PDA

View Full Version : Thread cutting, how many turns? Maths query...



The Artful Bodger
08-22-2009, 05:34 AM
If I understand correctly the accepted practice when cutting inch threads on a metric lathe is to use a 120/127 compound change gear to get the necessary ratio between spindle and feed screw revolutions. One must also stop the spindle at the end of each cut, retract the tool and reverse the spindle to 'motor' the carriage back to the start position keeping the half nuts engaged all the time.

Now it strikes me that with some sort of spindle revolution counter it would not be necessary to stop the spindle requiring only that the half nuts be disengaged at the end of the cut, the tool retracted and the carriage manually wound back to the start point all the while the spindle continuing to rotate. The next cut would be started by watching the counter and waiting until the spindle and feed screw approached the critical relative positions then reengaging the half nuts.

The question is, how long would one have to wait for the spindle and feed screw to get into position?

Black_Moons
08-22-2009, 07:31 AM
Depends on the pitch your cutting :)
While your idea could be pratical, at the very least you need a microcontroller and rotation encoder on the spindle/feed and program in the pitch you wish to cut and have it find common devisors.. could be thousands of revolutions before they line up again.

A slightly easyer method is to open the half nut, retract the tool and stop the lathe and take careful note of how many revolutions it makes before the lathe stops, then, turn the lathe on reverse, and reengage the half nut when the turn count goes back to 0, and let it motor the carrage back to the start of the thread that way. If your quick and observant, your threading dial alone should suffice.
Alternately, flip your threading tool upside down and cut the other direction if you have a pesky shoulder that your worryed about cutting up to.

Jim Shaper
08-22-2009, 03:03 PM
It's a whole bunch easier to just get a lathe that cuts both standards. :D

Forrest Addy
08-22-2009, 03:40 PM
Reverse back it's quick and foor proof. Otherwise you or a gadget will have to count 127 turns (this is the irreducible repeat independent of thread pitch) then close the half nut. Man's flighty mind was not made to count without a mistake to 127 for the number of passes it takes to cut a thread.

Reverse back and be glad of its simplicity.

The Artful Bodger
08-22-2009, 05:11 PM
Thanks for all the comments.

Forest Addy, I assume that would be 127 turns of the compound change gear? I am reassured that you seem to agree that the challenge would be independent of thread pitch.

The thought of making a counter does not faze me, electronics make such a thing very easy, easier still if I just turn on my laptop.

GadgetBuilder
08-22-2009, 05:49 PM
Martin Cleeve covers this in some detail in "Screwcutting in the Lathe". The gear ratio used when cutting a thread in the other language (English/metric) is an approximation so the leadscrew gets re-synchronized with the thread being cut periodically, where this period is much less frequent than when cutting a thread in the same language. (This also depends on how synchronized you want to be since there are close approximations to synchronism that occur more frequently.) The best synchronous distance can be large, one of his examples has an SD of over 9 feet. This implies that for some threads sync will take a while, perhaps more than is practical.

Cleeve suggests using a dog clutch instead since this works for both languages and picks up in one spindle revolution plus it allows threading at higher RPM than is common using manual control. I added a dog clutch to my 7x12 using Cleeve's concept and it does indeed synchronize both languages very nicely, see:
http://www.gadgetbuilder.com/Dog_Clutch.html

John

Carld
08-22-2009, 07:12 PM
What I want to know is,

Why make it harder to cut threads than is reasonable.

What is so hard about backing out the tool, reversing the chuck and setting up to cut the next pass?

What possible gain can be had by letting the lathe tick of revolutions while you pat your foot waiting to make the second pass?

Ahh, I see, it's to make it harder to cut the thread, that's the answer.

The Artful Bodger
08-22-2009, 07:40 PM
John, I was afraid synchronization may take an impractical time. The dog clutch is interesting and will certainly read up on that.

Carld, you ask why make something difficult? I take your point and from this day on I will buy everything I need and send out for pizza at every meal.

tattoomike68
08-22-2009, 07:51 PM
John, I was afraid synchronization may take an impractical time. The dog clutch is interesting and will certainly read up on that.

Carld, you ask why make something difficult? I take your point and from this day on I will buy everything I need and send out for pizza at every meal.

With practice and a lathe that dont blow out being slammed in reverse you can cut a metric thread faster than a standard thread.

I got stuck on a job doing lots of metric threads and got super fast at it, and I thread fast as is it is.

the three fastest ways to thread is threadroll , cnc, diehead. in that order.

gnm109
08-22-2009, 08:13 PM
On my Enco, the relay controlling forward movement won't drop out until the lathe chuck is completely stopped so there's no slamming it to get a quick reverse.

It would be nearly impossible to count turns and re-engage at the correct spot when cutting metric threads. I guess you could try. The only downside would be some scrap now and then.

On my lathe, I've changed the gears a few times to do a metric thread and it worked nicely leaving the half-nuts engaged and backing up without using the threading dial.

That's good enough for me, given that I seldom need an original metric thread. I do use metric now and then, but it's set screws and bolts that are available locally.

If I need a special metric thread on a rod, I'll just leave the half-nuts engaged and let it go at that.


.

