[LazyH]

Hello again everyone. Since my plans for the (polymer) concrete mill have been changing so much and put on hold for now, I figured I might build something a bit more practical in my small shop, a cnc lathe. It seems like there is a bit less information than on milling machines though, more specifically on software and hardware to sync spindle and tool speed/location for things like threading. I have three ideas on this and would probably prefer the most budget friendly one that still provides decent results. This is what I'm considering so far (from most to least expensive if I understand correctly):

1) Dedicated servo and hardware for the spindle for precise speed control and positioning, may even be able to use as a turn/mill machine if i decided to upgrade.
2) Simple DC motor without speed control (besides belts/chain/etc.) encoder to determine position and adjust steppers accordingly.
3) Dual shaft stepper on leadscrew to carriage, cnc program allows control for shaping, but connect a simple belt drive to other end of motor shaft connected directly to spindle with appropriate pulley sizes. Considering how fast the motor would have to move if fully cnc'ed, I doubt it would cause any damage to the stepper if I had the electronics manually disconnected for screw threading, and belt/pulleys disconnected for cnc'ing.

I like the simplicity of the third one, since programming isn't my specialty, cnc and g-code have been really difficult, but worth it for me to learn, but I wouldn't mind learning how to set up an encoder or servo system if it's affordable enough or the advantages of precision and ease of use once set up are worth it to start with. I'd like to learn the more complicated servo and encoder skills eventually anyways but I don't know how affordable it might be, and I can always switch around the motors and controls later if it does end up being too expensive at this point in time.

Now, as for my lathe design itself I'm planning something similar to my mill idea with a rebar space frame design and tapered rollers for the spindle. I did a quick search for cutting force calculator and determined for the kinds of cuts I'd be doing I'd need to be able to handle about 1,500 lbs of force at most. I found a couple of used ebay sets from IKO and NSK linear bearings, both 15mm and each rated at 2000 and 2500 lbs with all four bearing blocks together. I know that not all four will hold the forces equally so I might have to take some cuts a bit lighter and make sure I design this thing to put as much force evenly on the bearings as possible, so I have a couple of questions about a few arrangements I have to see what would be able to hold the most cutting force without sacrificing too much travel on the axes. I'll upload a few pics to see what you guys think, any help or reviews would be greatly appreciated.

[LazyH]

I'm planning on using some 1/4"+ flat iron welded to square tubing to make a place to bolt the rails, even if the curing epoxy does cause any warping, the steel can be made flat before drilling bolt holes for the rails, and the square tubing will allow room for the drill and tap without dulling on the epoxy mixture. The top of the steel should come just above the top of the epoxy surface.

I'm planning on using a single piece of hot rolled steel with the same bearings as in my mill thread, about 1 1/16th if memory serves correctly. I might use cold rolled to make a temporary spindle to turn the hot rolled between centers, but even though it's not as strong, I like the resistance to warp that you get with hot rolled, and since it's cheaper I can get a piece with the extra diameter I need to make the flange and threads for the lock nut. The lock nut design isn't mine, I can't remember where I saw it but someone made a lock nut with two slits cut into it and set screws travelling parallel to the spindle axis, when torqued down, these set screws would spread the slit metal apart, tightening around the threads and preventing the lock nut from moving. I've seen several other lock nut designs, but these usually require three hands (one to wrench the spindle down and one each for two lock nuts tightened together) were unnecessarily complicated, or pushed the setscrews against the threads, galling them up quite a bit. I haven't yet decided on either a pulley or chain drive, but I'm thinking chain since I have easy access to some good quality chains where I'm at, and you can get the same or better mechanical connection between the motor and spindle for the same or better price. Unless those with more experience think it's not a big deal, I'd be a little more concerned about belt slip, for some reason that's always freaked me out, whether it's machinery or cars or anything else, I just cringe at the thought.

Again, any assistance would be great, otherwise I'll just have to turn this into a build thread with my trial and error trying to get this machine working.

