DIY Mill Turn Center

Show off your latest creations, whether 3D printed or not
User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

DIY Mill Turn Center

Postby Generic Default » Sat Oct 31, 2015 8:43 pm

So over the last few years of 3d printing I've come to the conclusion that 3d printers are capable of making complex plastic parts better than most CNC machines can. With the exception of round parts.

Anything that needs to be very round must be made on a lathe or other turning type of machine. Even the most precise, high resolution printers cannot make parts that are nearly as round or dimensionally accurate as what lathes can do. There is no trend towards this, either, and none of the current 3d printing technologies seem capable of it.

I'm building a CNC mill turn center on a fairly low budget. The end goal is to have a working machine that can make complex parts in any/multiple material[s] with excellent part tolerances and smooth surfaces.
The goal is to make something capable of nearly all precision machining operations, so that raw materials are put in one end and a finished part comes out the other. Normally this would be prohibitively expensive but the cost of CNC has dropped to less than $100 per axis in the last few years, thanks to the growth of DIY 3d printers. The same stepper motors and controllers are fully capable of controlling a regular CNC machine.

Some features this machine will have;

Sliding headstock (Main spindle)
Sliding Subspindle for backworking
Live Gang tooling (radial and axial)
Live tool turret (radial and axial)
Full rotary indexing on all spindles
Through spindle bar capacity of up to 1 inch (25mm)
Chuck capacity of up to 3 inches
Automatic bar feeding/pulling/clamping
Tool position setting


A few of the features listed are not critical to the machine being able to make complex parts, and are only for convenience and productivity. So a few of them will be added on later after the base machine is built and working.
The first major step is to get two working axes that are perpendicular, plus a precise spindle that is aligned with them. This is all I really need to have a basic lathe. CNC'ing the axes is relatively easy since I'm sticking with established 3d printer control electronics that I'm familiar with.

A few pictures, the first being a stripped and simplified machine base so you can get an idea of what I'm going for;

lathe basic 1.jpg



Next is a rough prototype of the sliding headstock and spindle;
headstock rough 1.jpg


And a printed version of the tool turret that will go on the cross slide. The real one will be 4 inches across instead of 3 inches.
headstock rough turret.jpg


This is what the headstock spindle blocks looked like when I was machining the bearing journals, which are the only difficult part for me;
headstock face block 1.jpg

headstock face block 2.jpg


Notice in the picture above that I'm single point boring the journals for concentricity and roundness using my near zero budget homemade boring head and indexable insert tool from Harbor Freight.
Unfortunately I screwed up and bored it a hair oversize in diameter, so the bearing race is loose inside and will never be good enough for a final prototype. It still works for now though!


I also have the main machine base partially machined; all of the other parts and axes will bolt to it so I can align stuff within 0.001 over the length of the machine. More pictures in the next post, I forgot to take them so I'll get to it tomorrow or when I have time for more machining.

A few last notes on this post;

1) Nearly all parts will be machined metal in the working version; printed parts are temporary for fits and testing.
2) The headstock and subspindle will be precisely made and reinforced for rigidity. Slides will be longer with a lower moment.
3) The headstock and subspindle will have a clutch mechanism to switch between a stepper motor for indexing/positioning and a regular motor for continuous and fast spinning.
4) The tool turret is driven by a geneva mechanism; this has been tested successfully but not with huge cutting loads yet.
5) So far I've been stuck with 3 inch flat bar stock as a maximum size, which has limited my design. I'll get around it soon.

6) Here is a video of a mill turn center example part

https://www.youtube.com/watch?v=5kiLPUBAWLg

My machine will be less precise, fast, and productive, with a small maximum part size. But I want the same capability!
Check out the Tri hotend!

User avatar
Captain Starfish
Printmaster!
Posts: 960
Joined: Tue Dec 10, 2013 5:24 am

Re: DIY Mill Turn Center

Postby Captain Starfish » Sat Oct 31, 2015 9:08 pm

Good idea on the geneva mechanism, the angular load on the turret spindle would be pretty savage when you start loading her up.

Any reason you've gone for carriage feed under the head? Most designs, particularly for lightweight lathes, I've seen have the headstock and motor fixed as rigidly as possible to the X ways and the tool head does the moving. A little different when you get to two tonne machines.

