Lathe gear calculation

[Evan LewisParticipant @evanlewis60038]

I have written an online computer program for calculating gear trains to cut metric or imperial threads on metric of imperial lathes. You might find it helpful. This is the URL link:

RideTheGearTrain.com

I would love to hear whether people find it useful.

[JasonBModerator @jasonb]

Where you website says "link to program" it does not for me as it goes to RideTheGearTrain.com

[Thor 🇳🇴Participant @thor]

Hi Evan,

There is no link that works for me in your " 05/04/2022 10:26:05 " post, I guess it should be?:

***Link***

 I have just checked a few gear-trains and they seem to agree with my Chinese lathe. Thanks for sharing.

Thor

Edited By Thor 🇳🇴 on 05/04/2022 11:21:57

[Evan LewisParticipant @evanlewis60038]

JasonB,

Yes it looks as though you found RideTheGearTrain OK. It should even turn up on a Google search, as long as you don't insert spaces.

It has now had 1.2 million page loads. One company reported selling 10 million mini-lathes, and that was just one Chinese company!

Thanks Thor,

Your ***Link*** works. I just couldn't see how to insert an active link, but I just found that funny looking icon that looks like the back and of a petrol tanker doing 10 km/h. So here is another link to my program.

**LINK**

There was on older link via one of my other web sites HeroSteamEngine.com. But the program has been moved to its own site. The old address should forward you to the current address.

It is listed on http://www.lathes.co.uk/latheparts/page14.html but unfortunately they copied some of my very old code onto their web site. But when you try to run their program, it forwards to the correct site RideTheGearTrain.com.

Hope that is not too confusing!

Extensive help files have been written, with help buttons beside most entries. This includes a discussion of the weird method used to label gears in the gear trains of Chinese lathes.

You will see I have written some other apps on the same site. See the menu items with orange buttons. The app to calculate drill sizes based on % depth of thread is perhaps unique. After you specify the size of your thread it lists a range of metric, imperial and 'number' drills and calculates the percent thread depth for each one. The idea is that you might not have the exact recommended drill but can probably find one in your kit that gives a reasonable percentage.

Another set of apps are for cutting speeds. It shows tables of recommended speeds and provides a calculator and table generator.

Cutting speeds are dependent of hardness so there is a section on the hardness of various metals and then hardening and tempering of steel.

Another app is for calculating the tailstock offset required to cut a taper (eg Morse taper) by the tailstock offset method.

The next update will remove text from entries that should be numeric. I found people were trying to include units such as mm.

Evan

[Evan LewisParticipant @evanlewis60038]

PS For Programmers:

You may be interested to know that RideTheGearTrain was written using PHP. This is a general purpose programming language that runs on nearly all servers. It generates HTML code which is sent back to the users computer to display a web page. If the user responds to a question (or data INPUT) on the web site, the program has to run again from scratch, but the new and old data is transmitted back to the server in the URL. It gets very cumbersome, but has the advantage that the user can save the URL and use it again at any time. It means that the user does not have to run any code or run plugins. But it requires the user to be online.

[Nigel Graham 2Participant @nigelgraham2]

With great respect this thread on threads puzzles me.

Why do people keep re-inventing the [change-]wheel?

Or are they trying to cut very odd threads on lathes made to cut standard threads from standard lead-screws?

My Myford 7 and Harrison L5 came with change-wheels and tables for a wide range of Imperial and metric threads; but the sums are all simple. The ML7 can use its normal wheels for the smaller, common mm pitches, tabulated on a printed plate inside the guard; but I have treated it to an extra, 63T, wheel. The Harrison has a 127T wheel in a change-wheel set further enhanced by a 3-speed feed-shaftl/lead-screw gear-box..

I know many people use the small lathes sold by the likes of Axminster Tools, and I have one not yet "commissioned", These are metric but I believe you can buy additional wheels for cutting Imperial threads if wished. (I am not sure if Axminster still sells metalworking machines, though.)