[Eldon_Joh]

One of my projects that has sat outside for a year rusting, is a lathe. flat ways, 1 inch by 3 inch solid.. not yet bolted to a weldment consisting of 380 pounds of 9 inch wide channel iron, 7.5 feet long, welded like so: [[]] , with the ways on top, atlas style. the channel iron would be filled with dry sand of course and after that it would weigh 1000 pounds without the headstock. moment of inertia torsionally is on the order of higher than that of a similarly sized cast iron lathe, however, the chips won't fall through the bed which means they pile up between the ways. but I would never use such a machine for production so what does it matter.

I'm intending to use 3x 9117 bearings (aka 6017) bearings for the spindle, probably for a 16 inch swing.. some have said they aren't big enough.
I have two options for the spindle, a lightweight steel pipe which would leave me with a 3 inch spindle bore, or a very heavy pipe which would leave me with a 2.75 inch spindle bore... hmm.

anyhow point being: you mention 1500 pounds of cutting force.. that's more than i intend to deal with.

i will probably cast the headstock from aluminum even thought I really don't want to, due to the almost negligible dampening coefficient of aluminum. I may end up using a hypoid rear end differential ring gear and pinion from a small car to drive the spindle.

[LazyH]

One of my projects that has sat outside for a year rusting, is a lathe. flat ways, 1 inch by 3 inch solid.. not yet bolted to a weldment consisting of 380 pounds of 9 inch wide channel iron, 7.5 feet long, welded like so: [[]] , with the ways on top, atlas style. the channel iron would be filled with dry sand of course and after that it would weigh 1000 pounds without the headstock. moment of inertia torsionally is on the order of higher than that of a similarly sized cast iron lathe, however, the chips won't fall through the bed which means they pile up between the ways. but I would never use such a machine for production so what does it matter.

I like that channel iron idea, I wanted to cut up an old tractor with a front loader on it, way too small for most any work so it wouldn't be missed much, had 6 inch channel iron pretty thick too, but everytime I look at it I see my grandpa driving me around in it while I was a kid, so probably not going to happen. If I ever get a hold of an old truck frame or something like that I'd be tempted to fill it with epoxy aggregate mix just to prevent settling over time as the machine gets used, could probably use a long broom handle or something to tamp it all in place from the end.

I'm intending to use 3x 9117 bearings (aka 6017) bearings for the spindle, probably for a 16 inch swing.. some have said they aren't big enough.
I have two options for the spindle, a lightweight steel pipe which would leave me with a 3 inch spindle bore, or a very heavy pipe which would leave me with a 2.75 inch spindle bore... hmm.

Sounds good to me, especially since you say you're not going to use in a production environment. I'm no engineer so I can't say for sure if this rule of thumb works, but on some of the smaller lathes I was looking at, the spindle shaft was far less than 1/5th swing, so ~3 inch on a 16 inch swing sounds good to me at least, just take it easy on the cuts until you're familiar with the machines limits. Question though, is there an advantage to using angular ball bearings vs tapered roller bearings? It would seem the extra surface area of the rollers would improve rigidity and strength but I'm not sure. They're pretty cheap from chinese ebay sources but I guess if you'd rather buy local that's another factor too.

anyhow point being: you mention 1500 pounds of cutting force.. that's more than i intend to deal with.

That's what the online calc said, like I said I'm no engineer. I can't remember where I went but I just put down the biggest possible cuts I might get while using steel. It was kind of confusing though since in another thread I heard someone state at least 25mm for the main ways as a minimum, though I don't remember what size of machine they were advising on. The main ways on mine will reach about 26", so definitely not doing major production work on this thing either. I guess I didn't list on my earlier post, but the IKO rails are 26" and the NSK (cross slide) are just under 12 1/4" long, both 15mm. I still have to get an extra set of blocks for the IKO rails so I can have a tailstock to hold longer pieces, or to add a steady rest since a few friends were asking me to make a cannon barrel for them if I ever got this thing working.

i will probably cast the headstock from aluminum even thought I really don't want to, due to the almost negligible dampening coefficient of aluminum. I may end up using a hypoid rear end differential ring gear and pinion from a small car to drive the spindle.