Where's your live tooling??? :D

Keep posting progress please, be great to see this little beast develop. And nice to see something other than the usual 3 or 4 axis mill being built. My 4th axis is a little weak but, without a 5th axis, I can't orient it correctly to do axial boring operations. So they have to be done manually in the lathe before transferring the part to the CNC mill for the "circumferential" work.

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Sat Oct 31, 2015 9:31 pm

The geneva mechanism never slips, and can be operated with a single stepper motor instead of a clamping mechanism (air, hydraulic, or solenoid) and a separate motor. Plus the repeatability is excellent.
I can machine the entire geneva plate from one side using cheap aluminum flat bar. The live tooling is holding me up a bit. The plan is to rotate the tool into position with a regular two-bearing live tool holder design so that a gear on the tool shaft meshes with a fixed position drive gear. In total, that big rectangular block from the first picture will probably have three spindles through it; one for the main geneva turret, one for the indexing stepper, and another for the live tool driving motor.

I might get more complex if I end up clutching the live tool with both a stepper motor for rigid tapping and a continuous spinning motor as well. I may just go for a servomotor if the price is right. It would be simple to have a separate motor directly driving each live tool, but that would be very space inefficient. I want to drive all of the tools with the same motor[s].


The turret will never rotate from axial loads like face drilling, or from center-line loads with radial tooling, like cross drilling or whatever. The only time it will rotate unintentionally is from radial torque loads, such as fixed turning tools. I have to make the turret indexing shaft rigid enough to not deflect much from these torques. Anything over 1/2 inch shaft with 1 inch overhang should be good, but I haven't done the analysis yet. Just a gut feeling.


The sliding headstock is for future compatibility with a guide bushing. This is how Swiss type lathes work, since a guide bushing lets you make ridiculously long length/diameter ratio parts without getting any extra deflection from cutting. Keep in mind that having the headstock on an axis that is bolted directly to the machine base and having a similar setup for the cross slide actually makes for a more rigid machine, since there is no compounding deflection from having axis on axis on axis on axis on axis, ect.

This design minimizes axis stacking.

The first picture doesn't show fixed gang tooling, or the subspindle, or a bunch of other stuff I mentioned I want. I'm trying to nail the initial machine construction since the more complex features rely on a solid foundation. All of this while I keep the parts simple enough to make with my own cheap ass hobbyist tools on a low budget!
Check out the Tri hotend!

User avatar
Captain Starfish
Printmaster!
Posts: 960
Joined: Tue Dec 10, 2013 5:24 am

Re: DIY Mill Turn Center

Postby Captain Starfish » Sat Oct 31, 2015 10:00 pm

Someone's been doing his homework. I'm liking it.

I'll be interested to see how you manage the axial vs radial live tooling with an index via the geneva mechanism - the video you showed did a couple of facing operations on an arbitrary angle - and if you find a cheap servo that can be spun up fast enough to effectively run an endmill let us know, there's a reason most of the cheaper units have a clutch between a spindle motor and a servo or stepper...

User avatar
0110-m-p
Printmaster!
Posts: 456
Joined: Sun Oct 20, 2013 9:23 am
Location: Atlanta, GA

Re: DIY Mill Turn Center

Postby 0110-m-p » Sun Nov 01, 2015 12:08 am

Generic Default wrote: There is no trend towards this, either, and none of the current 3d printing technologies seem capable of it.


I've always wondered about the capabilities of this machine for round parts placed directly in the center...

http://polar3d.com/
Current Machines || Rostock Max (V1) | V3DR ||
Previous Machines || Flashforge Creator Pro ||

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Mon Nov 02, 2015 2:04 pm

Some of the bigger parts of the machine need to be made from aluminum flat bar. The linear motion parts have to be aligned and machined without re-clamping, since they have to be accurate to within 0.001 inch over 12 inches or so. The surfaces need to be very very flat.

Since I have a TAIG mill, I can only machine stuff in a small volume. The vise I've been using so far limits me to 3 inch flat bar, and I can't machine surfaces more than a few inches away from the vise without bad surface finishes or losing grip.

