[Hezz]

Hi, I don't think high quality CNC commercial lathes of an industrial nature were the topic of the thread even if they do use the methods for CNC screw cutting you mentioned..

. I quite understand the process required with CNC turning and screw cutting.......not the using of it though, and the "leadscrew", that pushes the saddle along the bed, which for any CNC work would need to be a ball screw, and as it has a pitch is still a leadscrew even if it has balls instead of a solid nut with threads.

BTW......a leadscrew on a lathe is called that because it generates the pitch required to be cut as a division of it's own pitch when connected to the headstock spindle with a gear train.

What I proposed, and wanted to find out the problems arising, is to have a CNC lathe and direct couple the leadscrew and spindle with gears instead of an encoder etc .

You would then not have to change anything virtually in the spindle drive train, apart from converting the rest of the lathe from Acme threads to ball screws for the saddle and crosslide as would definitely be required to do CNC turning...etc etc.

This does not include having linear rails on the bed or crosslide, but can be a retrofit as with the removal of a compound rest there will be bags of room to fit linears as required.

It's not conventional in the CNC sense, but simple......where are the problems I asked.
Ian.

OK, I'm sorry, I kind of misunderstood what you were getting at. But I guess I can't see the point in doing what you are thinking about. In a way, where threading is concerned, you kind of have a mechanical "CNC" connection while doing threading on a manual lathe, which essentially allows you to interpolate in more than one axis at the same time. This is really the point of CNC. Motion in more than one axis at nearly the same instant. The value of CNC is it's flexibility and repeatibility and enhanced accuracy and more then two simultaneous axis at the same time. This gets hard to do with mechaincal assemblies and when they get complicated they have a lot of backlash.

There are some kinds of lathes that are CNC but work more like a manual lathe. I have worked on them but I still believe they only synched the spindle and leadscrew/ballscrew electronically. It would probably be hard to build a ballscrew with as fine a pitch as a leadscrew and still be really strong. So there is always going to have to be some kind of electronic speed interpolation to get the right thread pitch relative to the spindle speed.

[handlewanker]

Hi, I quite agree with what you say.......CNC for CNC use, and if the coupling can be reliable without too much complication, that would be the ideal world.

A balls screw used as a lead screw for screw cutting would have a pitch of probably 5mm or 10mm,...... so doing a pitch of .75mm with a 5mm pitch ball screw would need a ball screw drive reduction of 6.66666 recurring etc........many manual lathes have 8 TPI lead screws or 3.175mm pitch...........mine has 6TPI.

I don't think many people with a CNC set-up would want to change pitch by moving levers or changing gears to cut a pitch, whereas with a program input the pitch is automatically generated.

The change gear method is something I can relate to mechanically, so I thought it might be a simple solution to get screw cutting with a CNC set-up without tears.

I'll be watching your build to see how it works out.......might go down that path too.
Ian.

[hanermo]

Does it ?
As you say .. for many years.
I believe there was a problem with sw based encoders being limited to very slow rates.

This was back when people were using 500 count encoders, to get around some limitations.
Today, basic encoders on basic servo drives at 10.000 counts, and with gearing, 512 kHz is the needed rate.

Please .. if you think something is useful, and works, and you know about it, provide data or links.
Have you in fact used encoders based threading, and with what equipment ?

Linuxcnc has supported encoder/index based threading for many years and it works fine

[hanermo]

All my stuff re_ threading is based on my trying to make good-quality threads for a long time, with not-acceptable results.
I am simply warning about the pitfalls, and about what did not work, and in many cases wont work.

Blight -
I disagree very much, I´m afraid.
Hobby use does not need to provide for speed, power, rigidity, or productivity.

*But*, where accuracy is concerned, accuracy is critical. Accuracy is accuracy.
Otherwise, assys wont mate, and fits cannot be built, and there will be wobble or binding.

In many cases, lesser results are ok, of course.
BUT, to be able to make a judgement on such, the first thing one must know, is how far "off" the results are, at least to some extent.
Or, to be able to confirm that the results are "good enough".

As such, some method, however basic, of measuring threads, or comparing them, is needed.
Examples:
A pair of go/no go gages for small threads, is not expensive, especially second hand.
A high-quality nut, made to 3A/B standards, is another example.
Neither fulfills "formal" quality standards - but formal quality standards are not needed for hobby use, indeed they are mostly not needed for industrial use.

The biggest mistake newbies can make is pretending they are building stuff to high accuracy (when its usually resolution), and not measuring/testing anything.
The second biggest is trusting gear to work, when most does not, in some areas, for many reasons.

