[siam]

..... measuring deflection with a static load (bucket of water on the end of the beam) isn't going to tell you anything about how the machine will perform.

An important design characteristic for a small machine is its resonant frequency.
Lets compare an intermittent cut on a part to a pushing a kid on a swing.
Each time the tool starts cutting, energy is put into the machine frame and it deflects a little. Each time the kid swings back, we put a bit of energy into him with a small push.
In both cases, if we push at the wrong time, the energy transfer is inefficient, but as long as we push during the correct half of the deflection/swing cycle, the total energy increases and the kid swings higher or the lathe deflects more.
In a short period of time you have a squealing cut on the lathe and a scared kid on a swing.

Cute analogy, simple & effective....In the physics world this effect is known as "constructive wave interference". And of course you're quite right about static testing being of limited usefulness, and perhaps not even worth the time and effort to perform. In light of this information, it seems quite likely that the only meaningful tests will be when I make real use of the lathe.

You avoid this on a lathe by ensuring that the resonant frequency of the machine is below the excitation frequency. This is where the "decrease the speed" part of "decrease the speed and increase the feed" anecdote comes from.
Increasing the mass of the machine decreases its resonant frequency and improves surface finishes because the 'pushes' happen on both sides of the swing. This is one of the reasons you don't see carbon fiber lathe beds (which could be made quite a lot stiffer than iron. But you do see granite machine bases.

Seems like vibration suppression may be of greater importance than stiffness; thanks to all the input from the readers of this thread, I have lots of ideas to attenuate vibration.

On the topic of mounting rails. I didn't follow too closely, but I assumed that you would weld up the frame, then machine the mounting surfaces flat and square?

You're partially correct; the lathe frame and "rail bed" are two separate pieces, so the rail bed weldment can be easily ground flat if needed before it's installed into the frame. The rail bed bolts onto the frame with ten M10 bolts, two bolts on each of the five 45 degree braces on the lower back side of frame. Multiple "jack points" have been welded into place to allow for easy adjustments of the bed in reference to the spindle line. If the 10 bolts alone prove inadequate to retain the bed's position under whatever usage it sees, I've allowed plenty of room to install multiple alignment pins in each of the 5 attachment bars. I plan to post a few pics of how all this fits together hopefully within the next few days.

[handlewanker]

Hi, one thing must be bourn in mind with CNC and that is it's a multi part operation........who thinks they can set up a CNC controlled machine for a single one off part and get it absolutely right at the end....it doesn't work that way.

All manual machining is a process of walking up to the dimension you want to end up with, correcting all along the way by testing with a micrometer etc for dimension changes.

When you go to a CNC solution, if it isn't right at the end, you do a program fix and make another one.....or two etc.

When I hear 10 " diam turning capacity and CNC in the same breath I realise something just hasn't been thought out thoroughly.

The design is catering for a 10 " diam work piece (over the combined height of the linear rails on the bed and crosslide + the thickness of the crosslide and saddle too)...... which also stretches the entire length of the lathe.....that is ludicrous to the extreme.......and that is to cut steel??????....I don't think so.

What is the actual centre height of the spindle to achieve the work piece dimensions stated?

Perhaps machining steel is limited to 150mm out from the chuck and wood and plastics the rest of the length.

CNC is very limited when it comes to waiting for the outcome and .02mm will scrap even the simplest part if only one dimension is out of tolerance.

Something is not quite right with this design concept........in my opinion, humble or otherwise.

Machining steel of any significant volume can get quite costly if the scrap rate exceeds the finished item rate......not to mention the time it takes to source more material.

How many people have turned common hot rolled (or cold rolled bright ) mild steel and curse the coils of swarf that wrap around the tools?

Most production turning, where steel is specified, is performed with leaded or free cutting steels just for that reason.......I doubt anyone would go to leaded mild steel in large diams just for a limited production run.....and hobby work rarely interprets as production.........leaded mild steel is also quite soft and doesn't have the same strength value, either in shear or tension, as hot or cold rolled.
Ian.