My solution is to make two custom low profile aluminum vises, bolt them down to the mill bed when they're somewhat aligned, then mill them in situ so that all faces and surfaces are parallel and perfectly flat to the mill. This means that every surface on the vise will be aligned perfectly with the mill axes; the vises will actually be cut in their final position.

Using this method I can get vise alignment to 0.0005" or less across the entire milling area. With two vises, I can put bar stock in and machine an area the size of the mill bed itself.
The whole point of doing this is so that I can cut the precision sliding surfaces of the mill turn center; the entire X axis and Z axis will be machined from one piece of aluminum flat bar with dovetails and stuff.
Plus, I'll finally be able to do 4" wide parts!

Making the vise body;

fly cutting 1.jpg

drillingvise 1.jpg


A semi-finished vise. I need to make a second one still, then bolt them to the mill bed and finish them in place.
vise 1.jpg



My lathe sucks. The tailstock is misaligned by about a millimeter with the headstock, which has been screwing me up for over a year now. I emailed their customer service and got a replacement tailstock, but it was just as misaligned as the original. So much for USA quality.

Anyway, I printed a tailstock slide and a tailstock mounting part from PLA. I can put different attachments on it, like live centers and drills and stuff. This one can be aligned concentric with the headstock. The purpose is so that I can machine the lathe spindle when I get to it eventually. You can see the big steel bar in the chuck. I can't turn the spindle without a tailstock, since the chuck has very weak gripping power. I need to turn the spindle so that the mating surfaces are within 0.001" of the bearing ID, then I'll press them in place and use thermal expansion to lock the spindle into the bearings. I don't have an arbor to press with so the spindle must be as close as possible to the bearing size, but it can't be any smaller.

lathestock1.jpg



And finally a picture of a finished spindle bearing block. I'll have to make another one for the back of the spindle. Actually two more since I overbored this one a bit. These are the most precisely made parts on the machine, and the most difficult for me to make. Most lathes use a one piece design for the headstock but I can't remove enough material to make that happen. I tried printing a one piece headstock but it failed at 15 hours in an 18 hour print.
headstock block 2.jpg




Hopefully all of this isn't too confusing for people reading it, I know it's different from 3d printer stuff. It will make more sense once you see how the parts fit together.
Check out the Tri hotend!

User avatar
Captain Starfish
Printmaster!
Posts: 960
Joined: Tue Dec 10, 2013 5:24 am

Re: DIY Mill Turn Center

Postby Captain Starfish » Mon Nov 02, 2015 5:35 pm

Nice work - I love pictures of bright fresh cut metal :)

User avatar
Eaglezsoar
ULTIMATE 3D JEDDI
Posts: 7185
Joined: Sun Apr 01, 2012 5:26 pm

Re: DIY Mill Turn Center

Postby Eaglezsoar » Tue Nov 03, 2015 3:41 pm

I agree, very nice work!

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Mon Nov 09, 2015 3:57 am

I was hoping to get stuff done this weekend but the cheap PSU wired to my CNC machines blew up. Not sure why exactly. I need to rewire all my things and fix my mill so it works reliably. Hopefully the smoothieboards didn't get fried when the PSU blew up, otherwise I'm screwed.

Anyway, I got a few pictures of the machine base. I made these parts before I had an edge finder so I was limited in my design.

headstock slide rough 1.jpg

headstock slide rough 2.jpg

headstock slide rough 3.jpg


I've already redesigned the dovetail guide. What you see in the pictures above was limited by part size and clamping capability, as well as a lack of indicating tools and cutting tools.

The base will mostly stay the same, but the dovetail slide will be made from a single 1x3x12 flat bar of aluminum. Everything needs to be accurate to within 0.001, so the entire slide will be milled without any re-clamping. It's more material removal and I can't make mistakes on it, so I need to get my CNC mill working reliably and error free before I start any cuts. It will probably be a few weeks away since I'm so busy and I need to fix and improve stuff before I can even start!

I plan on lining the linear slide surfaces will something that has very low friction and excellent wear resistance. 0.001 Kapton tape rips, so does PTFE tape. My next try will be UHMWPE tape that is nearly 0.03" thick! If that doesn't work I'll just make thin plates of a wear resistant plastic and somehow fasten them to the sliding surfaces. The idea is to keep low friction and make a linear slide that lasts forever and is still rigid and accurate.