If you make the casting with lost foam molds, you can have a ribbed structure using styrofoam sheeting that can later be filled in with sand or epoxy aggregate to improve damping. I like the idea of a machine that can align itself, since I read the gingery books I was planning a couple aluminum inserts and a temporary boring bar hooked up to the cross slide to machine a precise groove for the inner bearing race to press fit, just to avoid any potential shrinkage or template errors trying to align the inserts pre-casting. Since the aluminum would be so thin on mine, it shouldn't make it much, if any, less rigid or damping (I hope).

[LazyH]

Well, I wasn't expecting the rails here, but my 1 and a half year old found my computer and started pressing buttons everywhere, most notably the big blue "BUY NOW" button on ebay. So now I've got both sets of rails here with nothing I can use them for yet. Hopefully they stick around without damage until I can get everything together for this build. (please excuse the kids toy in place of the tool holder).

[LazyH]

Ok, I revisited the IKO bearing catalog and each 15mm bearing of the lwl series is rated at 3490 Newtons, which translate to 780ish lbs each, which times four would be about 3000 lbs that it can hold. The NSK bearings are rated at least 4410 Newtons which is about 990 lbs each or close to 4000 lbs total. This is with the absolute minimum rating for the rail size from that series, it's possible I have some of the higher rated models of the lwl or rsr series linear ways, but either way I think we'll be ok in terms of handling the cutting forces with these rails. I actually have nsk bearing blocks on a thk rail, and I was going based off of the thk rating, but the nsk bearing blocks look like they have even higher ratings than the thk's of the same size, based on their numbers at least.

I was thinking that the most cost effective method of leveling the rail mounts would be to hand scrape them with a surface plate reference. Since I'm aligning the bore with the rails once the rails themselves are set up I can focus on getting the ways parallel, straight, and level. I got myself a precision machinist level and a set of homemade hand scrapers I've used for smaller jobs like mating surfaces on a car (fixing warped throttle bodies and the like). All the forums say scraping steel is too difficult but personally I've never had any trouble with it. I just hope my surface plate is big enough for the whole ways or else I'll probably have to take parts at a time and overlap my previous work. Where I'm at, the dollar is harder to come by than spare time, so diying it rather than paying someone else to grind it is what would work best for me now.

[Eldon_Joh]

It might be interesting to think about epoxying in cast iron inserts into solid granite, epoxy granite, or just cement, then line boring (or grinding) the cast iron to fit the bearings for the headstock. The only problem being the differential thermal expansions.

anyhow my bearings aren't angular contact but just regular radial ball bearings which means their maximum contact angle might only be 15 degrees. two of them spring preloaded at the business end of the spindle with enough travel on the spring to cover the thermal expansion, and the third floats at the other end. (or light spring preload)


I scraped a south bend 9, 42 inch bed with a 38 inch aluminum straight edge and a dremel with 1/2 inch by 1/2 inch grinding stone. i had to remove .015" of metal from the front V way before it was flat, perhaps .003" from the rear v way. (38 inch AL straight edge only cost me 30$ off of ebay and it was guaranteed to be .003" or better. It turned out to agree with my dad's 56 inch long cast iron optical table to within .002", and i suspect both are concave.

Anyhow the dremel made very quick work of that soft cast iron. it would almost be easier to use a grinder on steel than cast iron because its harder to remove, more difficult to mess up.

[LazyH]

I've discovered the Fusion 360 program, since I don't make anywhere near 100k per year yet I can use it for free, and it really is an awesome program. I tried modelling my space frame design on its own, without the epoxy granite filling. The simulation part of Fusion 360 with low temper low carbon steel of the same diameter as the rebar I'd be using states less than 0.005" displacement at the highest possible force those bearings could be capable of, 10kN. That sim I ran with just a box frame, no triangulation of the space frame either, I might have to upgrade to larger linear bearings to make the machine more capable if the frame really is as rigid as the simulation states.