I used an encoder as a good solution for getting an index pulse.
The bad part is when people spend 30-50$ on something like an index sensor, for threading, that cannot possibly work properly.
Instead of spending the 40$ on an encoder, that *will* work at all speeds, and give a very good, clean, index pulse at all times.

ANY measurement or testing is useful.
The vast majority of machine tool testing can be done with instruments costing less than 100$, quite acceptable for anyone trying to build cnc stuff.
Examples:
For those of us with an oscilloscope, measuring the index pulses would/could show them being ragged or fuzzy. (Not me).
Some electronics-oriented people could build triggers to clean up the signal (not me), to an extent.

Me, I will solve it by using a simple, easy, cheap, industrial encoder that has a crisp pulse by default.
Actually, since I use servos, the pulse already exists on the Z/ channel, ie encoder home.
=>
The next problem then is that almost no motion-control stuff works well enough to actually use the index pulse (its very short).

Examples:
CSMIO-IPS and CSMIO-IPA work with both mach3 and mach4.
Hicon has been demonstrated by Brian to work with threading.

I believe the Pokeys will soon have threading - and the hardware is good enough (125 kHz) to read an index pulse.
(The sw exists, not released, I believe).

Some have plugin and driver problems - the SS/ESS is one to some extent.
Many of us, me included, have stopped using them.

Standard IO boards wont work at all. You do need a real motion-control board.

I think practically none of the USB stuff works, in general, for lathe or threading.

Hobby use doesn't have to comply to the same standards as industrial use.

[handlewanker]

Hi....BTDT........,many years ago I used tapped out nuts and "good" screws to gauge thread fits as it was all that was required then.

Being the nit picker that I was/am, I started using the three wire method that I found a formula for in the Machinery's Handbook and never had threads rejected for bad fits etc.

In those days we hand ground our screw cutting tools and I still have a set of hand made Carbide tipped screw cutting form tools for 55 deg Imperial and 60 deg metric external threads both left and right hand......the internal threading tools are all HSS.

The more I see a CNC screw cutting topic discussed the more I realise it ain't gonna happen for a lot of people....me included.

That would make screw cutting a non starter for a retrofit CNC'd manual lathe if the software was not able to handle the screwing part of it......boo hoo.....I wanna join the party but I ain't got a party dress....LOL.

Unless a simple interface is on the cards, it will be gears for me....I like to keep it simple.......it worked in the 50's so it'll work in the 21 st century too, alongside the digital machining evolution of CNC.
Ian.

[daniellyall]

you can do screw cutting on any machine but what the hardware and software are, makes a big differences. you can do it with M3 depending on what you are running CSMIO-IPS and CSMIO-IPA and Hicon make it easy to do as long as the signals are clean and you have a good encoder (plenty of evidences on youtube look up hood, Brian baker). you can do it with lesser boards but you will have to go slow as a wet week and if you use a single dent well if you're happy with +/-1mm well ok no problem. but a non cnc lathe can be better, if the person knows what they are doing.

anything can be done if you have the money and time and you know what you are doing. this is one of those around and around you go argument`s

[handlewanker]

Yes, I realise it's a program orientated thing with the spindle encoder generating enough pulses to "encourage" a stepper motor on the leadscrew to turn at the ratio for the pitch required.

How many pulses the stepper motor gets is dependent on the closeness of the lines on the encoder disc.and how much the stepper motor is geared down to the leadscrew.

If you understand how an encoder disc works together with the software and stepper motor, then it's easy peasy.......I just don't......hence the suggestion of the geared interface for the pitch variation.....I can understand that.

As I see it, if the spindle puts out a fixed 200 pulses, then the stepper motor on the leadscrew turns just once........pitch would be a geared combination as required.

When it comers to generating the pitch by varying the ratio of output pulses from the spindle.....that's where it gets tricky (for me), and you'll need more than 200 pulses to get a fine pitch from a 5mm pitch leadscrew or ball screw.

I'll leave it there, as the thread is about building a tubular steel slant bed lathe etc with screw cutting being achieved by total CNC means..
Ian.

[The Blight]

My statement: "Hobby use doesn't have to comply to the same standards as industrial use."

Your statement: "I disagree very much, I´m afraid."
And later: "Neither fulfills "formal" quality standards - but formal quality standards are not needed for hobby use, indeed they are mostly not needed for industrial use."