Side note, I found out that 2011 aluminum round bar is round to within 0.0002" at a 1 inch diameter. It's good enough to use as a bearing shaft before any machining!
Check out the Tri hotend!

stonewater
Printmaster!
Posts: 345
Joined: Mon Jan 06, 2014 1:24 am

Re: DIY Mill Turn Center

Postby stonewater » Mon Nov 09, 2015 11:25 am

that is a sherline 4000 right? with their CNC package you are using. nice! I have the mill and the lathe no CNC though. makes good small parts. sorry to hear about your tailstock being off. there will be a lot of people interested in your design.

Tom C
MAX V1
325 MM carbon arms trick trucks effector mount LED ring heat spreader
Corvair750
V6 Hotend
Robo 3D
Flashforge creator
http://WWW.TeslagenX.com" onclick="window.open(this.href);return false; - Bedini experimenter kits, books, DVD's

JFettig
Printmaster!
Posts: 821
Joined: Tue Nov 18, 2014 4:39 pm
Location: Minnesota

Re: DIY Mill Turn Center

Postby JFettig » Mon Nov 09, 2015 11:33 am

Nice work,

Consider finding yourself a couple sets of HSR or SHS THK rails from Ebay(the used ones that come out of S. Korea are decent) This way everything will move very smoothly, it'll be rigid and be quite accurate. 15mm rails should be plenty for this machine. If you mount those on a plate of Mic6, you can get them very accurate. Indicate rail to rail while using a mag base on a carriage.

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Mon Nov 09, 2015 6:38 pm

The lathe is a 4400 from sherline. I made custom stepper motor mounts and couplers, that was before I even had any measuring tools other than calipers so the mounts kinda suck. The lathe has been a PITA since day 1, and I realized early on that the tolerances and quality that go into making a lathe like it are nonexistent. I guess I just got a tailstock from a bad batch.

The Taig mill is great, I'm using a smoothieboard with the CNC ready one. It's very accurate and adjustable, unlike the Sherline. The Sherline lathe gave me an odd insight which is why I'm making this custom machine now, which is that even a terribly inaccurate and misaligned lathe with a weak motor and flexible frame can still turn steel accurate to a few thousandths!



Jfettig,

I've been considering using pre-made linear rails from the beginning. I might end up bolting linear rails to an 80/20 extrusion and shimming with aluminum foil, since foil is around 14 microns thick and T slot extrusions are very straight for how cheap they are. Pre-made rails would save me time but would be expensive, and I'm trying to avoid using regular steel on this lathe because it rusts and rust sucks. I kind of wanted to make a custom linear slide to see if it is viable on small machines.

My current lathe base is just a solid flat bar of 6061. To make everything more rigid and eliminate vibration, I'll probably end up filling a rectangular aluminum tube with granite and epoxy, then facing the surfaces flat and bolting the linear components to it. Using aluminum instead of hot rolled or cold rolled steel prevents thermal warping over time, plus the entire frame will be the same temperature rather than having gradients that flex it.

None of this is critical to getting a lathe that cuts metal, it's only for metrology purposes. I'll do some finite element analysis and post pictures here so you guys can get a sense of stiffness in machine tools, something that is nonexistent in 3d printers (unfortunately).


Also, absolute linear accuracy isn't as important on this machine as you might think. Most of the parts it will make will be less than 2 inches in diameter and less than 4 inches long. My plan is to use 1 inch ER collets and turn as much stuff from round bar through the spindle as possible. The small part dimensions allow the use of linear components that aren't perfectly straight or perpendicular, since cuts will only be made over short distances. I don't need accuracy to microns on this machine, I'm looking to hold around 0.0015 tolerances on most parts.
Check out the Tri hotend!

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1719
Joined: Tue May 14, 2013 12:52 pm

Re: DIY Mill Turn Center

Postby 626Pilot » Thu Dec 31, 2015 11:37 pm

This is fascinating. I've been thinking about building one of the OpenBuilds lathes, but I don't think any of them have complete plans, or have even been finished by their creators. What toolchain are you planning to use?