[handlewanker]

Going back to post #1 and the intended work piece diam is 14".max........you would need to have at least 10"centre height to achieve this.......the Colchester lathe I last worked on with that centre height weighed over 3 tons Imperial.
Ian.

[RotarySMP]

"I'm still waiting to get a quote back from one shop and will ask for a quote from at least one other shop that I've recently been told does good work. I also contacted a shop in India that specializes in making spindles, their price is $250 not including shipping. Unfortunately, the only method of payment they accept is wire transfer which adds another $40 to the cost. International shipping cost will be about $50, bringing the grand total to nearly $350. That's not a bad price, but I want to get quotes from the local shops before I make a decision. "

Shame these guys don't do Paypal, like the Chinese. Postage is also much cheaper from China. I just ordered a BoB, and it cost about $7 with free shipping to europe.

Are you going to direct drive, with your 2.2kw motor and VFD, or use some sort of gearbox or backgear setup? I'm working a new mill head for my Deckel G1L engraver frame, and have 1.7kw 400Hz 3Ph industrial servro which I'll run off a VFD. Not yet sure if I'll run a single poly V belt stage or two step poly V pulleys to give a little wider speed range, more torque down slow.

Keep the pictures coming.
Mark

[siam]

Is your weight constraint because you want to be able ship it back to the US after your stay in Thailand?

Thailand has become my permanent home, I truly enjoy the Expat life I have here, so no, future shipping is not the reason. My weight constraint serves three main purposes - 1) I currently rent a house, which has a typical people-size door (30" wide),...no easy job getting a large, 2 ton lathe in and out of my "shop". 2) I find it extremely useful to be able to easily move my machines out into the middle of the room when I'm working on them,...makes making changes, tweaks, or additions to my DIY CNC milling machine I made a few years back oh so easy. My workbench and mill are both on casters, which makes them quite easy to roll about my shop floor. 3) I actually enjoy the challenge of making things as light as possible.

What format are your drawing in? Do you have the ability to save then in step or iges format? I can run it through a simulation for you if you have a completed drawing in the drawing.

I use TurboCAD 18 Pro for all my CAD work, it has the capability to save files in both iges and two STEP formats, however, I doubt my drawings are suitable for modeling in their current state. I've drawn the tubes as simple 3-D boxes, not actual hollow steel tubes,...so any modeling software will see them as a solid.

[siam]

Are you going to direct drive, with your 2.2kw motor and VFD, or use some sort of gearbox or backgear setup? I'm working a new mill head for my Deckel G1L engraver frame, and have 1.7kw 400Hz 3Ph industrial servro which I'll run off a VFD. Not yet sure if I'll run a single poly V belt stage or two step poly V pulleys to give a little wider speed range, more torque down slow.

Keep the pictures coming.
Mark

First, let me say, WOW!!! I like the 400Hz 3 Phase motor with VFD you're using,...that must have cost more than a few pennies.

The 2 1/2 HP motor I'm using is a Permanent Magnet DC unit powered with a Pulse Width Modulator (PWM) with a rated RPM of 3600. This type of motor has very high torque starting at low RPM all the way through to max RPMs, so no need for a gear box. I decided to use a 2:1 reduction in the pullies so get more torque at the spindle and reduce the max spindle RPMs.

[siam]

Hi, one thing must be bourn in mind with CNC and that is it's a multi part operation........who thinks they can set up a CNC controlled machine for a single one off part and get it absolutely right at the end....it doesn't work that way.

All manual machining is a process of walking up to the dimension you want to end up with, correcting all along the way by testing with a micrometer etc for dimension changes.

When you go to a CNC solution, if it isn't right at the end, you do a program fix and make another one.....or two etc.

At the risk of sounding arrogant, I manage to make most all of my CNC milled parts on the first go-round. But to be fair, all the parts I've made to date have been relatively simple, and I haven't yet gone from a CAD drawing file to G-code,...instead I write the G-code myself. I also take the time to do an "air-run", which means that I set the Z zero reference in Mach3 so that none of the milling operations will actually cut any metal, but instead mill out a thin-air part just a few centimeters above the actual part,...sort of a test run.

