and bell crank to provide the shift from the leadscrew pitch.
 

john.k

Diamond
Joined
Dec 21, 2012
Location
Brisbane Qld Australia
The problem was sorted 200 years ago with a slave thread of the same pitch ,and any diameter as may be suitable rotating with the spindle ,which drives the threading tool.......or the thread unrolled into a suitable cam form to operate the tool.
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
Would you still have to keep the half nuts engaged the whole time? I think so.

Just got done doing a m45 x 2.0 internal about 4" long. What a pain that was. The lathe has metric change lever at least so swapping gears.

Would have been a piece of cake on a cnc.

Dave

No, you wouldn't. You could disengage the half nut, move the carriage back and forth and when you re-engage the half nut it will still be in perfect time. The lathe thinks it's cutting a pitch that is a division of the leadscrew. That timing won't change.
 

SBAER

Hot Rolled
Joined
Aug 21, 2006
Location
Kitchener, on canada
That is definitely some original thinking but you seem to be able to buy working CNC machinery for cents on the dollar, wouldn't you be better off buying a working flat bed lathe?

The 1990 Romi Bridgeport lathe I have is handier than any manual lathe I have ever used including the Hardinge that sits right beside it, not just for threading but also for roughing, chamfering and any turning any type of taper.

I've decided that I am going to spend my effort on learning how to do CNC retrofits so that I can make machines with out of date controls productive for my particular style of work.
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
I've seen a setup something like that, I think on a New England Model Engineers display. It used an additional slide to provide the advance or retard. There was a bar that could be angled like a taper attachment somewhere between the ways with a slider and bell crank to provide the shift from the leadscrew pitch.

I could see that working well also.

I was thinking the lever would be a reasonable approach because it's easy to put on and off the machine and one lever could have multiple positions to cover a range of pitches. The more I think about it, you could cover all common metric pitches in one lever arm if you have multiple fulcrum points.

The general idea stuck in my head because of how it would allow the half nut to function same as inch threading. That would sure be nice to have all threading jobs work the same way.
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
That is definitely some original thinking but you seem to be able to buy working CNC machinery for cents on the dollar, wouldn't you be better off buying a working flat bed lathe?

The 1990 Romi Bridgeport lathe I have is handier than any manual lathe I have ever used including the Hardinge that sits right beside it, not just for threading but also for roughing, chamfering and any turning any type of taper.

I've decided that I am going to spend my effort on learning how to do CNC retrofits so that I can make machines with out of date controls productive for my particular style of work.

I have a shop full of nice CNC stuff. I have a nice CNC lathe with a 5.5" spindle bore and live tooling.

Do you know how much quick change soft jaws cost for a 16" Schunk power chuck? So when I need to thread one hydraulic cylinder part it takes hours to program and setup for it. Plus I have to consider the time it takes to set the machine up for it's regular work again.

My manual machines are great at what they're made for. Just would be nice if they did metric threads.
 
Last edited:

CalG

Diamond
Joined
Dec 30, 2008
Location
Vt USA
The problem was sorted 200 years ago with a slave thread of the same pitch ,and any diameter as may be suitable rotating with the spindle ,which drives the threading tool.......or the thread unrolled into a suitable cam form to operate the tool.


I like the way you think!

(see the eta on post #3) {-)
 

EPAIII

Diamond
Joined
Nov 23, 2003
Location
Beaumont, TX, USA
I am having a bit of difficulty imagining this in my head but it sounds like basically you are having the tool and it's holder move at a different speed than the lathe carriage. And this is accomplished with a carefully calculated lever arm that pivots on a fixed pivot on the rear way. That lever is driven at one distance from the pivot by the carriage's motion (thanks to the lead screw and gearing) but the tool and holder are driven from another point on the lever which corresponds to the metric motion needed.

First, this could work. And it sounds a bit like some mechanisms that were developed to generate lead screws of increasing accuracy by using a correcting mechanism which was set after the error of the earlier lead screw was measured. This measure and correct action allowed the new lead screw that was being cut to be more accurate than the one that was doing the cutting.

