Screw cutting is over complicated

I've watched a few YouTube machining videos of cutting threads on the lathe with a single point form tool. Many of the YouTubers make a point of setting over the top slide by half the included angle for the thread to be cut, then go through a complicated dial setting exercise to cut the thread. Cosine error is often mentioned too!

This got me thinking! In all the years I was turning in machine shops, I made hundreds of threads using single point HSS tools, ground to the correct included angle and radius.

I never once set over the top slide.

The method is: Touch off the tool on the diameter. Zero the cross slide and top slide dials. Move the cross slide (for example) 0.006" for the first depth of cut. Second cut is 0.009" on the cross slide, but move the top slide approximately a third of the depth of cut (0.002". Take a cut. And repeat until nearly down to depth. The final couple of cuts are taken without moving the top slide, this will give a good finish on both flanks of the thread. This will ensure that the cutting edge of the tool is always advancing and cutting, with minimum rubbing on the trailing edge.

Nice and easy method, still do it this way, and it allows the true thread depth to be cut. I never had any problems doing it this way. Can't see me doing it the YouTubers way any time soon, far to complicated!

Ian

I've been banging this drum for years, but I put half the cut on the top slide, which generates an included angle of 53 degrees, so you get a light scrape down the right hand flank

Well here we go again, this is an old chestnut which has often been debated without a definitive answer, I suggest try both methods and see what works best for you. I think amateur machinery may begin to struggle with larger pitch threads using the plunge method. As long as you are aware of both methods then you have an option if things are not going well, three tons of DSG is a different animal to 150lbs of Myford.

Mike

I'm sure I've got some notes somewhere with this method, I'll dig them out, post them up and give a definitive answer. Never got taught the set over method as an apprentice, and I don't know where the YouTuber method came from, I can't see the point in doing it that way.

And you must be a psychic Mike I did work on (amongst other machines) a DSG, it was usually a 48"x6" centre lathe, with all the gizmos you could wish for on a manual machine, the Rolls Royce of the machine world!

Ian

I was taught the same/similar way as an apprentice in the toolroom. But advancing the topslide by half the amount the cross slide was moved in rather than a third.

There are several advantages to this method, over "the Yankee method" of swivelling the topslide at an angle.

One is that you set your topslide to turn dead parallel and leave it there, so it can reliably be used for turning short lengths up to shoulders etc. thus saving time on small and/or fiddly jobs.

The other is the you can manipulate the top slide at the end of the screwcutting process to take a cut down one flank of the thread at a time to clean it up, or to create a little extra clearance if needed, without exceeding the correct thread thread depth. This becomes more critical on larger threads where you are most likely fitting the male thread to an existing female thread, nut, etc.

However, the "Yankee method" of swivelling the topslide does have the advantage of keeping the topslide clear of the tailstock, more of an issue on smaller lathes such as the Myford etc. The larger lathes eg DSG etc don't seem to have this problem so much because the larger toolbits they use have more overhang and better designed layout of topslide.

ISTR there is a standard number that you mulitply the thread pitch by to arrive at correct thread depth when using the topslide swivelled to 29 degrees a la "Yankee method".  So it's no more complicated than the non-swivel method. Just set the cross slide to zero and use that to retract tool at the end of cut and then reset it to zero and then add the cut depth with the topslide. I have that number written in my old thread chart book but dont remember offhand what it is.

I use both methods, depending on what I am doing, which lathe I am on, the way the wind is blowing etc. Both seem to work ok. But I do prefer the non-swivel option, providing tailstock clearance is not an issue (which it can be on the Myford etc at home.)

Re "YouTuber's method" there seems to be a great variety in quality of information on machining on YouTube. Much of it is good. But much of it is any old self-taught beginner posting up their new-found expertise. (Not saying that is the case in this particular instance, note.) The Dunning-Kruger effect can run rampant. To wit: one example on servicing a lathe chuck shows the two halves being separated by belting the chuck key with a shifting spanner! All the while ignoring the three threaded jacking screw holes especially provided for this purpose.