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Fri Jan 01, 2016 1:15 am

I just got access to a camera again so I'll post some pictures soon. I didn't even know about that openbuilds website until a few minutes ago!

I'm planning on using a smoothieboard to control it, either the 5 axis one or perhaps the new smoothie V2 pro when it comes out. For full automation, my machine is up to 9 stepper motors now. Lawd!

Most of the gcode can be written by hand or with turning CNC software. The machine will have an LCD running smoothieware, probably using repetier host or whatever to control it from a PC. Gcode for turning and lathe operations is a lot more simple than 3d printer gcode.


I recently took a class in embedded systems and I learned a ton, and I feel like I have the skills to do some serious firmware re-writing to make a custom operating system for the mill turn center. Of course I would start with something that works, since the lathe is controlled exactly like a 3d printer and can work without modifications. But to operate all 9 axes in the way I want, I will have to make some changes.....


I'll be disclosing a lot more in a few weeks, tomorrow I'll get some basic pictures of my progress. This machine is turning into a beast.
Check out the Tri hotend!

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Sun Jan 03, 2016 5:53 pm

OK so first for the machining setup.

To make ultra precise parts for this mill turn center, I had to upgrade my Taig mill so it can make big enough parts. So I made two aluminum vises. They have a fixed jaw, a fixed screw thrust block, and a sliding jaw that gets bolted down after being pressed against the workpiece by the thrust screws. Pictures make more sense.

IMG_1521.JPG

IMG_1522.JPG


To get the level of precision I needed, I milled the vises as separate blocks, then assembled them, then bolted them down to the mill table. Once they were somewhat aligned on the mill table to within a thousandth or two, I milled all mating faces flat in situ. This means that the surfaces in contact with the workpiece were milled flat and level and square relative to the mill table itself, which theoretically gives perfect alignment. So I know everthing is aligned as accurately as the mill itself, and the mill is pretty accurate!


One of the largest and most precise pieces of the mill turn center is the Z axis linear ways, which the headstocks slide on. They have to be extremely accurate, smooth, flat, ect. They are milled in two setups from 1x3x12 aluminum flat bar. The first setup is just to face one side perfectly flat. The second setup is to place the flat side face down on parallels in the mill vise[s], then to machine all of the holes and mating surfaces without re-fixturing. This guarantees parallelism within a few ten thousandths across the entire part. The 60 degree dovetails are perfectly aligned with the top sliding surfaces of the rail.

IMG_1512.JPG

IMG_1513.JPG

IMG_1514.JPG


Unfortunately I messed up a bit on this piece. I made it a day before I bought an air compressor to clean stuff. When I flipped the part there must have been a tiny spec of metal between the parallel and the flat milled surface, which caused it to not sit flat before machining. As a result the whole part is only accurate to a couple of thousandths instead of ten thousandths. The top side mating surfaces are still parallel and stuff, but I'll have to re-flycut the bottom surface relative to the top to fix the problem. It is still a working part as of now, and it's still accurate to 2 parts per 12,000. I checked the corners with a micrometer to figure this out. The spec was under one of the corners between the face and the parallel. Metrology stuff here!

I'll show more stuff in the next post.
Check out the Tri hotend!

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Sun Jan 03, 2016 6:34 pm

Now onto the linear motion!

Ballscrews have been the norm in precision CNC machines for a few decades. The goal is to get linear motion with excellent stiffness and repeatability to within a few ten thousandths of an inch or less.

Just for reference to everyone using the metric system who's reading this.....

0.001 inch = 25.4 microns = 0.0254 millimeter
About the smallest thing you can see by eye.


So for my mill turn center, I need an ultra precise linear motion system. I can't spend hundreds or thousands of dollars on precision ball screws, or ball nuts. I have like 6 linear motion axes on my machine right now. It would be too expensive!