For anything else, hardly likely more than once in a Blue Moon Leap Year, the calculator would suffice, but I did create an MS 'Excel' spreadsheet to determine close BA and ISO-M threads for my third lathe, an EW Stringer, with its own 25 to 65 X 5T wheel set.

No need to write computer programmes unless you enjoy programming for its own sake, and arithmetically all my spreadsheet does that would be time-consuming by calculator or slide-rule, is facilitate homing in on least-error wheel combinations for cutting up to perhaps 10 turns to acceptable accuracy – so covering the majority of tasks. I simply, manually tweaked the tooth-counts by "inspection" and rejected those results too inaccurate for fine work. Oh – and it then lets me print it.

'

*Holtzappfel also recognised chasing-lathes, which basically copy the thread from a master-screw rather than generate it by gearing; and a cam-type fitting for short thread-lengths, The latter controlled the tool travel by a cam plate or bar set at the appropriate angle across the spindle axis, but I don't know how it related that to the spindle revolution. I have a copy of ME from sometime in WW1, in which a contributor describes a chasing attachment he made for his simple lathe, from a length of threaded bar.

[Ady1Participant @ady1]

26.000000 -0.000000 n[40] n[50] n[25] n[65]
26.000000 -0.000000 n[40] n[60] n[30] n[65]
26.000000 -0.000000 n[40] n[65] n[25] n[50]
26.000000 -0.000000 n[40] n[65] n[30] n[60]
26.000000 -0.000000 n[40] n[65] n[35] n[70]
26.000000 -0.000000 n[40] n[70] n[35] n[65]

[Martin ConnellyParticipant @martinconnelly55370]

Nigel, I think a lot of people have, or are going to get, second hand lathes that are missing some gears and almost certainly missing the original manuals. Added to this is the fact that some old lathes were never intended to cut metric thread pitches. So there is a constant demand for information on how to do this with either what people have got or what is the minimum they need to buy or make to enable them to do what they want.

Martin C

[Howard LewisParticipant @howardlewis46836]

In this instance, the requirement is to cut a 26 tpi thread.

The picture of the chart on the lathe seems to show that 26 tpi is one of the pitches available on line. A.

All that is needed is to set up a gear train as the chart indicates.

Howard

[SillyOldDufferModerator @sillyoldduffer]

Posted by Nigel Graham 2 on 05/04/2022 13:23:06:

With great respect this thread on threads puzzles me.

Why do people keep re-inventing the [change-]wheel?

… the sums are all simple. …

Can you share your method please Nigel? I for one don't find the sums simple!

For example, if my lathe has:

H A
… |
C B
|
D H

Leadscrew 3:1
Headstock 40t
Gears Available: 20, 30, 45, 50, 60, 60, 65, 70, 75, 80, 85

What's the closest gear combination you would recommend for:

• 2BA 0.81pitch 31.25tpi
• 3.25tpi BSW?

Which of the two example does my lathe get closest too?

And, for extra points, what gear(s) should be bought for my lathe to get the threads spot on?

Please show all your workings!

Dave

[Evan LewisParticipant @evanlewis60038]

Hi Nigel,

Yes, despite being retired, I am a programming addict like others on this thread on threads. Of course many absolute beginners using lathes don't know how to calculate the gear train gear ratio. That included me when I started on this path.

Martin Conelly pointed out a common use for the programs. People often have missing or damaged gears and manuals and want to find the best and cheapest way to cut metric on an imperial lathe etc.

A number of users have contacted me and I have helped them out by adding their lathes with gearboxes. This included a 16 foot imperial lathe being used to overhaul metric hydraulic cylinders for diggers up to 800 tonnes in Canadian open caste oil-sand mines…etc.

But probably the most common use is trying to cut odd threads on minilathes, eg a camera thread 27 TPI.

The problem is the weird labelling used on minilathes. It is OK if you just want to do exactly as they say, but if you want to actually understand what it is doing, it is extremely confusing. As mentioned above, they never mention the spindle gear that has been machined into the spindle (usually 40 teeth). If the user is not aware of this, it is impossible to calculate gear ratios.