So it doesn't have to comply to industrial standards, but it helps? Then how is my statement wrong? Hobby machines don't have to comply to industrial standards. The cost of developing a robust system that can machine precise threads can be a costly affair. The cost alone isn't just in the tools needed to measure the threads, but also the time invested in finding out where the machine lacks work to have some precision. Perhaps you need a higher quality ballscrew, or a better encoder, or an entirely new controller as well. The cost benefit analysis might not fit the average hobby user, and so the trouble of machining a few threads far outweigh the cost of investment (time and money) at the start of the project.

Back up your claims with real data. Show us the methodology of your studies, and the data you have collected instead. Just because you haven't gotten it to work properly doesn't mean it can't work for someone else. If you post the data, someone else might find a hole in it and can find a way to work around it. Right now I'm only seeing someone post to a forum saying that A is better than B. Someone else might come in and say that the opposite is true and you would never know unless data has been posted to back up the claims.

For some people the criteria for success might be as follows: Does part A fit part B, if yes = Success! If no = make more room. Accuracy is a broad statement as well, and should be narrowed down to a set of criteria that matches the users needs. Basing this on industrial standards is a good thing, but might not be the best end goal. Cost vs benefit should be considered.

It would be fun to see someone build a machine that can machine high quality threads and only use it to turn candle holders.

I would love to see a topic on the subject where every part of the machine has been analyzed. Please start a topic on the subject and fill it with your data and experiences. The OP of this topic has already stated what he is trying to do, and so I won't derail the topic further than this.

--------------------------------------------

Siam: Have you though about filling the tubes with sand, or lead shot to control vibrations? Once the machine is in place you can fill the frame, and if you need to move it you can just take it out again.

[siam]

Siam: Have you thought about filling the tubes with sand, or lead shot to control vibrations? Once the machine is in place you can fill the frame, and if you need to move it you can just take it out again.

I have considered filling the the tubes with sand, but I must admit, I hadn't thought of using lead shot,...though it should be a good vibration dampener. I've also considered using a two part foam or even oil. However, although it's always good to plan for the worst, at this early stage in the build, I'm not going to spend too much time developing contingency plans for a problem which may never develop. At this point in time, these ideas will simply be noted in my list of "what ifs".

[m_c]

I think practically none of the USB stuff works, in general, for lathe or threading.

Dynomotion's KFlop is USB and manages threading perfectly fine with full encoder support. It can even handle rigid tapping.

[handlewanker]

I have considered filling the the tubes with sand, but I must admit, I hadn't thought of using lead shot,...though it should be a good vibration dampener. I've also considered using a two part foam or even oil. However, although it's always good to plan for the worst, at this early stage in the build, I'm not going to spend too much time developing contingency plans for a problem which may never develop. At this point in time, these ideas will simply be noted in my list of "what ifs".

Hi, I think the amount of damping you'd need or get with those long rails for a bed would not make a hapoth of difference to any turning that happens.

The bed might flex or bow a bit if it's flimsy and has to resist drilling forces, mainly on long pieces of work where the drill point is halfway down the bed..... so design to cater for that need......that is, heavier wall bed section tubing......do you really need a long bed lathe?

Most of the time the saddle is near the chuck doing short jobs without tailstock support, and in that mode the stiffness right near the chuck end would be at it's maximum, so no need to over engineer for a condition that isn't present.

If the bed, frame whatever, is felt to be flimsy and needs damping, go to a thicker tube wall.....it's the simplest way without tears.

You'll get the most vibration or deflection from having a high centre distance (swing), in an attempt to cover all needs for work diam, whereas you may never turn large diams..........if this is important, include a gap in the bed near the chuck, or the mass of the headstock and bed would need to be increased a few fold.

Manual centre lathes ( some, not all) have removable gaps in the bed right where the chuck rotates to cater for large diams etc.

My two bobs worth would be to design for a conventional work diam/length and live with the fact that a lathe, especially a CNC one, for all seasons is impractical.
Ian.

[Tkamsker]

Siam have a look of how thread and linuxcnc it is quite easy i use Amt 102 encoder. . Read through it and decide yourself and i also use steppers -

Gesendet von meinem SM-G800F mit Tapatalk

[Eldon_Joh]

So I recently found out that the modulus of elasticity of steel being 30 million psi.. is actually more than twice that of cast iron. somewhere long ago I read that they were the same.. nope. anyhow: So you should be able to take a steel weldment, and take an equal weight of something with a very high dampening coefficient and get something with a much higher dampening coefficient than cast iron. Now I don't suggest you buy 500 pounds of silicone when there are other solutions but the point is clear.