Black_Moons
08-22-2009, 09:01 PM
I don't think the sync distance is allways 127 turns, after all theres the rest of the gearbox to play with the ratios, Not all ratios repeat after 1", Infact TPI's like 11 1/2 pipe thread only repeat every 2 inchs. Im not sure exactly how to do the metric math but I don't think it comes out exact either.

Im impressed by the dog clutch idea, as long as your output gear = spindle, your set, since a thread will allways be the same spot for every spindle revolution, reguardless of TPI.

Forrest Addy
08-22-2009, 10:16 PM
The 127 turns is referenced to the spindle. Monkey around counting the turns of an intermediate gear and you can get into trouble. The lead screw indexes from the spindle revolutions, therefore you have to calculate based on 127 +/- 0.0 spindle revolutuions to get a precise leadscrew close and thread pick-up when cutting metric thread pitches on an English lead screw.

I don't know why you guys are turning reversing back to preserve leadscrew timing into something to obsessively avoid. I've cut oddball leads with 5 start multiple threads that took over 400 passes to make a finished threads by reversing back. The carriage went back and forth like a planer table. Very long threads may take a while to reverse back. Most people thread far too slow and in most materials they could speed it up considerably.

Since my lathe has a VFD, I can thread at a lower speed like 200 RPM and when reversing back crank it up to 600 or more.

The secret is to have a single Fwd/Rev control. I have that right on the apron of my lathe and most lathes are made similar either via a reversing clitch or by electrically reversing the motor.

People having single phase motors may be in the soup because most single phase motors won't instant reverse.

Black_Moons
08-22-2009, 11:08 PM
My apron has the forward/reverse control.. But it has the WONDERFUL danger in that if you slam it into reverse when its going forward.. it just.. keeps going forward -_-; Can hardly decide if I want a VFD for my mill or lathe first.

(of course you have to move the control in a ~ sorta shape to go from forward to reverse. (ie, down to 'off' and then go far left before it will go further down into reverse))

The Artful Bodger
08-23-2009, 03:55 AM
Forest Addy, I like to think of the best way to make use of what I have which does not include 3 phase power, a VFD or even anything more than a barely adequate single phase power supply.

Stopping and restarting a single phase motor repeatedly puts a strain on the start winding and with a Chinese motor I tend to be aware of such things.

I have concluded from the various comments here that synchronisation via turns counting may not be practical, oh well, it was at least interesting for a while!:)

Paul Alciatore
08-23-2009, 04:05 AM
Many lathes are equiped with other, approximate transposing gears instead of an exact combination that uses a 127 tooth gear. These lathes will NOT synchronize in 127 turns and you will need to find the exact number of turns that will work with them. It will probably be less than 127, but this is not necessairly so.

Forrest Addy
08-23-2009, 04:43 AM
Bodger. As usual I can be witlessly arrogant and dismissive. I was not polite to scoff even indirectly at the equipment of those with priorities like feeding a family and paying down a mortgage.

The Artful Bodger
08-23-2009, 05:00 AM
Forrest Addy, no problems!:)

J Tiers
08-23-2009, 11:35 AM
The 127 is exact, not approximate.

A metric threading lathe can have a thread dial.....

It will have 4 gears, which you must choose between for any thread. it will have several sets of dial marks, which you consult a large table to discover which to use. probably easier to wind it back regardless, even with a metric machine.

All this because metric uses an insane system of "pitch based" instead of ratio-based" threads.

The US system of ratio (so many per inch) threads directly gives a single gear and one single set of marks, with a simple rule to choose marks.

If metric had agreed on a system of threads per cm, metric also would have a simple thread dial system.

it is the one huge failing of the metric system, which is otherwise very much like decimal inches.

Carld
08-23-2009, 02:52 PM
Carld, you ask why make something difficult? I take your point and from this day on I will buy everything I need and send out for pizza at every meal.

I have to take issue with that comment. Trying to find a short cut for some work has nothing to do with buying a pizza or anything else.

With the unknowns that can happen with opening the halfnuts on an imperial lead screw while cutting a metric thread you could spend a lot of time watching the lathe spin and still end up with a damaged thread when you could be cutting the thread. It's really a matter of being efficient and practical when doing a job.

The point is, you may never find the EXACT number of revolutions to get back to the correct engagement point and if you do, can you count that off each time without a mistake? I doubt it.

As far as finding the best and easiest way, I do that all the time and my wife makes the pizzas for us, we don't buy them.

The Artful Bodger
08-23-2009, 05:27 PM
I have to take issue with that comment. Trying to find a short cut for some work has nothing to do with buying a pizza or anything else.

Is that intended to be a sermon or do you truly believe that? What you call finding a short cut is how progress is made meanwhile I presume you are still cutting threads with a file.



With the unknowns that can happen with opening the halfnuts on an imperial lead screw while cutting a metric thread you could spend a lot of time watching the lathe spin and still end up with a damaged thread when you could be cutting the thread. It's really a matter of being efficient and practical when doing a job. Everything is unknown until it is learned and it is apparent from responses to the question I asked that there is no one here who could tell me how to calculate just how long one would have to wait, there is however a general opinion that it would be a long time.