Secondly, they do not label the gears in the order that power is transmitted through the gear train. Instead they are labelled by the order in which you mount the gears on the studs. So you cannot simply say that eg Z2 drives Z3 because that may not always be the case. It is so confusing that I recommend minilathe users who want to use my program should give up on that crazy labelling and use the labels shown in my program, which are in the correct order for calculating the gear train ratio.

I have extensive discussions of these issues, and many others, in my help files (probably too verbose).

I had even forgotten why I started writing RideTheGearTrain. I suddenly remembered! I was considering making an Archimedes Screw for the Hero's Steam Engine and wanted to know the maximum pitch that could be achieved with the gears I had. It grew like Topsy!

Many users contacted me wanting special features and most of these have been added, including how to cut worm gears to a particular modulus or diametric pitch. I had to teach myself.

Feed rates are less important but have been included.

The ability to look up various metric and imperial standard threads from tables has been added, eg bicycle threads, Tyre valve threads, Edison light bulbs etc (just for fun). These tabels gave a lot of extra information in the specifications and these have been used to calculate torque/stress etc as shown in the results.

Any further suggestions for RideTheGearTrain.com ?

Nigel

I found your note about a system using a leadscrew set at an angle to the axis of the lathe to modify the effective pitch of the leadscrew. At least I assume that is how it works, but I had never heard of it. I suppose the effective pitch would be the cosine of the angle times the pitch of the leadscrew. The variation in pitch would be limited by the offset angle that could be achieved. Perhaps they had a range of leadscrews to choose from. VERY INTERESTING ! Of course, very early lathes had a range of leadscrews and did not rely on the gear train. It may have been easier to produce leadscrews than gears. A cam system might just use a cam to nudge the carriage a bit further as the half-nuts move along the leadscrew. Again I am guessing.

Evan

[Nigel Graham 2Participant @nigelgraham2]

My point was that people seem to be approaching what should be a relatively simple task in rather roundabout ways.

The ratios can be awkward to find, I agree, if you need cut a thread that does not give neat ratios of integers, and for those you do very often need a calculator at least.

Writing a special programme is fine as a challenge in itself, if you can do that, but otherwise a spreadsheet like "Excel" will give you everything you need to create a table you can print, laminate and keep by the lathe.

What really matters is our being able to cut the intended thread to acceptable accuracy over the length necessary.

I have cut a "bastard" thread on my EW: a non-standard metric thread on a fitting for a special purpose, the manufacturers' trick to force buying only its range for a very different purpose! It took a lot of trial-and-error on paper as this inch-based lathe in standard trim cannot accommodate more than a 2-step compound train. The thread, in brass, was a bit rough but the fitting fitted suitably well.

'

Dave –

I am not very numerate and having to calculate change-wheels does make me think hard. I can't usually do the sums in my head, but I don't cut threads very often, don't cut non-standard threads unless really necessary, and have various text-books to help me if the thread is not in the manual. (My L5 can catch me out because its spindle pinion is not the size given in the hand-book!) Even so I don't usually find it too difficult. Sometimes the book explanations seem to make it harder than it should be, but I keep in mind 2 things:

1) The thread being cut is normally always of more TPI than the lead-screw. So that dictates which way round the fractions go: the wheels get bigger from spindle to lead-screw. (If coarser, it needs care in thinking how to do it without undue cruelty to the machine.).

2) The starting values on inch-sizes are the leadscrew TPI and the cut TPI: not the pitches. Using inch pitches can immediately make the ratios very awkward. I think for all-metric lathes, pitches may be better as they tend to be of simpler numbers than inch-pitches.

Then my first move: multiply the ratio's numerator and denominator by 5, because most of the wheels are in steps of 5.

You won't find exact matches for many threads not forming simple ratios with the lead-screw, and I never claimed you can. I stressed calculating a table giving the least error within the 10 or so turns that should cover most purposes, such as stud ends.