This is what I came up with;
IMG_1517.JPG

IMG_1520.JPG


SIDE VIEW
Linear Positioning Setup.jpg


TOP VIEW
Linear Positioning Setup Top Cross.jpg


It's a 0.375" fast travel precision ACME leadscrew, passed through two radial ball bearings in a thrust block, locked in place by two 0.375 clamping shaft collars. Before locking, the assembly is preloaded for stiffness by tightening a nut onto one of the shaft collars. The ACME leadscrew has a pitch of 1/6 inch. So each revolution pushes the linear slide 0.166666 inches. With 200 steps per revolution and 16x microstepping, the machine will have a positioning resolution of 0.000052 inches, about half of a ten-thousandth of an inch, or just over one micron. When turning cylindrical parts, the diameter step size would be 1 ten thousandth of an inch. Well within my goals for precision!


The most important thing to the setup is the nut itself. It does not have any moving or adjustable parts. Backlash is zero. By printing the nut vertically with extremely close thread fits, and using a long nut length, backlash can be reduced to zero. I have tested this a bit with many nut designs, and right now I have a damn good one! The nut is around 4 inches long and has many thread faces in contact with the leadscrew at any time. Backlash is less than 10 microns (at worst). With grease and a bit of thermoplastic creep, the friction drops to near zero but no gaps develop in the nut. I'll see how well this design holds up to constant, high load usage. It can't run nearly as fast as ballscrews at any real load, but in slow axis travel it is excellent!

It's a good fit for my machine, since the slides will not travel more than about 7 inches and will only do so at low speeds and moderate loads. The cost per linear slide is in the range of 10 dollars, excluding electronics.


Also, I've been thinking about this quite a bit....Is there any reason at all to keep the screw ball bearings in a tight fitting bore on the thrust block? As long as they can't move axially with the screw, they will work fine, right? I'm thinking about oversizing the thrust block bore so the bearings float radially for easier alignment. They would still clamp to the flat surfaces once preloaded, but it would make assembly a lot easier, and machining the blocks with be cheaper too.

NOTE: The thrust blocks here are just prototypes, the real ones will be machined from aluminum with larger flats for the bearing faces. They are bolted to the machine with two 1/4-20 socket head cap screws so they can take large thrust loads.
Check out the Tri hotend!

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Sun Jan 03, 2016 7:06 pm

Three posts for today should be enough. I'll machine some more stuff later so I can show you a working sub-assembly. For now, its onto the spindle workholding!


I decided long ago that this machine will be a mix of a mill turn center and swiss type lathe. Many parts are around or under and inch (~25.4mm) in diameter, so I kept that in mind when designing stuff.

This machine's primary method of holding work will be ER-40 collets on two headstocks, each adjustable around the axis of rotation of the spindles. ER-40's hold work pieces up to one inch in diameter. With oversized collets, they can hold round bars up to 1.1875 inches in diameter. My spindle will limit the size to 1.125 inches, or 1-1/8th round bar, which is easy to get. Metric sizes in this range work too. ER-40 collets are mass produced and extremely accurate (0.0003" for ~$22 each collet). By sticking with two 1.125 ER-40 collets (one in each headstock), any part at or under 1.125" (28mm) diameter can be machined.

This is what the collets look like. Keep in mind real ones are precision ground steel. All of the parts will be metal, I just have prototypes for now to check fits and sizes and stuff.

IMG_1524.JPG

IMG_1525.JPG


The two pictures below are not my machine, they are real ER-40 collets so you get an idea of what they are;
Image
Image


The tool in the picture has a 0.375" shank, the aluminum bar is 1 inch diameter. You can see how these collets hold stuff; the machine will be able to transfer work between opposing headstocks so parts can be completely machined in one setup without re-fixturing. It will all make more sense later when I show you the entire machine, but that's a secret for now!


Oh yeah, the main headstock spindles will have a backplate that all this stuff gets bolted to. So while ER-40 collets are the main method, the machine will also accept 6 inch chucks for much larger workpieces.
Check out the Tri hotend!