[hanermo]

I actually build and have built tools, but ...
Suggest you upgrade your (impact) loads 10x.

A tiny 10x lathe, has 1.500-2000 kg or 10.000 lbs impact loads.
Size 15 linear rails are not approriate.

A very good small lathe, of size 12x20-30", has structural strength of approx 100 metric tons (before deformation ie failure).
Monarch 10EE, Hardinge HLV-H, Geeler, Sharp (clones).

A tiny toy minilathe 7x, has bearings and spindle/hs for 1200 kgf load, == 3000 lbs.

[LazyH]

That's not too bad, Pretty much anything I'd be using the lathe for would be about the same size as what I would have on a mini lathe. The extra travel is for light duty milling in aluminum or for turning longer shafts which will have a smaller diameter and therefore less torque, so those ratings should be just fine. Until I get my purpose built mill project started, I'll be adding a milling head to the back of the lathe, since I don't have much room in my shop (4'x5' storage room) so combining both into one would be very helpful in terms of space. Taking really light cuts wouldn't be a problem for the work I'd be doing and I can always upgrade to larger bearings later. Like I said, they're just the bearings I'm stuck with for now, so I'm not against recycling them for other projects when I find a better set.

[Hezz]

That's not too bad, Pretty much anything I'd be using the lathe for would be about the same size as what I would have on a mini lathe. The extra travel is for light duty milling in aluminum or for turning longer shafts which will have a smaller diameter and therefore less torque, so those ratings should be just fine. Until I get my purpose built mill project started, I'll be adding a milling head to the back of the lathe, since I don't have much room in my shop (4'x5' storage room) so combining both into one would be very helpful in terms of space. Taking really light cuts wouldn't be a problem for the work I'd be doing and I can always upgrade to larger bearings later. Like I said, they're just the bearings I'm stuck with for now, so I'm not against recycling them for other projects when I find a better set.

LazyH,

If I were you I'd aim a little higher than the performance of a mini lathe. I have a mini lathe and it is useless for any kind of accurate work. The ways and the carriage are so flexible that even cutting plastic to an accurate tolerance is difficult. I think that the only way to alleviate this problem with the mini lathe is to bolt it down to a very flat and thick piece of steel to improve the rigidity of the whole lathe. Even deflection of .001 of an inch will adversly effect how the cutting tool operates. I'm not sure of the size of the tubing that you are using but it look like only 1 inch or so. I know it's a bit more expensive but if you can go up to a 2 inch size it will make the frame a lot stiffer. Then if you put a piece of rebar and EG fill in a few of the critical frame members that will really up the dampening and also increase the stiffness.

Even if you don't do any EG and rebar at least go to a 2 inch size tube with the thickest wall you can afford.

Essentially, you have made a steel tubing torsion box which can be very stiff. But you might consider skinning the bottom with a thin piece of steel and the top with a little thicker piece of steel and as many of the sides as will allow you to get a welder in there.

In fact, your approach has given me an idea for a pretty stiff and relatively light lathe frame. If you take your frame and add a 1/4 inch thick top plate and 1/8 inch bottom plate and 1/8 sides plates but leave some holes on one side plate. You could then stress relieve or not and then have the top plate ground to be very flat. But first you could fill the whole entire inner cavity with expanding structural foam. It will damp the entire structure and keep all of those inner braces from resonating. At least to a large degree.

In fact, the way to get the ultimate strong and light structure would be to make the frame work as you have shown and then weld in square pieces of flat 1/8 inch steel with holes cut in to allow expanding foam to pass through. You could use any cheap used 1/8 plate because it could have holes in it and even be slightly warped. I think that you could place them in there by hand one at a time and weld them in there.