Usually I screw-cut the bulk of the thread then finish to profile and size with a chaser, die or die-nut. In any case, a single-point tool alone gives a sharp profile, though the crest can be blunted by shaving a tiny flat along it.

I don't know what your lathe is, but it must be pretty massive to be able to cut a 3.25tpi thread happily. My L5's lead-screw is 4tpi and we know lathes should not really be expected to screw-cut coarser than their own lead-screws.

I looked at cutting BA threads but really, would use a die. Same with small ISO-mm threads. I suppose I would screw-cut threads for very fine clock and scientific-instrument making; but would equip my workshop appropriately.

'

Evan –

If change-wheels are missing, the manufacturer's manual won't help calculate alternative combinations. The only real solution may be to obtain replacements that may have to be "stock" gears, with the choice greater and prices lower for metric (module) wheels. However, it might enforce a set so the tooth-geometry is compatible within each combination, if as likely, the original tooth-counts and pressure-angles cannot be replicated.

As a general point though, cutting mm threads on an Imperial lathe, or vice-versa, normally relies on a 127T or (slightly less accurately) 63T conversion wheel in the train, but on an 8TPI leadscrew, some common ISO-M pitches are obtainable within sensible accuracy with a compound train.

'

The system Holtzappfel describes does not use a lead-screw at an angle.

It is a long time since I had read it, in Vol.2 of the treatise, pub. in facscimile by TEE Publishing. Re-reading, I had misunderstood it as a lathe attachment.

In fact it was a rather complex, manual machine dedicated to cutting up to 10 turns of any reasonable pitch, left or right-handed, on the pre-turned work-piece. Developed from fusee-cutting engines (conical "threads" ), a pinion on the work-spindle engaged a rack at right angles to the spindle, to move a follower along a cross-guide set an angle to the work axis, the follower itself traversing the tool-slide. As the lead cut is a function of the angle (the helix angle?) it can be set to any lead, LH or RH, within the machine's range. The writer points out potential sources of inaccuracy: the standard of the machine itself, rack-and pinion backlash and the accuracy of setting the guide.

Holtzappfel also shows a simpler machine whose inclined cutter formed the first turn, then a follower engaged the growing thread in a self-copying arrangement – like a tap and die. It is hard to see how this was set to a definite pitch; but until Maudsley, Whitworth, Thury et. al, screws were never standard, very consistent, or mutually compatible

The book covers a wide range of ways to form threads, and even shows a 16C precursor to our familiar change-wheel system, using cords and change-pulleys instead.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Nigel Graham 2 on 06/04/2022 12:44:09:

My point was that people seem to be approaching what should be a relatively simple task in rather roundabout ways.

…

Sometimes the book explanations seem to make it harder than it should be, but I keep in mind 2 things:

1) The thread being cut is normally always of more TPI than the lead-screw. So that dictates which way round the fractions go…

2) The starting values on inch-sizes are the leadscrew TPI and the cut TPI: not the pitches. Using inch pitches can immediately make the ratios very awkward. I think for all-metric lathes, pitches may be better as they tend to be of simpler numbers than inch-pitches.

Then my first move: multiply the ratio's numerator and denominator by 5, because most of the wheels are in steps of 5.

…

I looked at cutting BA threads but really, would use a die…

'

…

Ah, I see Nigel's method isn't a general purpose way of calculating the change gears needed to achieve a particular ratio. Rather, the approach restricted to Imperial Threads and a Gear set stepping 5 teeth at a time, which happens to suit Nigel's equipment, but is less useful to many others.

The computer programs on offer solve the change gear ratio problem in a much wider sense, doing TPI and Metric with no assumptions about the change wheels available or the lathe.

My Far-Eastern lathe has a metric lead-screw, but it cuts imperial threads with the same change gear set. And the Imperial Version works the same way. Neither machine has a 63 or 127 toothed conversion gear: the necessary ratios are generated by the ordinary gears provided. On machines of this type, and whenever the owner has lost the manual, or a second-hand gear set is missing a few, it's useful to know what can be done with the gears available.