The point is, you may never find the EXACT number of revolutions to get back to the correct engagement point and if you do, can you count that off each time without a mistake? I doubt it. I think you are wrong on a couple of points there, the synchronisation will most certainly repeat at some point and those with the approriate mathmatical skills could calculate it, but I would need just a little more study to do it and that is why I asked the question here. I can easily make an accurate counter that will run at much faster than spindle revolutions.


Enjoy your pizza.

Carld
08-23-2009, 05:37 PM
Bodger, it appears your somewhat of a smart ass and I will keep that in mind. With your attitude there is nothing more to say to you.

The Artful Bodger
08-23-2009, 05:47 PM
As you wish.

tattoomike68
08-23-2009, 05:59 PM
you boys kiss and make up..

carl is right, counting turns is silly.

The Artful Bodger
08-23-2009, 06:55 PM
.........counting turns is silly.

Maybe, but you dont know until you know how many turns might be involved.

The Artful Bodger
08-24-2009, 06:32 PM
After considerable head scratching I have decided to do the initial calculation to find the answers to the questions I raised.

This is a list of inch thread pitches from 10 through 60 threads per inch, the equivalent metric pitch and the number of revolutions of a 3mm pitch metric feed screw between one synchronisation point and the next.


10Tpi = 2.540000000mm 50 turns
11Tpi = 2.309090901mm 55 turns
12Tpi = 2.116666667mm 540 turns
13Tpi = 1.953846155mm 195 turns
14Tpi = 1.814285711mm 210 turns
15Tpi = 1.693333333mm 225 turns
16Tpi = 1.587500000mm 80 turns
17Tpi = 1.494117642mm 85 turns
18Tpi = 1.411111111mm 90 turns
19Tpi = 1.336842106mm 95 turns
20Tpi = 1.270000000mm 100 turns
21Tpi = 1.209523801mm 315 turns
22Tpi = 1.154545455mm 110 turns
23Tpi = 1.104347826mm 345 turns
24Tpi = 1.058333333mm 1080 turns
25Tpi = 1.016000000mm 125 turns
26Tpi = 0.976923077mm 390 turns
27Tpi = 0.940740741mm 675 turns
28Tpi = 0.907142857mm 420 turns
29Tpi = 0.875862067mm 145 turns
30Tpi = 0.846666667mm 450 turns
31Tpi = 0.819354837mm 155 turns
32Tpi = 0.793750000mm 160 turns
33Tpi = 0.769696967mm 165 turns
34Tpi = 0.747058822mm 170 turns
35Tpi = 0.725714286mm 1575 turns
36Tpi = 0.705555556mm 180 turns
37Tpi = 0.686486486mm 185 turns
38Tpi = 0.668421059mm 190 turns
39Tpi = 0.651282051mm 585 turns
40Tpi = 0.635000000mm 200 turns
41Tpi = 0.619512191mm 205 turns
42Tpi = 0.604761905mm 630 turns
43Tpi = 0.590697675mm 215 turns
44Tpi = 0.577272727mm 220 turns
45Tpi = 0.564444444mm 225 turns
46Tpi = 0.552173918mm 690 turns
47Tpi = 0.540425534mm 705 turns
48Tpi = 0.529166667mm 2160 turns
49Tpi = 0.518367346mm 2205 turns
50Tpi = 0.508000000mm 250 turns
51Tpi = 0.498039215mm 255 turns
52Tpi = 0.488461538mm 780 turns
53Tpi = 0.479245288mm 265 turns
54Tpi = 0.470370377mm 1350 turns
55Tpi = 0.461818182mm 275 turns
56Tpi = 0.453571429mm 840 turns
57Tpi = 0.445614039mm 285 turns
58Tpi = 0.437931039mm 290 turns
59Tpi = 0.430508471mm 885 turns
60Tpi = 0.423333333mm 900 turns



The first thing to notice is that although there are som rather long intervals there are a suprising number around 100-200. The 49TPI interval is about 20 feet and the much more common(?) 48TPI is almost as long.


So, how to use this information? If my thinking is correct one could engage the half nuts at the start of a thread cut and begin counting turns of the feed screw. The first cut would be made, the half nuts disengaged, the tool retracted and the carriage wound back to the start position and the half nuts renengaged at any multiple of feed screw turns according to the above list for the pitch being cut.

Manual counting is obviously out of the question but it is really easy to make an electronic counter from a basic pocket calculator (just bring out wires from the '=' button) but it really needs a counter that automatically repeats and for this a PC is ideal, even an old one.

We have seen several comments expressing the opinion that this is an inappropriate way of going about thread cutting and depending upon one's equipment this may well be so. However it may be of interest to the others, and this includes all those with lathes that are not easily or quickly reversed.

franco
08-24-2009, 08:48 PM
Bodger,

If you can get hold of the April 09 issue of Model Engineers Workshop magazine there is an article which discusses the reverse case to yours, i.e. making thread dial indicators for cutting metric threads with an 8 TPI leadscrew (Myford), at some length. If I read the article correctly the author made two thread dial indicator types, a conventional one with a 40T wheel driven by the lead screw for whole number metric pitches, and a spindle driven one incorporating a 127T gear and change gears for the half and quarter mm pitches, with different change gear combinations incorporated in the indicator for each thread pitch.