Much more reliable to spend the little bit of money it takes to get a couple of the classic books on using the lathe by past masters such as LH Sparey (my personal favourite), Martin Cleeve, Duplex, Ian Bradley, GH Thomas etc. Many of these were guys who had many years of day-in day-out professional machining experience behind them.

Edited By Hopper on 16/09/2019 05:31:14

Good to see you back on the forum Hopper, hope you have had a nice break rather than been off sick.

Mike

I really like the Hardinge top slide method. As it has a cam that retracts the tool on the flank angle.

The way I set it, is set the compound at 1/2 the thread angle. Have the lever in the forward position, and the compound wound to the forward stop position. Wind the cross slid to a touch, and then go in the full finished depth. Leave the cross slide at this position. Use the retract at the end of each cut. Wind back on the compound. Take all cuts using the compound slide only.

Without the retract lever, I do the same on my Myford, and do all the feeding to zero on the compound. For most threads, 1 turn on the compound is the max required amount of retraction required.

I have done the feed only on the cross slide method as well. But prefer the feed on the compound. Unlike the youtubers, my method does not require and trig to get the depth of cut. That is set from the cross slide , the same as if you were not incrementing the compound slide.

Neil

I think my pinup girl Ruby did it Ian's way:-

Ruby Loftus

She's left her compound parallel, but she's getting a single curl of dull-red swarf coming off, so she's only cutting one side of the tool.

In my view the most important thing about screwcutting methods for normal use is that it should be possible to easily figure out what went wrong in the event of the thread not coming out right. This is especially true for methods to be used by neophytes and even more specially so for those learning on their own with no mentor to point out errors of understanding, implementation or practice.

From this viewpoint the only satisfactory method is the one often called "Zero-to-Zero" or "Zero-2-Zero" which I have several time previously described on the forum. Fundamentally it's a variant of the angular indeed method advocated by Duncan refined by Neils feeding to zero on the compound approach.

Basically the topside is set at an offset angle close to and slightly smaller than the thread angle and the threading tool set perpendicular to the work after it has been bought to the correct diameter. The cross slide is fed forward until the tool touches the workpiece and both dials set to zero. Then move the saddle so the tool is clear of the workpiece and feed the cross slide forwards by the desired thread depth as specified in your data book. Reset the cross slide dial to zero and pull back the topside to clear the work. Cuts are put on via the top slide and threading passes made with the cross slide at zero. Use the cross slide to pull back the tool for clearance on the return pass. Cross slide manipulation is a lot easier if you have a stop system, whether universal of simple threading one as per SouthBend and Boxford. Final pass is when both dials read zero.

The major advantages are :-

1) The lathe does all the infeed calculations for you so the top slide offset angle merely needs to be close, not exact. My top slides live at 25° off perpendicular to the bed which not only works fine for both 55° and 60° threads but also keeps the hand wheels out off each others way and the slide itself clear of the tailstock.

2) All there set-up is done to data book values before starting the job so when finished you have cut exactly what you set-up to do. So if things don't fit you have a solid base from to start working out what went wrong. Frequently with home ground tools the tip radius doesn't correspond to that assumed by the data books, usually too sharp so the thread is too shallow. In that case its easy enough to make extra passes with bit more in feed on the cross slide until a good fit is achieved. If you have a bunch to do re-setting the cross slide dial zero once the first one is right means all the rest will be right. There maybe some merit in deliberately making the tool a touch too sharp knowingly creating an overtight thread before finally correcting it with a touch more infeed. Especially if you need a close fit. No different to using partial profile carbide threading inserts really.

All other methods involve either some in-job manipulation or very careful set-up to ensure the calculations correspond to the reality. The basic angular in feed method needs the feed calculated for the angle at which the topside is set. A value not generally available in data tables.