Solving ratios for any lathe and any gear-set requires either difficult paper and pencil maths (continued fractions), or calculating all possible combinations and eliminating irrelevant answers. Calculating all the permutations is a mass of tedious arithmetic that would take a human several days to do: a computer does the job in milliseconds. The brute force calculation is so fast it's rarely worth programming the continued fraction method that allow humans who understand it to find approximations in tens of minutes. Computerised continued fractions are useful if you need to know which new tooth count is needed to produce an unusual ratio – like BA, or one of the antique systems.

Using bought in dies is always good, but what happens if I need to make a BA tap or die myself?

May seem roundabout, but the programs are valuable to anyone who has a more than a trivial ratio to calculate.

Dave

[Nigel Graham 2Participant @nigelgraham2]

Dave – I'm afraid you've misunderstood me.

What I suggested is not limited to specific threads and change-wheel sets at all, even if the set in your own workshop is incomplete. I used a spreadsheet but pushed it to BA and ISO metric threads to see what was possible and sensible, with the equipment I have – perhaps extended with some intermediate wheels if they are obtainable and of correct form!

Firstly, consider what you threads you need cut, and whether ones like BA are better produced with a die anyway. Producing a table that offers BA threads was more speculative exercise than practical proposition, partly due also to the difficulty of grinding the tool to proper form.

What is not in the information I have for my lathes, is calculable even for a non-mathematician like me, and looking at the finished spreadsheet I realised a lot of it was really of little practical value.

.

Secondly, unless your engineering really does involve generating a lot of very odd threads indeed, whose calculations do not shake out with with just a calculator, is it worth the added work of writing a complete programme for something readily done in 'Excel'? Obviously both need the requisite computer-using ability, but using a published MS 'Office' or 'OpenOffice' spreadsheet template is simpler and the donkey-work has been done for you.

Creating an entire programme requires its language compiler and your understanding how to write it. Fine if you write lots of programmes anyway but hard to justify for a single task; and besides, how many of us have that specialist knowledge? Or indeed could run it on our own computers if written by someone else?

'

Most lathes have lead-screws of very common tpi or mm size, so the loss of the chart for one make of lathe is no great problem. If you know its lead and the tooth-count of the spindle pinion then the change-wheel chart for any machine with the same two values will at least overlap your own lathe's wheel-set.

I do agree that manually calculating screw-cutting for awkward threads, including BA, can be a chore manually, but even with the aid of a computer it still must match the gears you have or can obtain. I doubt you've be able to gain more than a close match needing finishing with a die anyway, unless you can obtain suitable additional gears.

The problem with replacing missing, or adding extra, wheels is not tooth-counts though. The original sets are mainly in equal steps so most of the gaps are fairly obvious, but may include others like 32, 38, 42, 63 or 127. The real difficulty is of ensuring replacements of compatible tooth-form (DP or Mod, and pressure-angle), and you may even need buy an entire set.

Manually calculating wheel combinations for odd threads can be a chore where the TPI or pitch really is a strange fraction of the lathe's lead, but how often do you need do that?

I have seen "continued fractions". I forget the context but it was not screw-cutting and none of my engineering reference-books old and new use it for that. They show just simple vulgar fractions – the operation may need only one, or you may need multiply it by one, perhaps two, more. The chore comes when you need start manipulating tooth-counts to reduce an error to acceptable level; caused by a starting ratio whose denominator is itself a fraction, and by not having suitable intermediate change-wheels anyway.

(E.g. 2BA = 31.3tpi, difficult on an 8tpi lead-screw. The nearest likely match for a small lathe, 20/75 then 60/63, still needs that 63T wheel but gives 0.002" error per turn… I used a calculator but it did take several iterations. However, BA threads are a geometrically-progressing metric series, and 2BA's 0.81mm pitch might be fairly easy on a metric lathe. 0BA is of 1mm pitch – as is M6 Coarse, though of different form.)