Paul

[LazyH]

According to the sim at least, the base itself will be rigid enough to hold over 100kN with less than 0.00002" deflection translated to the carriage/cross slide when filled with plain epoxy, not even epoxy granite or even triangulating the box frame to make it a space frame (I love experimenting with Fusion 360, make a box surrounding the whole frame with enough room for the ways on top, cut while keeping tool with the frame, and set material characteristics to epoxy and contact type as bonded no slide or separation). The main problem with rigidity now is the cross slide and carriage itself now, they distort a total of 0.0008" at 10kN, and that's not including the tool holder, as you can see from the pic I've included it almost seems like a twisting motion which will cause the tool itself to move even further considering it's on the end of what is basically a long lever. Like I said, I'd be more than happy to upgrade to stronger bearings when I can find them at an affordable price, that's why I left the square tubing under the ways, to make it easier to tap and drill with hand tools if the new ways have different pitch between holes, I can fill in the old ones with welds and drill/tap a new set for the new linear bearings. I've been looking at some 25 and 30mm linear ways too.

Larger diameter tubing can be more rigid, I know in tubular car frames, larger diameter is more rigid than larger thickness for a given weight. The difference with this scenario is that when filled underneath with epoxy there is more of a crushing distortion than bending distortion, in which case a larger diameter tube would be less rigid.than smaller diameter, and this theory is supported by my simulations as well. I went with 1 inch because I'm thinking 25-30mm ways would be the ideal size when I upgrade and 1" tubing gives a good balance in terms of room and rigidity.

[Hezz]

According to the sim at least, the base itself will be rigid enough to hold over 100kN with less than 0.00002" deflection translated to the carriage/cross slide when filled with plain epoxy, not even epoxy granite or even triangulating the box frame to make it a space frame (I love experimenting with Fusion 360, make a box surrounding the whole frame with enough room for the ways on top, cut while keeping tool with the frame, and set material characteristics to epoxy and contact type as bonded no slide or separation). The main problem with rigidity now is the cross slide and carriage itself now, they distort a total of 0.0008" at 10kN, and that's not including the tool holder, as you can see from the pic I've included it almost seems like a twisting motion which will cause the tool itself to move even further considering it's on the end of what is basically a long lever. Like I said, I'd be more than happy to upgrade to stronger bearings when I can find them at an affordable price, that's why I left the square tubing under the ways, to make it easier to tap and drill with hand tools if the new ways have different pitch between holes, I can fill in the old ones with welds and drill/tap a new set for the new linear bearings. I've been looking at some 25 and 30mm linear ways too.

Larger diameter tubing can be more rigid, I know in tubular car frames, larger diameter is more rigid than larger thickness for a given weight. The difference with this scenario is that when filled underneath with epoxy there is more of a crushing distortion than bending distortion, in which case a larger diameter tube would be less rigid.than smaller diameter, and this theory is supported by my simulations as well. I went with 1 inch because I'm thinking 25-30mm ways would be the ideal size when I upgrade and 1" tubing gives a good balance in terms of room and rigidity.

It appears your whole space frame is full of epoxy. If that's the case you can see why the results show such rigidity. Of course without the aggregate fill the cost will be very high. I believe the aggregate fill will increase the compressive strength over plain epoxy. If your planning this massive EG base then I agree that 1 inch is certainly good enough. For some reason I was thinking that you were trying to build a pure space frame structure.

[LazyH]

It appears your whole space frame is full of epoxy. If that's the case you can see why the results show such rigidity. Of course without the aggregate fill the cost will be very high. I believe the aggregate fill will increase the compressive strength over plain epoxy. If your planning this massive EG base then I agree that 1 inch is certainly good enough. For some reason I was thinking that you were trying to build a pure space frame structure.

I should make a quick correction, I made a silly mistake in Fusion 360 and put 100kN PRESSURE on the ways, which apparently distributes the force very evenly across the entire surface. Instead, I should have placed blocks on the ways and put 25kN on each. Either way, the amount of compression distortion is still pretty negligible even with plain epoxy. I think there's a way to add custom materials in Fusion 360's local database, I'll have to find the properties of the aggregate I'd be using and see if I can add the material properties for a more accurate result. I was advised by someone, I believe on the thread with my gantry mill idea (a design I'm not sure I'll follow through with, after much more learning since I conceived the idea) that I should attempt some very light cuts with only the space frame to make sure the design is solid, prior to adding the EG.