If you can find a copy of this article it might help to produce some ideas on how to approach the problem of making a thread indictor for TPI threads using the metric leadscrew.

franco.

The Artful Bodger
08-24-2009, 09:30 PM
Thanks Franco, that will be an interesting article to read and I am sure my local library carries that magazine.

Don Young
08-24-2009, 11:03 PM
I find it easier to understand lead screw sync by considering that sync when stationary is no different than when running. If you consider a very long leadscrew alongside a very long correctly metric threaded rod it soon becomes obvious that sync always occurs every 127 pitches of the metric thread and no place else. Assuming that the gearing is the same as when the rod was threaded, it makes no difference whether they are rotating or what the numbers of teeth on the gears are. Just take any metric pitch, multiply it by 127, convert to inches and you get the idea. 1MM X 127 = 127MM = 5 inches = 40 threads on an 8TPI leadscrew. A 40 tooth gear makes a perfect thread dial for that example. All other cases work just the same. If approximate gearing is used, the sync error is the same as the pitch error, which may be very significant over 127 pitches.

tattoomike68
08-24-2009, 11:22 PM
Maybe, but you dont know until you know how many turns might be involved.
After considerable head scratching I have decided to do the initial calculation to find the answers to the questions I raised.

This is a list of inch thread pitches from 10 through 60 threads per inch, the equivalent metric pitch and the number of revolutions of a 3mm pitch metric feed screw between one synchronisation point and the next.


10Tpi = 2.540000000mm 50 turns
11Tpi = 2.309090901mm 55 turns
12Tpi = 2.116666667mm 540 turns
13Tpi = 1.953846155mm 195 turns
14Tpi = 1.814285711mm 210 turns
15Tpi = 1.693333333mm 225 turns
16Tpi = 1.587500000mm 80 turns
17Tpi = 1.494117642mm 85 turns
18Tpi = 1.411111111mm 90 turns
19Tpi = 1.336842106mm 95 turns
20Tpi = 1.270000000mm 100 turns
21Tpi = 1.209523801mm 315 turns
22Tpi = 1.154545455mm 110 turns
23Tpi = 1.104347826mm 345 turns
24Tpi = 1.058333333mm 1080 turns
25Tpi = 1.016000000mm 125 turns
26Tpi = 0.976923077mm 390 turns
27Tpi = 0.940740741mm 675 turns
28Tpi = 0.907142857mm 420 turns
29Tpi = 0.875862067mm 145 turns
30Tpi = 0.846666667mm 450 turns
31Tpi = 0.819354837mm 155 turns
32Tpi = 0.793750000mm 160 turns
33Tpi = 0.769696967mm 165 turns
34Tpi = 0.747058822mm 170 turns
35Tpi = 0.725714286mm 1575 turns
36Tpi = 0.705555556mm 180 turns
37Tpi = 0.686486486mm 185 turns
38Tpi = 0.668421059mm 190 turns
39Tpi = 0.651282051mm 585 turns
40Tpi = 0.635000000mm 200 turns
41Tpi = 0.619512191mm 205 turns
42Tpi = 0.604761905mm 630 turns
43Tpi = 0.590697675mm 215 turns
44Tpi = 0.577272727mm 220 turns
45Tpi = 0.564444444mm 225 turns
46Tpi = 0.552173918mm 690 turns
47Tpi = 0.540425534mm 705 turns
48Tpi = 0.529166667mm 2160 turns
49Tpi = 0.518367346mm 2205 turns
50Tpi = 0.508000000mm 250 turns
51Tpi = 0.498039215mm 255 turns
52Tpi = 0.488461538mm 780 turns
53Tpi = 0.479245288mm 265 turns
54Tpi = 0.470370377mm 1350 turns
55Tpi = 0.461818182mm 275 turns
56Tpi = 0.453571429mm 840 turns
57Tpi = 0.445614039mm 285 turns
58Tpi = 0.437931039mm 290 turns
59Tpi = 0.430508471mm 885 turns
60Tpi = 0.423333333mm 900 turns



The first thing to notice is that although there are som rather long intervals there are a suprising number around 100-200. The 49TPI interval is about 20 feet and the much more common(?) 48TPI is almost as long.


So, how to use this information? If my thinking is correct one could engage the half nuts at the start of a thread cut and begin counting turns of the feed screw. The first cut would be made, the half nuts disengaged, the tool retracted and the carriage wound back to the start position and the half nuts renengaged at any multiple of feed screw turns according to the above list for the pitch being cut.

Manual counting is obviously out of the question but it is really easy to make an electronic counter from a basic pocket calculator (just bring out wires from the '=' button) but it really needs a counter that automatically repeats and for this a PC is ideal, even an old one.

We have seen several comments expressing the opinion that this is an inappropriate way of going about thread cutting and depending upon one's equipment this may well be so. However it may be of interest to the others, and this includes all those with lathes that are not easily or quickly reversed.


If you are on drugs please stop, if you are not on drugs plaese start. ask a doctor.