[Howard LewisParticipant @howardlewis46836]

Being a coward, I cut BA threads, (And where possible, threads smaller than 13 mm ) with Taps and Dies.

Without a program to help, selecting the right change gears can be a time consuming, and intellectually challenging, business, and sometimes call for a gear that is not in stock, or commercially available. In which case, it may be possible to buy or to make such a gear, which will be useful, if ever the process needs to be repeated at some future time..

EXCEL can be a time saver in such situations,, occasionally, because of the speed with which the result of a "What if?" can be seen..

Sometimes, a "non standard" gear for one's machine can be produced by modifying one from a different machine. On other occasions, if the need drives, one has to accept a train that produces a thread with an acceptable error.

Brian Wood's book shows table for various lathes and where the translation is not exact, the percentage error, so that a decision can be made as to acceptability for a particular application.

For short threads, such errors may be lost in the clearance between the internal and external threads. Errors become more important when long lengths of thread are required.

How often do we need to, produce studding nor a Leadscrew?

Unless it is impossible to change or modify a combination, the solution may be to use a slightly different thread form mor size.

After all, in some instances, M6 may well clamp things together just as well as 1/4 BSF, or vice versa..

Even if it means using / producing a hybrid with the theoretically correct hexagon, but a different hread. This happens in industry, either to provide clearance or convenience (On vehicles, I have encountered a 12mm nut with a M8 thread, or a 5/8" hexagon with a 5/16 BSW thread )

Unless it is imperative to be correct, cosmetically, the ultimate objective is to provide a clamping force, or adjustment between two components.

Howard

[duncan webster 1Participant @duncanwebster1]

BA threads were devised by a committee, I reckon none of its members had ever driven a screw cutting lathe. I could say the same for metric threads, not quite as bad but you need three gears for your thread indicator wheel rather than just the one for imperial. If only they had specified metric as threads per 24mm, then you'd only need one gear

[SillyOldDufferModerator @sillyoldduffer]

Posted by Nigel Graham 2 on 07/04/2022 10:48:32:

Dave – I'm afraid you've misunderstood me.

…

Possibly, but you said you had a simple method, so simple it's not worth writing a program that anyone can use. I wondered what the method is and asked to see a worked example. Instead I'm told it's a spreadsheet. That's fine, but can you explain the maths please?

Dave

[Michael GilliganParticipant @michaelgilligan61133]

It appears that there are a few typo’s in the transcription of this chart …but it’s a very good start if you want to cut BA threads on an Imperial lathe :

**LINK**

https://www.greenbuildingcentral.us/machine-tools/screwcutting-ba-threads.html

MichaelG.

[Evan LewisParticipant @evanlewis60038]

Interesting discussion!

Nigel, you said:

"Creating an entire programme requires its language compiler and your understanding how to write it. Fine if you write lots of programmes anyway but hard to justify for a single task; and besides, how many of us have that specialist knowledge? Or indeed could run it on our own computers if written by someone else?"

My computer program gets around all of these issues. When you go to my web site the computer program on the server runs automatically. You don't need any computer skills other than general use of web sites. It does not download the program to your computer and does not require a local compiler.

I would appreciate it if some people could run the program and then give some feedback. I believe you will not want to use a calculator, spread sheets or pencil and paper once you have run the program.

RideTheGearTrain.com

There are a number of entries required because this is a general purpose program that can be used for almost any lathe, to cut any thread. The only real restriction is that if your lathe has a lead-screw gearbox, the program needs to know the gear ratios in the gearbox. There are about 25 gearboxes pre-programmed including common lathes like Myford, Boxford and all other South Bend 9" clones. If your lathe is not on the list you can contact me and I will add it. Furthermore, there is a custom lathe option which walks you through the process of adding your gearbox. This just uses the thread cutting table usually attached to the gearbox, so you do not need a manual, and you do not need to take it apart to count gears.

If your lathe is on the list, or if it does not have a gearbox the process is much easier. There are 5 data-entry steps. Originally they were all on one web page, but one person complained there was too much scrolling. So I split it into 5 reasonably simple pages. After you have completed the 5 steps and run the program, you can come back and alter any one of the questions without going through the whole lot.