When I hear 10 " diam turning capacity and CNC in the same breath I realise something just hasn't been thought out thoroughly.

The design is catering for a 10 " diam work piece (over the combined height of the linear rails on the bed and crosslide + the thickness of the crosslide and saddle too)

What is the actual centre height of the spindle to achieve the work piece dimensions stated?

Well, not quite all those pieces,...take a look at this drawing,

Attachment 294122

The design allows turning a 14" diameter round with only the need to clear the cross slide linear rails. The use of linear rails & a single ball screw with anti-backlash nut means that the saddle can be made to look more like a flat plate. The saddle need only be thick enough to prevent flexing caused by the loads on the tool post.

[LeeWay]

Now those drilltaps are way cool,...I got to try to find those here in Thailand. And thanks for the pics,...Thais have their own names for tools, in the Thai language of course, so whenever I go shopping for just about any tool, I've learned to take a picture or two with me that I show the clerks,...helps enormously.

I used to do the same thing in Germany. I was pretty fluent in German, but not technically so. Just when I thought I knew what the German's call something, I was way off base.

[siam]

I used to do the same thing in Germany. I was pretty fluent in German, but not technically so. Just when I thought I knew what the German's call something, I was way off base.

Living in a foreign country, where you don't speak the language, can indeed be a challenge. I take my Thai wife with me to act as translator whenever I go shopping for tools, nuts and bolts, etc, but even she doesn't always know the words used for the item I want, and sometimes isn't familiar with the item itself. Just the other day, I told my wife I needed to buy some lock washers for the M10 x 1.5 x 30mm bolts I had just purchased,...my wife is attorney, and had no idea what a lock washer was, nor what the Thai name for them might be. Yep, it can get interesting over here.

[handlewanker]

[QUOTE=siam;1770158]At the risk of sounding arrogant, I manage to make most all of my CNC milled parts on the first go-round. But to be fair, all the parts I've made to date have been relatively simple, and I haven't yet gone from a CAD drawing file to G-code,...instead I write the G-code myself. I also take the time to do an "air-run", which means that I set the Z zero reference in Mach3 so that none of the milling operations will actually cut any metal, but instead mill out a thin-air part just a few centimeters above the actual part,...sort of a test run.
Well, "if the tool path looks right as it moves above the job, then it must be right when the tool cuts the metal"?

That is weird........if the machine is incapable of accurate work, no matter how accurate the tool path is, the actual work piece will not be to drawing specs.

Summed up, my point is that if the machine flexes even .02mm in the spindle to bed area (and a high centre height exacerbates this).......the tool point is doing that too.......results are the job is not going to be to drawing and rework is either impractical or not possible if the part is undersize on any dimension.

I'm new to the CNC world, but I have experience with manual stuff to know that machinery cannot be flexible if you want to produce parts to drawing....CNC is machining with your fingers crossed, as the final dimension is totally up to the machine and the operator just takes a back seat.....there is no capability to correct a dimension once the stepper/servo motor....whatever.... does it's thing......your sizes are all pre ordained by the in built accuracy of the machine itself.

I would almost go so far as to say you need two CNC lathes.......one to produce the small parts and is able to be made rigid enough to maintain tolerances etc, and the other for the larger diams.....with a fairly wide tolerance span.

BTW, the drawing link you posted cannot be opened.... or so it says.
Ian.

[siam]

Well, "if the tool path looks right as it moves above the job, then it must be right when the tool cuts the metal"?

No, of course not, but it is a great way to insure that you haven't forgotten to lift the tool bit out of the 20mm deep pocket you just finished milling before you move 80mm over to begin cutting the next pocket.

I'm new to the CNC world, but I have experience with manual stuff to know that machinery cannot be flexible if you want to produce parts to drawing....CNC is machining with your fingers crossed, as the final dimension is totally up to the machine and the operator just takes a back seat.....there is no capability to correct a dimension once the stepper/servo motor....whatever.... does it's thing......your sizes are all pre ordained by the in built accuracy of the machine itself.