So I've been building an atlas style lathe. I have the bottom three layers of the attached image welded together, had I started with clean stock* and welded it properly I could have ended up with 2mm warp, instead it was quite a bit more but I was able to fix some of it. the fourth, top layer will be separately constructed and welded very carefully.. i may preheat all of it to 500F before welding it to the bottom layer, and after welding take it up to 1200F without letting it cool down.

anyhow I've not decided on 1 inch by 3 inch ways, or .75 by 2.5 inch ways. they will bolt to the weldment (after the mounting surfaces are ground within a reasonable tolerance) and i haven't decided what steel to use or how to process them, or heat treat them.

currently the cross section is designed to look like this:
Attachment 292364


Anyhow the reason I'm posting this here is because I'm concerned about the stiffness of the frame in OP's proposed lathe.
mine is on the order of 200in^4th in both axis. or maybe it was 150, i can't recall. anyhow the channel iron is 13.5 pounds per foot and its relatively easy to calculate.

mine will have 72 inch long ways and the lathe bed is 92 inches long. lathe carriage will need to be at least 18 inches long, so 54 inches of travel.

I will have to add about 100 pounds of steel to the headstock end of the lathe bed to provide surface area for the headstock to bolt to. But even after that.. the bed will "only" weigh about 450 pounds. (not including the ways, carriage, tailstock, and head stock, these items may indeed double the weight)

So i'm wondering, should i fill it full of sand, or add more steel?

*i am building it from channel iron scrap.. ~24, ~23 and ~19 inch long pieces. so the length, 92 inches, was a compromise because i could make all the beams fit to within an inch.


anyhow, OP i suspect should add two thick steel plates to make a solid triangular box section under the lathe bed, or fill the space with cement. it looks fairly lightweight to me,

[handlewanker]

Hi, any welding will produce heat, and steel expands with heat........if you subsequently weld another section to a previous section that is hot, when it cools down you'll get contraction and that is where the problem lies when the two sections try to pull each other apart.

When welding a structure, I use a cold welding method........that is, keep the section you're welding as cold or as cold as the ambient as you can, by welding in short stretches and then cooling the part down, even by quenching it in water.......if it isn't hot, it isn't bigger or longer than when it was cold.

Welding a collection of metal parts where some of them are hot or are going to get hotter than the rest is making a jigsaw of pieces under stress..... as you fix more expanded parts together, those with the most mass will pull the hardest as they cool down and contract.......that is distortion.

Steel tubing, by nature of the unconnected 4 sides will bow all over the place if it's allowed to get hot during welding.......it will get hot, but it must be cool when the next bit is added to it.

Interpret that as you see fit, but expansion and contraction is not an interpretation of physics, it's a fact of physics, so knowing the beast allows you to overcome it's idiosyncrasies.

BTW, if you speak for yourself, no doubt a complete structure can be completely assembled and welded in one hit and then annealed at a temperature that allows the stresses to dissipate, but I doubt the average or above average constructor would go down that path and still have money in the bank.
Ian.

[Eldon_Joh]

handlewanker i should have been more clear, to minimize warp i may decide to heat the entire 450 pounds of metal to 500F before welding, then take it up to 1200F immediately without letting it cool down.
yes i have considered welding it cold in short sections, but i'll have to heat it up to 1200F anyways, so what's the point. i can heat treat it with 4$ worth of electricity and I might destroy 20$ worth of fiberglass insulation. half of it by weight (and perhaps moment of inertia) has already been heated to 865F which should have reduced any residual stressed by 10-20% but that's about it.

that said, with regard to steel tubing in OP's build: if he wants to have it all tack welded to within 1/32" that's fine. it won't be within those tolerances after its welded regardless how many welders were welding on it at the same time: it simply doesn't have torsional stiffness.

[handlewanker]

Hi........all at the same time??.......that would make it glow with the heat build up which is contrary to what I said about the amount of expansion that would occur......hoping it would simultaneously contract at the same rate when cold means anyone working with structures and doing that hasn't understood the reason they warp and distort.

You could tack weld a structure together completely to get it to a close dimensional final shape, but progressive and continuous welding with a subsequent heat build up will play havoc with your final structural shape etc etc

It goes without saying that if the assembly is kept as cool as it was when it was tacked together, it will not distort.....the emphasis being on "as cool" etc,...... hence the need to cool it down progressively and keep it cool to maintain the structural shape, and that means short welds on alternate sides as you progress......time consuming but necessary to control heat build up.

I think I've said enough on the weld side, as no doubt everyone has their own point of view, and mine is just one of them.

BTW, I use a plasma welder, and that at 8,000 deg C is like Tig welding on steroids...... welding with heat like that is quick but is a factor I'm very much aware of.
Ian.