I like to get high but, I say to you... dont smoke ping pong balls

Don Young
08-24-2009, 11:39 PM
In regard to the numbers posted by The Artful Bodger, I may be all wet but it seems to me that 127 turns of a 3MM leadscrew would represent 381MM which is exactly equal to 15 inches. All whole numbered English threads would exactly complete whole threads in this distance and would properly re-synchronize. For 35TPI that would be 35 X 15 = 525 threads exactly.

The Artful Bodger
08-24-2009, 11:54 PM
Thanks Don, you have given me a bit to work through there.:)

The Artful Bodger
08-25-2009, 12:13 AM
I have just read post #30 Don and it appears the answer to my original query is 127 in all cases?

I have no idea where my logic went cranky but your's looks pretty good!:)

Paul Alciatore
08-25-2009, 03:39 AM
Consider this, a mechanical or electronic dial could be made for any thread you can cut. The difficulty is in waiting for the 127 or perhaps 2 or 3 or 4 times that number of revolutions to come around. And if you miss it, you have to wait all over again.

Such a dial should be made to work both forwards and backwards. A mechanical dial would do this automatically while an electronic one would have to be designed to do so, probably with a quadrature encoder. Now, after you make a pass and DISENGAGE the half nut, you run the lathe backwards and move the carriage back for another cut. The dial also moves backwards. Go a few turns extra and then reset the tool depth and go forward again. Engage the half nut at the point of synchronization and cut away. This will be a lot faster than waiting for 127 or more turns.

This could be done with a mechanical dial if it had perhaps a secondary dial that indicated a larger number of turns geared down from the main one, perhaps several dozen or so. Enough turns to prevent going back too far. (This would only work if you use reverse for synchronizing.) You would line both up before starting the cut and them look for both to line up again as you run in reverse and then forward again.

If you wanted it to work running in the forward direction only, you would need to use a 127 tooth worm for the secondary dial and a large diameter dial to allow reading to 1 division in 127. At least, I think this could be made to work.

The Artful Bodger
08-25-2009, 06:39 AM
Paul, enlightened by Don, I can say that in the case of turning imperial threads on a metric lathe you must either reverse the lathe or wait while the feed screw turns 127 times. OK, there is the other option of using a dog clutch or some mechanical means to move the feed screw in relation to the spindle.

By the time the cut is made, tool retracted and the carriage wound back and the tool advanced 127 turns is probably not very long to wait depending of course of how much thread is being cut. I think it would certainly be quicker than stopping and reversing a lot of lathes.

A mechanical counter or dial would certainly be possible but if it is a simple device with a 127 tooth wheel engaging the feed screw it would be about 5 inch diameter (on a screw of 3mm pitch). That might not be a problem depending on how clever the design is. I am not sure how a dial could be geared down as 127 is a prime number but thats not to say it is impossible.

My thoughts for an electronic counter would be a simple forward counter reading feed screw turns independant of the carriage, because that would be the easiest way to mount a counter but it has the (maybe slight) limitation that the carriage would have to be in the same position each time the half nuts are engaged as I think this is something that is compensated for by mounting the conventional threading indicator on the carriage.


John

oldtiffie
08-25-2009, 07:58 AM
I find it easier to understand lead screw sync by considering that sync when stationary is no different than when running. If you consider a very long leadscrew alongside a very long correctly metric threaded rod it soon becomes obvious that sync always occurs every 127 pitches of the metric thread and no place else. Assuming that the gearing is the same as when the rod was threaded, it makes no difference whether they are rotating or what the numbers of teeth on the gears are. Just take any metric pitch, multiply it by 127, convert to inches and you get the idea. 1MM X 127 = 127MM = 5 inches = 40 threads on an 8TPI leadscrew. A 40 tooth gear makes a perfect thread dial for that example. All other cases work just the same. If approximate gearing is used, the sync error is the same as the pitch error, which may be very significant over 127 pitches.

Well done Don.

I've been watching this thread and staying out of it until someone actually used some identities and constants as well as some applied shop math into the discussion.

For those that missed it:
- there are 25.4mm in an inch;
- 127 (a whole number) is 5 x 25.4

Don's formula may be expressed as: thread pitch (mm) x 127 = co-incident distance (mm)/25.4 = co-incident distance (inch).

An example:
thread lead = 0.75mm
lathe lead-screw = 1/4" (0.250")

Co-incident distance = (0.75 x 127)/25.4 = 95.25/25.4 = 3.75"

As the lathe pitch is 0.25" then the co-incident distance = 3.75"/0.25" = 15 pitches of the lead-screw.

If the lathe lead-screw was 6 tpi then the lead would be 1/6" = 0.166666"

In that case in the previous example 3.75"/0.1666666 = 22.5 pitches of the lead-screw.

If the lathe lead-screw was 8 tpi then the lead would be 1/8" = 0.1250"

In that case in the previous example 3.75"/0.125 = 30 pitches of the lead-screw.

That's all very well and good as a math solution but how to apply it and use it in a practical case on the lathe in the shop?

Easy - sort of (maybe??).

Set the gear train - and quick change gear-box (QCGB) if there is one - for the metric thread required.

Mount the screwing tool, engage the lead-screw and half-nut/s, and rotate the head to move the screwing tool to the point where it is to cease screw-cutting. Mark the lathe bed at the left end of the carriage position.