So here are the 5 steps to use the program. These steps are also listed in a quick-start guide at the top of the menu.

1. Choose a lathe from the list (eg metric or imperial without a gearbox or choose a lathe with a gearbox).

2. Change the lead-screw pitch if necessary. 

Enter your own list of change gears, ie number of teeth on each gear e.g. 20,30,35,40,45,55,60,63,70,80.

Decide how many compound gear pairs you can fit on your quadrant/banjo. Usually one is enough. Two is more complicated.

Then there is a question about whether you want 'Automatic' pairing of compound gears. Some older lathes like my 1955 Boxford A, have a set of change gears and a separate set of compound gears and they require different shaft sizes so they cannot be mixed easily. However, most lathes, like the Myford, do allow you to make compound pairs by mixing and matching change gears. In those cases you choose 'Automatic' and the program calculates all the combinations, and you can leave the space for separate compound gears completely blank.

3. Choose whether you want a metric thread, imperial thread, feed rate or worm gear thread calculated.

4. Choose the percentage error you are willing to accept eg 0.5%

5. The program obviously wants to know what pitch or TPI you require. There are 3 ways of doing this, and you only need ONE of them.  The simplest is 5A, to just enter the number. Then you can skip 5B and 5C

5A. Enter the pitch or TPI of the thread

5B Allows you to produce a table of results for a range of pitches or TPI.

5C Provides drop-down lists of standard threads eg BSW, UNF ISO-metric, BA etc so that you can look up the pitch or TPI.  Currently this does not automatically enter the pitch. You would have to copy it into 5A.  I am working on a new version to fix this.

6. Finally you can RUN the program.

The program goes through every permutation and combination of gears calculating the pitch or TPI for thousands of combinations, sometimes millions. If it finds any that are perfect (0% error) or within the range of error you specified), it displays the set of gears that will produce your thread. The results can range from 0 to thousands of possible solutions, If there are too many you can go back to the main menu and reduce the amount of error you are willing to accept, possibly even entering zero. The results are sorted so that the ones with the least error are listed first.  You can experiment, e.g. try adding or removing gears from your set of change gears.

When you choose the gear set you like, click the DRAW button and it will produce a scale drawing showing you how to assemble the gear train. This also helps to show you potential clashes between gears, and the program tries to warn you about possible clashes, using *** and ### beside the best solutions.

If you chose 2 compound gears instead of one, there are many more possible solutions, because there are 6 gears in the chain, so the first page of results shows a list of 3 driver gears and 3 driven gears. Click the SOLVE button beside your choice and it shows 12 to 36 ways of putting these together, all producing the same gear ratio and thread. Each solution has its own DRAW button. Again you can click DRAW to see a scale drawing of the gear train.

So, this may look daunting at first but reasonably easy once you use it. So see what you think and let me know!

Happy turning and all the best from

Evan Lewis in New Zealand

1.2 million page loads can't be wrog!

 

Edited By Evan Lewis on

[Evan LewisParticipant @evanlewis60038]

I just wrote a very long post and clicked back to an earlier post so that I could refer to it, and my message was deleted.

Howard Lewis,

My namesake! My family came from Basingstoke / Southampton area. They moved to NZ in 1875, but flitted back and forth under sail (3 months) and later steam. We also flitted back to England (about 24-48 hours) when I worked on medical research at the Department of Mechanical Engineering in the Stephenson Building at the University of Newcastle upon Tyne. I failed to get into engineering school in New Zealand because I didn't get a high enough grade in mathematics, but I got my way in the end! Doing mathematical simulation of haemodialysis patients in 1978.

Duncan,

So you are the famous Duncan who beat me to the online program idea! I didn't find your work until after mine was already functioning. There is a lot to be said about a minimalistic approach. It looks so much easier to use, but not quite as flexible. I knew about the nthread.exe programs from lathes.co.nz, but I am an Apple Mac user and these .exe programs will not run without Windows… taking us back to Nigel's point. The downside of "cloud based computing" on servers is that it requires internet access and some have complained that they do not have internet access in their workshops.