I'm trying to add the spindle, another part I'm 100% certain I'll upgrade once I get better materials. Right now I'm recycling some tapered bearings for a one inch bore, but I'm thinking something a bit more substantial would be more effective at providing the rigidity I need for good cuts. I'm attempting the simulations of the spindle tonight, but does anyone know how accurate the distortion would be, considering that in practice the spindle and workpiece will be rotating? I don't know but it seems like a load against a stationary shaft might cause different amounts of deflection vs the same force on a rotating shaft. There's so many little pieces to build when simulating an actual spindle though, hopefully I'll get it right. In the meantime I have someone who wants me to build a sound system for them so my time might be taken up for a week or two until I get that project done. Hopefully I'll post back soon, maybe with the wife convinced to invest in some stronger linear bearings too (doubt it but fingers crossed).

[Eldon_Joh]

do you have any information on the stiffness of the bearing blocks.

[LazyH]

do you have any information on the stiffness of the bearing blocks.

I couldn't find anything on the manufacturers site, just static and dynamic load ratings (and torque of course) I'm not sure even contacting the manufacturer would help though since the shorter ways are THK blocks on NSK rails, that's another reason I'm willing to upgrade whenever I find better rails, I can always reuse these for a router or wet grinder or something like that.

[ebrewste]

NSK lists stiffness in the preload and rigidity table. Probably not a bad source to make rough estimates even on other manufacturers products.

http://www.jp.nsk.com/app01/en/ctrg/...Down&pno=E3329

http://www.jp.nsk.com/app01/en/ctrg/...Down&pno=E3332

[Hezz]

I should make a quick correction, I made a silly mistake in Fusion 360 and put 100kN PRESSURE on the ways, which apparently distributes the force very evenly across the entire surface. Instead, I should have placed blocks on the ways and put 25kN on each. Either way, the amount of compression distortion is still pretty negligible even with plain epoxy. I think there's a way to add custom materials in Fusion 360's local database, I'll have to find the properties of the aggregate I'd be using and see if I can add the material properties for a more accurate result. I was advised by someone, I believe on the thread with my gantry mill idea (a design I'm not sure I'll follow through with, after much more learning since I conceived the idea) that I should attempt some very light cuts with only the space frame to make sure the design is solid, prior to adding the EG.

I'm trying to add the spindle, another part I'm 100% certain I'll upgrade once I get better materials. Right now I'm recycling some tapered bearings for a one inch bore, but I'm thinking something a bit more substantial would be more effective at providing the rigidity I need for good cuts. I'm attempting the simulations of the spindle tonight, but does anyone know how accurate the distortion would be, considering that in practice the spindle and workpiece will be rotating? I don't know but it seems like a load against a stationary shaft might cause different amounts of deflection vs the same force on a rotating shaft. There's so many little pieces to build when simulating an actual spindle though, hopefully I'll get it right. In the meantime I have someone who wants me to build a sound system for them so my time might be taken up for a week or two until I get that project done. Hopefully I'll post back soon, maybe with the wife convinced to invest in some stronger linear bearings too (doubt it but fingers crossed).

LazyH, some other Autodesk products have concrete as an engineering material. It's probably similar in mass and compressive strength to EG you might check Fusion to see if there is a concrete option.

[RotarySMP]

Does Fusion let you plug in these numbers for E/G?
Youngs modulus of around 40 GPa (higher is unlikely in the hobby environment),
Density around 2.3 g/cm³
Poissons ratio around 0.25.

Mark

[rowbare]

A tiny 10x lathe, has 1.500-2000 kg or 10.000 lbs impact loads.
Size 15 linear rails are not approriate.

That is interesting. I would have thought size 15 rails to be sufficient for that size lathe. What size would you recommend?

bob