Put a machinist ruler on the lathe bed with the end of the ruler against the right/rear/back (tail-stock) end of the carriage.

Mark off the number of lines required on the lathe bed with the first line = co-incident distance and the rest of the lines each at the co-incident distance further on as well.

You have now marked off the limits of travel for the carriage.

The spindle must be stopped when the carriage gets to the left mark (ie when the cutting tool has reached its limit of left-travel).

Then and then only, release/dis-engage the half-nut/s, move the carriage right until it gets to the first of the marks for it on the lathe bed.

Engage the half-nuts.

Start the lathe spindle to commence screw-cutting.

There are both the principal and the application of it.

The lathe does not need to be reversed and the half-nut/s do not need to remain engaged.

This is counter-intuitive for those that are used to the "leave the half-nuts engaged and reverse the lathe spindle etc?" procedure.

It requires total concentration and lots of practice.

But it works.

The "lines" on the lathe bed can have "carriage-stops" instead if required.

I'd also suggest putting a mark on the lathe chuck and head-stock as "register marks" to stop the spindle at as it will make things a whole lot easier.

Frankly - and from experience - this is a good project to try and get right as an exercise, but I'd still opt for the traditional "reverse the spindle and keep the half-nuts engaged" process as it requires less concentration, is less error-prone and it works too.

This discussion thus far has been to do with cutting a metric thread on an inch lathe/lead-screw - as it seems is mostly the case in the USA.

The same principals and application applies to those of us in the "metricated" countries who have to cut "inch" threads on a metric lathe/lead-screw.

Barrington
08-25-2009, 09:00 AM
EDIT: Oldtiffie posted while I was preparing this, so there's a bit of crossover but hopefully this might be useful to someone...

I find thinking in terms of distances along the leadscrew makes things more difficult to visualise - it's easier just to imagine a gear on the spindle meshed with a gear on the end of the leadscrew, each with an index mark. (Real world intermediate gears don't change the calculations, only the overall ratio of spindle to leadscrew matters.)

Starting with both marks at 12 o'clock we are simply trying to work out how many turns the spindle/leadscrew will make before both marks are simultaneously at 12 o'clock again.

A way to visualise synchronisation :-
e.g. if gear one has 100 teeth and gear two 75, then they must synchronise after 75 turns of the first (when 7500 teeth have passed) and 100 turns of the second (again 7500 teeth have passed) - however in this example they first synchronised at 3 turns (300 teeth) to 4 turns (300 teeth), because 100:75 = 4:3. - It is important to reduce the gear ratio to the smallest numbers possible.

Why is 127 so fundamental?

For an 8tpi leadscrew, 1 turn (= 1/8" = 25.4/8 mm = 254/80 mm) = 127/40 mm

(Which is the same as saying the carriage travels 127 mm for 40 turns.)

So to cut a pitch of 1mm, we must gear 127 turns of the spindle to 40 turns of the leadscrew.
(For simplicity imagine this as a 40 tooth gear on the spindle meshing with a 127 tooth gear on the leadscrew.)

To cut a pitch of 2mm then we need to double the speed of the leadscrew and so would gear 127:80 and so on:-

2.0_____127:80
1.75____127:70
1.5_____127:60
1.0_____127:40
0.75____127:30
0.7_____127:28
0.5_____127:20
(mm)____127: (mm*40)
etc.

These are all simple ratios, not reducable. The common feature is that these will all synchronise after 127 turns of the spindle.



For the opposite case of cutting imperial threads with a metric leadscrew:-

For a 3mm leadscrew, 1 turn (= 3mm = 3/25.4 " = 30/254 ") = 15/127 "

(Which is the same as saying the carriage travels 15" for 127 turns.)

So to cut a pitch of 1 tpi i.e 1" we must gear 15 turns of the spindle to 127 turns of the leadscrew.
For simplicity imagine this as a 127 tooth gear on the spindle meshing with a 15 tooth gear on the leadscrew. (Don't try this at home !)

To cut a pitch of 2 tpi i.e 0.5" then we need to halve the speed of the leadscrew so would gear 30:127 and so on:-

1___1________15:127
2___0.5______30:127
4___0.25_____60:127
8___0.125____120:127
16__0.0625___240:127
tpi__1/tpi_____(tpi*15):127
etc.

These again are irreducable ratios and the common feature is that they will all synchronise after 127 turns of the leadscrew.


For any ratio including 'approximation' ratios (those not involving a 127 changewheel) simply calculate the overall ratio and express it as the simplest ratio of (spindle turns:leadscrew turns). The first number number provides the turns of the spindle between synchronisations and the second the turns of the leadscrew between synchronisations.

Cheers
.

oldtiffie
08-25-2009, 09:42 AM
These again are irreducable ratios and the common feature is that they will all synchronise after 127 turns of the leadscrew.

Thanks Barrington.

That certainly improves the answers and widens the scope of the application.

The reason that I put multiple "co-incidence marks" on the lather bed to the right of the carriage was to provide multiple "127" marks for the cases where the thread being cut was longer that the first or a single "127" distance.

That is to say that co-incidence would occur at:
- 127 = 127 x 1;
- 254 = 127 x 2;
- 381 = 127 x 3;
- 508 = 127 x 4;
etc.