You may want to look up some of these correcting mechanisms. They were not limited to short threads but applied the same correction for the full length of the lead screws. They can be found in texts that discuss the history of precision in cutting threads.

BUT, because of the mathematics of the action of a thread dial, I think you will find that you will not be able to synchronize the lead screw with any of the markings on a standard English thread dial with a metric thread without counting the number of revolutions of the spindle or lead screw before engaging the half nuts a second time. You will still need to rely on keeping the half nuts engaged and backing the spindle up in order to synchronize the metric thread being cut.

Going a bit further on this subject of synchronizing the thread after opening the half nuts and re-closing them, contrary to popular belief, a thread dial can be built to function in this manner when cutting metric threads with an English (inch) lead screw. What is not possible is doing this with a standard English thread threading dial. The standard threading dial for English threads is based on four inches of travel on the thread being cut. This four inches allows only threads with certain fractional threads per inch to be synchronized and those fractions ONLY include numbers that are whole, those with half fractions, and those with quarter fractions. Any whole number of threads per inch will return to synchronization after any whole number of inches of travel by the thread being cut. So 1, 2, 3, 4, 5, etc TPI threads can be synchronized. Also any TPI expressed as a whole number plus a 1/2 fraction will return to synchronization after 2 inches of travel. So any TPI with a 1/2 fraction can be synchronized by using any even or any odd pair of numbers on the threading dial. Likewise, any TPI with a 1/4 or 3/4 fraction will return to synchronization after FOUR inches of travel of the thread being cut.

BUT, THAT'S IT! If your English thread has a 1/3 or 2/3 fraction, then it needs THREE inches of travel to achieve synchronization. I mentioned counting revolutions above and that could be done for a 1/3 fraction. You would need to advance one number on the threading dial for each three inches of travel of the thread being cut and that would be for every three revolutions of the spindle between the time you opened the half nuts and the time you wanted to close them again. Keeping track of this would be almost impossible while doing all the other things needed for cutting the thread. But, IN THEORY, it could be done.

Likewise, the difference between an English thread and a metric thread can be measured in a rational ratio. That means with a fraction that is composed of two whole numbers. When that fraction was reduced to it's lowest terms, one of those numbers would be 127 and the other one would depend on the exact thread being cut. Again, it would be possible, IN THEORY, to count the spindle revolutions and close the half nuts at a point where the thread would synchronize. But, just like in my 1/3 example, you would need to advance the number on the threading dial according to the number of revolutions of the spindle and it would not be a simple relationship as in the 1/3 example. Worst case, you would need to wait for 127 revolutions of the spindle before closing the half nuts on the same number again. Some metric thread pitches would allow this to be done with fewer revolutions, but all in all, you are still going to be waiting for a while. And if you miss one such point, then your wait starts all over again.

My point is that it is not what the lathe and/or it's gearing think you/it are doing. It is the basic mathematical relationship between the English and metric threads that determines when that English threading dial can actually tell you it is time to engage the half nuts for the metric thread you are cutting. It is possible. But it is NOT easy.

Two other points about threading and threading dials.

Fact one: the four inch nature of the standard threading dial and the fact that all English threads are either whole numbers or 1/2, 1/4, or 3/4 fractions go hand in hand.

Fact two: the reciprocal relationship of English threads, which are measured in TPI or threads PER inch and metric threads which are measured in linear distance or mm is why the threading dials for English threading is so simple while the threading dials for metric threads need to have different gear ratios to cover the range of each type of thread. English threads, measured in TPI are a lot simpler to synchronize than metric threads measured in mm.
 

Billtodd

Titanium
I had a similar idea a few decades ago before I got a hlv h and never worried about threading again 8-)

The idea was to replace gears with levers and slides such that by changing the angle of the slides an infinitely variable different ratio could be defined .

I should not have been surprised when a Google patent search revealed several similar (and at least one better) ideas.

Somewhere on Tony's lathe site are screw-thread correction lathes that use a rotating 'half-nut, to correct their lead screws in a similar manner.