My first steps into the CNC world was when I finished my mill build a few years back, for me, it was love at first use. I absolutely enjoy being to write the G-code that tells the machine to move X to exactly 3.452", then move Y to 4.500", and at what feed rates to make each move,...etc, etc. So much easier than counting handle turns.

BTW, the drawing link you posted cannot be opened.... or so it says.

Not sure whats going on there,...it opens for me with no trouble. I'll post it again and see if works...

[handlewanker]

Hi....OK, part now seen.

It has oft been said, especially on the forum, that a ball screw is capable of moving the tool in ever so fine increments in many directions........provided the infrastructure also supports the tool in a rigid tool path.

I think you are being ambitious in the spindle height....the higher you go the more leverage you get to the bed and the tool will not be able to stay on the work piece even though the ball screw moves it there.

In my opinion, you may be able to machine the big diams, but not with any accuracy, and that is the same for anything smaller.

The height is the centre of the spindle to the sliding ways of the bed that carries the tool,...... no matter if the bed with sliding ways is lying at an angle

If you sacrifice rigidity for manoeverability....something has to give.

Make it and prove me wrong.......proof is in the pudding....LOL.
Ian.

[siam]

The following pics show what I call the "rail bed" (the green part) bolted into the lathe frame with ten M10 X 1.5 X 30mm bolts. The adjustment bolts worked beautifully, allowing me to align the rail bed with reference to the spindle line to within 0.5mm, top and bottom rails, and headstock to tailstock, which is all I was trying for on this initial alignment. Now I can proceed to the next step of mounting a flat plate to the rail bed, insure it's flat to within the best tolerances I can get, and then mount the linear rails to the rail bed assembly.

Attachment 294194

Attachment 294196
Above is a close up of 3 of the adjustment/alignment screws,...3 additional screws (bolts) are located at the other end of the rail bed.

Attachment 294198
If you look closely, you can see some of the M10 bolts in the lower corners of each of the mounting plates.

Attachment 294200
I used a length of Aluminum channel I had laying around the shop and aligned it to the centerline plane. For this first, initial alignment I just wanted to insure that the spindle centerling alignment was in the ball park, and could be adjusted as I had planned. No need to do a better alignment until I have a spindle & chuck mounted.

[pk7639]

OK I get it now. The rail bed is where you need to add your stiffness and mass. The mounting points you have partially isolate it from the mass of the rest of the chassis. So mass and stiffness in the rest of the machine won't make as much of a difference...

If you have any problems getting that plate it flat. You could consider replacing it with some artificial stone kitchen benchtop. That stuff is surprisingly flat, massive and has excellent dampening characteristics. It's a bit of a ***** to get threaded inserts into, but it can be done... I've used it on two machines so far...

[lcvette]

It appears that the bokts are all on a flat steel tab.. Which would tell me that the whole rail table can flex in the z axis but will be less flexible in the X axis. What happens if you pop it with a hammer? Does it ring? Try it in both the z and x direction, if there is a different tone it will tell you which is the better secured.

My guess is the z direction will resonate more as it is depending on 5 flat tabs which will allow some more movement and a different frequency. The suggestion about adding a stone top is a good one, it would need to be epoxied to whatever sub structure you might connect it to too also so it will help quiet the frame down. You can use a diamond bit to bore holes through it and use a T type insert from the back to the top face and use epoxy on them as well and you have a nice flat mounting surface with threaded holes for your rails.

It is actually a great suggestion, you might consider working it into your build.



Sent from my XT1080 using Tapatalk

[The Blight]

"I'm new to the CNC world, but I have experience with manual stuff to know that machinery cannot be flexible if you want to produce parts to drawing....CNC is machining with your fingers crossed, as the final dimension is totally up to the machine and the operator just takes a back seat.....there is no capability to correct a dimension once the stepper/servo motor....whatever.... does it's thing......your sizes are all pre ordained by the in built accuracy of the machine itself."