My thanks again to Don Young.

The Artful Bodger
08-25-2009, 04:12 PM
Oldtiffie, thank you for that clear explanation of the process however it does not directly answer the question I asked which was to do with cutting imperial threads on a metric machine.

If I understand you your method effectively counts feed screw turns by measuring carriage travel? This would be accurate in theory but in practice it may be impossible when doing inch threads on a metric lathe with 3mm lead screw as the distance is 15" and depending on the size of the workpiece there may not be that amount of carriage travel available between the headstock and the tailstock.

Furthermore, and this is in consideration of those who have shown such concern at having to wait for 127 turns of the feed screw, your method is even slower as it involves not only that time interval but also the time to stop, wind back and restart.


Barrington, thank you for shedding more light on this subject too.

Peter.
08-25-2009, 04:38 PM
I guess I'm lucky with my variable-speed belt drive, I could run a slow threading pass, back out the tool, reverse the motor and swing the speed lever to whizz the tool quickly back the the start of the thread ready for another pass.

Since threading is usually done at slow speed, thread-counting seems like a tedious way of dong it unless you have no option.

oldtiffie
08-25-2009, 07:51 PM
Oldtiffie, thank you for that clear explanation of the process however it does not directly answer the question I asked which was to do with cutting imperial threads on a metric machine.

If I understand you your method effectively counts feed screw turns by measuring carriage travel? This would be accurate in theory but in practice it may be impossible when doing inch threads on a metric lathe with 3mm lead screw as the distance is 15" and depending on the size of the workpiece there may not be that amount of carriage travel available between the headstock and the tailstock.

Furthermore, and this is in consideration of those who have shown such concern at having to wait for 127 turns of the feed screw, your method is even slower as it involves not only that time interval but also the time to stop, wind back and restart.


Barrington, thank you for shedding more light on this subject too.

Thanks TAB.

I addressed the issue in terms of cutting a metric thread on an inch lathe as most seem either more familiar with and have more need for that as they seem to have inch machines.

As I said, and as Barrington showed, the principles are universal either way ie cutting metric threads on an inch lathe or vice versa.

I take your point regarding the length of the "127 turns", but it is quite possible to use "62.5" turns and to advance or retard the cutting tool by 1/2 lead by using the cross-slide set at 90 degrees. But it would be hard to keep track of and is one more source of possible error and/or confusion in a process that is quite difficult enough as it is.

Before I even attempted this process, I would see if I could find a thread that was "native" to the lead-screw that would suffice. For example:
- on an inch lead-screw:
a pitch of 3mm is 0.1181" which at 1/0.1181 = 8.4667 tpi ~ 8.5 tpi which means that if that small error is acceptable, the 8.5tpi screw and the say 8 tpi lead-screw will co-incide at 2" lead-screw travel which is 16 pitches on the lead-screw and 1 full turn on a 16-tooth geared thread-chasing dial. So, that being the case you could have the half-nuts engaged anywhere along the available carriage travel provided that you "dropped in" at the same point on the thread-chaser dial.

In other words. you would be very closely approximating a metric thread by cutting an inch thread.

That being the case, you can keep the spindle running and disengage and engage the half-nuts anywhere you like - providing you use the chasing dial correctly.

The converse also applies when cutting inch thread on a metric lathe with a thread-chasing dial with the right number of teeth on it.

I rather hope that Barrington sees this and "tunes in" and reads it and adds his comments as he always does a fine job there as he brings up points that I've either missed or expands on what I've said so that between us we get a better answer that I get on my own.

This is all just applied "Shop Math" and a bit of logic and lateral thinking.

The Artful Bodger
08-25-2009, 08:22 PM
Oldtiffie, I see what you mean about using a close approximation of the thread using a ratio from the 'other family' but would the error not increase (i.e. accumulate) with each pass? Granted, it might be insignificant anyway.

It appears to me that the challenge of imperial threads on a metric machine are a little different to those cutting metric threads on an imperial machine and this comes about from the way the threads are defined. In the inch world threads are 'threads per inch' whereas the metric threads are expressed in the form of 'millimetres per thread'. No doubt the principles are the same but the thinking required is different, on the metric machine cutting inches the interval is always "127 turns" as shown by Don but on an inch machine cutting metric the intervals differ and as you have shown may be quite short. (I hope I got that right! :rolleyes: )

In my case I have a metric machine but there is a lot of inch material in our environment so that is the direction I have been thinking in.

I still think a 127 counter (and a carriage stop) is the practical way to do this, arranging the counter is the challenge!:)

oldtiffie
08-25-2009, 08:50 PM
Oldtiffie, I see what you mean about using a close approximation of the thread using a ratio from the 'other family' but would the error not increase (i.e. accumulate) with each pass? Granted, it might be insignificant anyway.

It appears to me that the challenge of imperial threads on a metric machine are a little different to those cutting metric threads on an imperial machine and this comes about from the way the threads are defined. In the inch world threads are 'threads per inch' whereas the metric threads are expressed in the form of 'millimetres per thread'. No doubt the principles are the same but the thinking required is different, on the metric machine cutting inches the interval is always "127 turns" as shown by Don but on an inch machine cutting metric the intervals differ and as you have shown may be quite short. (I hope I got that right! :rolleyes: )

In my case I have a metric machine but there is a lot of inch material in our environment so that is the direction I have been thinking in.