[Eldon_Joh]

Hi........all at the same time??.......that would make it glow with the heat build up which is contrary to what I said about the amount of expansion that would occur......hoping it would simultaneously contract at the same rate when cold means anyone working with structures and doing that hasn't understood the reason they warp and distort.

i can't give you a percentage of how much less distortion you'll get if you heat the metal evenly to 500F then weld, then heat to 1200 F after (not letting it cool down) but it will be a lot less than welding it cold.
but i can tell you 500F is the minimum pre-weld heat treat i've ever found.. for speciality metals.*

When you weld a metal cold, if you weld it such that you lock in the stress, and you do so evenly and properly, yes it is true it wont move on you. but it still has locked in stress.
But it is difficult to do this. And you can't touch it afterwards or it will move on you. you can't even drop it on the floor or hammer on it... it will move. the highest stress points will yeild and it will move.

I heated the 260 pound weldment i had to 860F thinking 800F is enough. it isn't. you're supposed to heat mild steel to 1200F, then let it soak at 800F for 4 hours before letting it cool to 400F. then after its at 400F you can douse it with water (you cannot cool it rapidly until its less than 500F)

*it is probably true that mild or non speciatly metals don't care. but the authors of those sources are concerned about post weld strength, ductality, etc. they are not concerned about stress relief on the order of 1%.. which would be enough to move a lathe bed beyond reasonable tolerance.

if you were to heat mild base metal to 1200F and then weld with the usual stick welder.. you would probably have almost no distortion at all.
but no one wants to weld in hell lol...

[siam]

to minimize warp i may decide to heat the entire 450 pounds of metal to 500F before welding, then take it up to 1200F

If you plan to heat the entire structure to 1200F, have you considered "wick" brazing it; you pre-flux all the appropriate surfaces, strategically place brazing rods or wire such when they melt, the liquid metal flows and wicks into all the joints? I believe there are brazing alloys that melt at 1200F. The result would be a structure that is essentially one solid piece. Should be much stronger than welding, and you wouldn't have to sit in front of 450 lbs of metal which has been pre-heated to 500F.

[Eldon_Joh]

If you plan to heat the entire structure to 1200F, have you considered "wick" brazing it; you pre-flux all the appropriate surfaces, strategically place brazing rods or wire such when they melt, the liquid metal flows and wicks into all the joints? I believe there are brazing alloys that melt at 1200F. The result would be a structure that is essentially one solid piece. Should be much stronger than welding, and you wouldn't have to sit in front of 450 lbs of metal which has been pre-heated to 500F.

you're at another level of workmanship here to suggest such a thing, no offence.

welding on 500F steel is of no concern, not when half of its wrapped up in insulation and you only have a few square feet exposed. if it were 1000F which is a black red heat in daylight*, yeah that's warm. but 500F is nothing.

(*and you have to supply a few dollars an hour to keep it that warm)


1200F is minimum required to anneal a weldment as far as i can tell, but it really depends on what you want to do with it.

[siam]

with regard to steel tubing in OP's build: if he wants to have it all tack welded to within 1/32" that's fine. it won't be within those tolerances after its welded regardless how many welders were welding on it at the same time: it simply doesn't have torsional stiffness.

Let me state straight away that I don't weld; I have never learned that particular skill. That said, I have had a number of tubular steel structures professionally welded over the years. I've actually watched experienced welders assemble a steel frame for a home-built aircraft using 4130 thin-wall tube. The entire structure was held rigidly in place by a robust jig during the welding process, and after all the welding was completed and the frame removed from the jig, the frame didn't appear to move even a fraction of an inch; there was NO discernible warping or twisting. Good, experienced welders have learned how to weld large complex structures with a minimum of post weld distortion.

Ahh, but here's the rub, all those superb welders I know all live in the USA, and I now live in Thailand and so will be using local welders of unknown skill level. I've asked around a bit at local machine shops, which welding shops they recommend for complex tubular structures, but I won't actually know for certain until the shop finishes the weldment & I have a chance to examine and measure it for myself.

Knowing that it would be the wildest stroke of luck ever to get the completed structure back from the welders and find that it all lines up to 0.001", I devised a method to adjust the entire "rail bed" in both the X & Y planes, and then securely bolt the whole bed in place. Should the lathe bed or frame ever "creep" out of alignment over time, I will always be able to re-align it.

Keep in mind that my build is a relatively small, bench-top lathe, no where near the size of the one you're building. I'm quite confident that the finished frame (or cage) will have more than adequate stiffness for its relatively short 1 meter bed, and if I do get unwanted twisting during heavy cuts, I can always add more metal at that time. One of my main goals of this build is to keep the weight down, so I will only add more metal if absolutely needed.