In my view the method originally advocated by Ian is an invention of Beezlebub accurately aimed at generating vast quantities of shop esperanto. All that futzing about with the top slide set parallel to the bed becomes acceptable only when shaving an ACME or similar feed screw thread to zero, or at least close to zero, backlash. Its still a PIA tho'.

Clive

If you weren't taught all methods you should claim back from your school for incompetent instructors and also learn to read a book or two.

As Mike said it's an old chestnut topic and belongs in the bin with the 'Why do people take ten thou cuts because when I was in industry I used half inch depth of cut (on my 50 HP lathe)'.

You don't need trig as it is done for you in done for you in Machinery's Handbook and on the web but anyway just keep taking cuts until the test nut fits.

Good to see her again!

She is, of course, cutting an internal thread where I imagine setting round is less common.

I have just noticed that, as the supposed star of the shop, she is wearing a green head scarf but her mates are all in red.

Better not let Joe Pieczynski (The Texan Spiderman) hear you talk about "thread depth" - you'll get a smack on the wrist!...

https://www.youtube.com/watch?v=IqJWoAM5eRQ&t=888s

IanT

Yes, but it's a shortish thread in a wide bore, so half-angle could still be practical if desired.

I hadn't noticed the green snood. When I was working in machine shops in the 70s, I had long hair and had to wear a blue one - but there was no grade distinction I was aware of...

Edited By Mick B1 on 16/09/2019 17:13:19

Joe is of course right in what he says.

Tony

I used the Ian Johnson method recently for a different reason. I had a leadscrew for a Tom Senior X axis, which I added about 4" of extra thread on one end. This meant that the uneven wear on the leadscrew was further compounded by some brand new thread. With the compound at zero (parallel), I used the compound to creep to one side of the most worn part of the thread until the tool started rubbing. Then I carried on over the less worn parts in several stages, and re adjusting the compound to cut the other side of the thread. Eventually I ended up with a thread with all the wear equalised. The wear was not so bad as to be noticeable visually, but was quite obvious when using two nuts to minimise backlash. Now, the backlash is minimal, with no tight or slack areas.The thread is 3/4" X 5 ACME, I had to use a travelling steady and a live centre in the tailstock, plus bored soft jaws.

I usually consult thread charts for thread depth, which give an idea of how far to go and also helps to plan the number and depths of the cutting stages. The last threads I cut were of no particular size, the M and F had to be a nice fit, and I made the thread depth to suit the pitch using charts.

Edited By old mart on 16/09/2019 18:59:48

Right! Had a rummage around today when I should have been doing something else, and found this 'Newnes Engineers Pocket Book' its the posh 'Metricated' version. I bought it around 1980 ish.

Here are a couple of pages describing the method I and many others use. Annoyingly the text is spread over two pages. There are some very good methods in these old engineering books which are still very relevant today.

I still think there is no need to over complicate things, keep it simple, if it's good enough for Ruby it's good enough for me!

Ian

It is interesting that nobody in the workshop with Ruby is wearing safety glasses. I suppose that was quite common years ago but it did make me shudder.

Doug

First thread I ever screwcut was a 16tpi square thread and I have not bothered to swing my topslide for any others since.

This method would also save all those beginners making teh mistake of setting the angle at 29.5 deg from the wrong axis

Good comment. They were obviously concerned - at least nominally - about hair, but not eyes. In the turning shop with surface speeds around 100 ft./min. the risks were probably not that great, but amongst the bench grinders in the background... ?

Perhaps all the perspex was reserved for cockpit canopies and tank vision blocks. British industry didn't really get safety conscious until much later - a lot of casually-unsafe practices were still common enough in the 70s, and we used to joke that the incoming H&S rules seemed to be meant to ensure that no sharp tool could ever actually cut anything...

Agreed! (With Ian, for clarity).

And set the dial to zero at the right DOC on the first run, so on further threads you can just do in two, across one until it reads zero again.

Neil

Edited By Neil Wyatt on 16/09/2019 21:13:42