I still think a 127 counter (and a carriage stop) is the practical way to do this, arranging the counter is the challenge!:)

Thanks TAB.

A couple of points.


Oldtiffie, I see what you mean about using a close approximation of the thread using a ratio from the 'other family' but would the error not increase (i.e. accumulate) with each pass? Granted, it might be insignificant anyway.

Nope. In the 8tpi example I was cutting an 8tpi thread on an 8tpi lead-screw. The error is constant at ((8.500 - 8.467)/8.500)/100 = 0.39%

- on an inch lead-screw:
a pitch of 3mm is 0.1181" which at 1/0.1181 = 8.4667 tpi ~ 8.5 tpi which means that if that small error is acceptable, the 8.5tpi screw and the say 8 tpi lead-screw will co-incide at 2" lead-screw travel which is 16 pitches on the lead-screw and 1 full turn on a 16-tooth geared thread-chasing dial. So, that being the case you could have the half-nuts engaged anywhere along the available carriage travel provided that you "dropped in" at the same point on the thread-chaser dial.


It appears to me that the challenge of imperial threads on a metric machine are a little different to those cutting metric threads on an imperial machine and this comes about from the way the threads are defined. In the inch world threads are 'threads per inch' whereas the metric threads are expressed in the form of 'millimetres per thread'. No doubt the principles are the same but the thinking required is different, on the metric machine cutting inches the interval is always "127 turns" as shown by Don but on an inch machine cutting metric the intervals differ and as you have shown may be quite short. (I hope I got that right! :rolleyes: )

It is true that "inch" threads are usually specified as "threads per inch" - eg 1/2"-13 - and that metric threads are specified as the lead - eg. M10-1.5 - but it is of no consequence.

Lead = 1/tpi and tpi = 1/lead.

The problems are best solved by reducing all threads to their respective leads.


In my case I have a metric machine but there is a lot of inch material in our environment so that is the direction I have been thinking in.

Answered previously.


I still think a 127 counter (and a carriage stop) is the practical way to do this, arranging the counter is the challenge!

An odometer from a car will be a good place to start. Or have a look at the simple counter (analogue) on a "pedometer" (aka (Surveyors) "Measuring wheel") which measures distance as a function of number of wheel turns) at any good tool shop - or eBay.

http://en.wikipedia.org/wiki/Surveyor%27s_wheel

http://www.justtools.com.au/category319_1.htm

http://www.google.com.au/search?hl=en&q=measuring+wheel&btnG=Search&meta=

http://en.wikipedia.org/wiki/Trundle_wheel

The Artful Bodger
08-25-2009, 09:45 PM
Oldtiffie, all good gen. Thanks.

Don Young
08-25-2009, 10:13 PM
There are a lot of methods which work to re-establish sync between metric and English threads. For an English thread and metric leadscrew, I think the best arrangement would be a 127 pitch thread dial. I think such a critter actually exists for some lathes, using fine pitch internal gearing since a 127 tooth leadscrew gear would be so large. I have seen what looks like a veeder-root (odometer style) counter on some import lates and I suspect these might be used to count 127 leadscrew revolutions (or pitches when the screw is stopped). As someone mentioned, you can always use the normal thread dial with the half-nuts open if you use the same mark and keep track so you can back up the spindle to put the dial on the same revolution (not gone completely around). That puts everything back exactly the same as before the half-nuts were opened.

When cutting metric threads you could install a special gear on the leadscrew. The common 4 inch dial (32 teeth on an 8TPI screw) works for 0.8, 0.4, and 0.2 MM. For 1.0, 0.5, and 0.25 MM you need a 40T gear.

Don Young

oldtiffie
08-25-2009, 10:48 PM
There are a lot of methods which work to re-establish sync between metric and English threads. For an English thread and metric leadscrew, I think the best arrangement would be a 127 pitch thread dial. I think such a critter actually exists for some lathes, using fine pitch internal gearing since a 127 tooth leadscrew gear would be so large. I have seen what looks like a veeder-root (odometer style) counter on some import lates and I suspect these might be used to count 127 leadscrew revolutions (or pitches when the screw is stopped). As someone mentioned, you can always use the normal thread dial with the half-nuts open if you use the same mark and keep track so you can back up the spindle to put the dial on the same revolution (not gone completely around). That puts everything back exactly the same as before the half-nuts were opened.

When cutting metric threads you could install a special gear on the leadscrew. The common 4 inch dial (32 teeth on an 8TPI screw) works for 0.8, 0.4, and 0.2 MM. For 1.0, 0.5, and 0.25 MM you need a 40T gear.

Don Young

Thanks Don.

That 127 tooth gear might not be all that large - depending on the user wanting to make it as well as the inconvenience of it!!

For a 3mm pitch lead-screw (as the OP seems to have) the Pitch Circle circumference would be: 3 x 127 = 381mm/pi = 121.276mm Pitch Circle Diameter (PCD).

Add say 3mm for 2 x addenda (guessing here as I haven't checked) and the outside diameter (OD) is in the order of 125mm ~ 5"