Evan Lewis

[CalumParticipant @calumgalleitch87969]

Posted by Calum Galleitch on 10/07/2021 14:00:26:

John, I quite take your point about the output requiring some intelligent interpretation. There's a balancing act between making such a tool capable of handling every case correctly, and it becoming too difficult to use for anyone but an expert – who probably doesn't need such a tool to begin with!

[Nigel Graham 2Participant @nigelgraham2]

Evan –

Re your last paragraph.

Those you say they can't use an Internet-based programme in their workshops are rather missing the point. No-one needs do that. Perform the calculations in the house then take the results on paper to the workshop.

Even better, use the programme just once one evening to create a printed table for all the threads you are likely to want, then laminate it or wrap it in a polythene bag to keep next to the lathe.

Same with CAD: I work from the paper print, not the screen image. My PC and printers are nowhere near the workshop.

[Evan LewisParticipant @evanlewis60038]

Hi NIGEL,

You mentioned that you have an L5 lathe. I assume that is a Harrison lathe. If it has a gearbox and you would like me to add your lathe to the program I would be happy to do so. What I need is:

A photo of the thread cutting table, usually shown on the gearbox.

The pitch of the leadscrew, which I believe is 4TPI.

The standard gear train setup used with the gearbox label. Number of teeth on the spindle and lead-screw gear.

Fudge factor if you need it.

These are the basics for the gearbox, but we can add

Add a list of change gears in your set,

Gear ratio for longitudinal power feed rates, eg feed per turn of the leadscrew

Same for power cross-feed. ie feed per turn of the leadscrew.

[Evan LewisParticipant @evanlewis60038]

Hi Nigel,

re Holzappfel:

I found the books by our Mr Wooden Apple on Mr Googles collection and you can download it as a PDF ! So I have been reading all afternoon. Here is the link to volume 2 about old methods for cutting threads: The blue link below will save you from copying and pasting the full URL:

https://books.googleusercontent.com/books/content?req=AKW5QaeFhanvnTl_D4nQ0btJKs_ONycVFdRvJg40qlMcvrKc2MX3hn1aSMP6BPzq7usdAD5jpAHgmgFfvKA-O-Awh7NdCVfjGlAXjMOMuu5PWhUpU0ZV2y9fRR3q-wSDL42dLVeFJRwBYJT6nLeDRjVhNkNFWZtxOR6Z9pYbODBHOifwvH7NyeTTQJ4Y-5Re51sdEQrrATu7wgUHmgs5Cb-t1XiS3m95mtVYxwXi0ZNpwZbxCkuovd3e-VhW1P1nHrz_n0bJfM4XaEQPDCab7DbxPBc_Lwn3ZQ

Holzapffell

The relevant part is about p577 to p655. The system you were referring to is in image 602.

As the spindle or mandrel in their example is rotated it turns a small pinion gear which drives a rack placed at right angles to the axis of the spindle (as you described Nigel). But I couldn't see how this rack would move the tool along the axis of the spindle.

The answer is that the rack is connected to an adjustable wedge. A bar sliding along this wedge, and positioned at right angles to the wedge and pinion, moves the tool along the axis of the spindle.

In their example it takes 10 turns of the spindle to move this wedge along its full length. If the thickness of the wedge increases by one inch along its length, then the tool moves one inch with 10 revolutions of the spindle, thus producing a perfect 10 threads per inch, eg for a clock mechanism.

Since the taper is adjustable it could be set to say half an inch to produce 20 threads per inch or anywhere 10 to infinity threads per inch. The catch is that when you get to the end of the wedge the tool cannot go any further and it is limited to 10 turns of the thread. ie 10 turns results in the full length of travel of the pinion and its wedge. Probably good enough for most watch and clock mechanisms. I copied the drawing of this device and will try to insert it later.

[Author]

Posts