Not true at all. You can do the same in a CNC machine as you do on a manual. Just run multiple passes where you measure in between as you would on a manual lathe. I have done so several times without scrapping any parts. The trick is to know how the machine you are working on behaves under certain conditions.

[siam]

OK I get it now. The rail bed is where you need to add your stiffness and mass. The mounting points you have partially isolate it from the mass of the rest of the chassis. So mass and stiffness in the rest of the machine won't make as much of a difference...

If you have any problems getting that plate it flat. You could consider replacing it with some artificial stone kitchen benchtop. That stuff is surprisingly flat, massive and has excellent dampening characteristics. It's a bit of a ***** to get threaded inserts into, but it can be done... I've used it on two machines so far...

I'm considering using either 6mm flat Aluminum plate (not sheet) or G-10 (Epoxy Fiberglass). I've used both materials in the past and both are extremely flat. G-10 is an amazing noise dampener,...it simply will not resonate, but I'm concerned about it's flexibility. A sheet of 6061 Aluminum plate would also be very flat and I could count on it not flexing while I glue it to the steel tubes, but it's not nearly as good at noise dampening as G-10. Maybe pricing each one will help me decide.

[pk7639]

Aluminium is generally considered to be about the worse material to make machines out of. Low mass, reasonable stiffness, low damping is a bad combination. The G10 sheet could be good. But it's expensive. If the piece of plate you need is less than 600mm wide then you may be able to get it for nothing in synthetic stone. Small pieces aren't useful in kitchens and bathrooms, so they tend to just get thrown out. I bought a square meter at full price recently. It was only AU$150.

[RotarySMP]

Hi Siam,

There is a guy on the Austrian Equivelent to Craigs list who gets Lenze Servomotors from some industrial application, where they are replaced on a fix interval contract. I asked about getting one, and he made me such a good offer for two, that I now have a spare.
https://www.willhaben.at/iad/kaufen-...tor-134384129/

If you are interested, contact him. I am sure he will be fine with english.

How are you planning to do the headstock on your lathe.
Mark

[handlewanker]

"I'm new to the CNC world, but I have experience with manual stuff to know that machinery cannot be flexible if you want to produce parts to drawing....CNC is machining with your fingers crossed, as the final dimension is totally up to the machine and the operator just takes a back seat.....there is no capability to correct a dimension once the stepper/servo motor....whatever.... does it's thing......your sizes are all pre ordained by the in built accuracy of the machine itself."

Not true at all. You can do the same in a CNC machine as you do on a manual. Just run multiple passes where you measure in between as you would on a manual lathe. I have done so several times without scrapping any parts. The trick is to know how the machine you are working on behaves under certain conditions.

Hi......I may be wrong but when the tool is cutting multiple diams at each pass......just one diam/length dimension out of tolerance and the whole job is scrap......unless the customer is benevolent and overlooks a few out of spec sizes for a concession on the price.

If the machine is so wonky that it takes multiple goes and operator hands on input to get it right.......that ain't CNC.

CNC is after all a production tool, as in multiple parts to drawing, scrap all those that don't measure up, just keep making until the order is complete.

The last firm I worked at had quite a few Nakamura Tome CNC lathes, and with one of the jobs we had an acceptable scrap rate that filled a 44 gallon barrel....frequently.........the loss due to scrap and the recovery with scrap metal made the loss at cents in the dollar......we gauged parts at 1 in 10 on the run and scrapped the previous 9 if one part started to show out of tolerance.....no rework or reclamation was done.

Now when I hear someone mention resolution of the ball screws and linear rails, I have to wonder if the parts produced are ever measured before delivery.

The build of a machine can make any accurate peripheral null and void if it allows deviation from the programmed path......that's why tonnage does count a lot when you get serious.

With any one off build, a lot does depend on what the acceptable outcome is compared to what it should be.

BTW.......I love the beige colour of the machine frame work......our Nakamuras were that colour before the works manager....... after the consultants advised him to brighten the place up a bit....... decided that they would look better painted a pale baby blue.....hand painted that is.......geesus wept.
Ian..