Bill

US2492251A - Lead screw variator
- Google Patents

US4092884A - Lathe adapter for non-standard thread machining
- Google Patents


and many others
 

David_M

Hot Rolled
Joined
Sep 30, 2014
Location
Midway, GA, USA
No, you wouldn't. You could disengage the half nut, move the carriage back and forth and when you re-engage the half nut it will still be in perfect time. The lathe thinks it's cutting a pitch that is a division of the leadscrew. That timing won't change.

Take a look at this video about half-nut engagement timing:

Threading Dial Indicator for Metric AND Imperial Threads - YouTube

He adds this to correct a mistake:

In the example at 9:15 I messed up the numbers, sorry guys. The fraction 4.8 is for a thread pitch of 1.25mm, not 1.75mm.

It's more complicated than it appears at first glance.
 

gbent

Diamond
Joined
Mar 14, 2005
Location
Kansas
Take the lead screw out of the compound and set it parallel to the spindle and you have your sliding tool holder. Now you just need to build the lever.
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
Take a look at this video about half-nut engagement timing:

Threading Dial Indicator for Metric AND Imperial Threads - YouTube

He adds this to correct a mistake:

In the example at 9:15 I messed up the numbers, sorry guys. The fraction 4.8 is for a thread pitch of 1.25mm, not 1.75mm.

It's more complicated than it appears at first glance.

I think you're implying the pantograph idea won't work with the half nut? It's a good video, but I didn't catch anything relevant to this discussion in there.

Think about it like this-

With this, lets call it "Pitch offset device" you have an offset distance that is linear and locked to the carriage motion. Has zilch to do with any rotating gearing in the machine.

If you disengaged the half nut and ran back to the start you haven't lost your place. The mechanical linkage is what is keeping your timing. Anywhere you lock that half nut onto the leadscrew it is going to be 100 percent in time with the spindle to cut the inch pitch that it's set for.

In this way, YES, you can use the half nut for metric threading because instead of the metric pitch you are cutting being some prime number ratio division of the leadscrew it is a perfectly even leadscrew division PLUS an offset percentage. The offset percent is ALWAYS connected. The connection is never broken. The linkage maintains the timing that leaving the half nut engaged provides for cutting a metric pitch with gearing.

Keep in mind here we are talking pretty small numbers. The difference between 12.7 and 13 TPI is .00181" per thread. That's about 47 thousandths over 2" of thread length.

So you would in effect be cutting a 13 TPI thread with a linear 2.3 offset percentage.
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
Take the lead screw out of the compound and set it parallel to the spindle and you have your sliding tool holder. Now you just need to build the lever.

I like this idea. It would be nice if the compound could be run off the end of the screw, but I don't think either of my lathes will let that happen.
 

CalG

Diamond
Joined
Dec 30, 2008
Location
Vt USA
One should be overjoyed to understand that "gears" are nothing more than lever arms. The "leverage" between readily understood though tooth count. (Just don't mix tooth pitch!)
 

Garwood

Diamond
Joined
Oct 10, 2009
Location
Oregon
One should be overjoyed to understand that "gears" are nothing more than lever arms. The "leverage" between readily understood though tooth count. (Just don't mix tooth pitch!)

In a lot of ways, sure.

Levers differ from gears with how a lever doesn't make a full rotation and the ratio of a lever can be easily adjusted- No fixed center distance to deal with.

You could do the same thing as the lever arm with compound gears moving a rack attached to the toolholder, but then you'd have to change gears for every pitch.

I think the lever on the bed is a simple solution to the problem and it could be superior to changing gearing as the half nut function is retained.

To prototype the concept I think I will replace the toolpost with a short piece of linear guide rail fit to the compound T-slot and a threading tool mounted to the guide truck.
 

steve45

Stainless
Joined
Jan 16, 2012
Location
Midland, Texas
I've seen a setup something like that, I think on a New England Model Engineers display. It used an additional slide to provide the advance or retard. There was a bar that could be angled like a taper attachment somewhere between the ways with a slider and bell crank to provide the shift from the leadscrew pitch.
The Home Shop Machinist magazine (gasp!) had an article about a device like this a couple of years ago. By using an angled bar, like on a taper attachment, you end up multiplying the leadscrew speed by the cosine of the bar angle, or something like that.
 








 
Top