User avatar
DavidF
Printmaster!
Posts: 778
Joined: Thu Aug 29, 2013 12:12 pm
Location: Delaware

Re: DIY Mill Turn Center

Postby DavidF » Mon Jan 04, 2016 11:23 pm

neat project. Ill be following along
http://www.youtube.com/user/aonemarine
Lost pla castings? see me

JFettig
Printmaster!
Posts: 821
Joined: Tue Nov 18, 2014 4:39 pm
Location: Minnesota

Re: DIY Mill Turn Center

Postby JFettig » Tue Jan 05, 2016 7:55 am

ER40's are awesome for work holding. I'm planning to get an adapter for the lathe at the shop. You should also consider 5c. You can buy a spin indexer for $40-60 and steal the spindle from that. It has a 45mm OD on the back side, a few modifications, $150 in quality bearings and you'll have a super nice spindle! I have mine all modeled up in Solidworks if you need a copy of the spindle. I'm looking at turning mine into a 4th axis indexer for my CNC mill.


Jon
Attachments
spindexer spindle.PNG

User avatar
Captain Starfish
Printmaster!
Posts: 960
Joined: Tue Dec 10, 2013 5:24 am

Re: DIY Mill Turn Center

Postby Captain Starfish » Tue Jan 05, 2016 3:16 pm

What you're talking about with the use of ACME threads in radially floating arrangement with a non-adjustable nut (as opposed to preloaded halfnuts) goes against pretty much every admittedly limited thing I've learned about CNC motion control design for repeatable accuracy and precision.

I'd be careful of designing in radial slop. Check your trigonometry for starters, what does a millimetre or two (for example) end deflection on the opposite side of a right triangle with an adjacent side half the length of your travel turn into on the hypotenuse? What are the limits of slop there before the error becomes significant? Then look at the kinetics of it. Once a lead screw starts whipping around on rapids or under load, radial play in the thrust block will transfer into oscillating radial loads into the stepper spindle and carriage nut - significantly reducing the life of both.

Depending on the materials, feeds and speeds you're working, I'm going to be interested to see how long it lasts. Aluminium is IMO too soft for the ways and the plastic carriage nut I would have thought would flog out pretty quickly.

I will be very interested to see how it all works out - and really, REALLY hope I'm wrong in my initial guesses. I hope it does work well for you. Coz then I might be tempted to try some of this myself instead of going down the well travelled path of buying Hiwin rails, ball screws etc to make my next machine at some point. :D

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Tue Jan 05, 2016 3:59 pm

Thanks for the quick and thorough feedback guys.

JFettig,

I thought a lot about the workholding stuff. 5-C collets and other types like it do not have slits in both directions. To clamp them you pull them back into a tapered tube, which squeezes the front on the tool or workpiece. ER collets will clamp evenly along the entire length of the workpiece, which is much stiffer when you start putting moment/torque far out on the workpiece. Maybe I'll switch to 5-C if the ER system doesn't work out. I do have a spindle design already, but it's nice knowing that they can sell a spindle at retail for $50. If I make a bunch of these, I think I can get my spindle cost around that point. Mine has to be able to pass a 1.13" round bar through the bore.


Starfish,

I know it goes against normal machine design. My leadscrews are 0.375 diameter and only a foot long. The nut itself is 4 inches long, and I might upgrade that to 6 inches to take up the entire length of the carriage when it's bolted to it. My earlier designs had back-to-back locking nuts to eliminate backlash. I realized that a very smooth, burnished, fast travel leadscrew sliding against printed thread flanks has almost no wear compared to other systems. The ridiculously long nut has so many threads in contact that there is no axial slop, and the load is distributed across a very large area. I think this will give the nut a very long service life. They have no backlash as of now, my main concern is if they will wear out quickly.



The thrust blocks, on the other hand......
Radial play in the bearings is [nearly] eliminated by preloading them. It makes them stiffer, too. I noticed that on my Taig mill, the leadscrews are 15 inches long but can rotate with no visible wobble on the other side of the table, which is like 16 inches away. The bearing faces are ground flat, and once they are preloaded against each other, they will hold in position. Allowing a radial clearance in the thrust block is simply for easier assembly and alignment, since it prevents binding from misalignment near the thrust block. I'll test it soon. If it doesn't work out I'll go back to the normal way of fitting the bearings into precise holes.

Since I'm using fast travel leadscrews and direct drive steppers, there will be a maximum force that the linear axes can generate. They will never spin the leadscrews over 600 RPM and the thrust force will probably be below 30 kilograms peak. Most cutting doesn't take that much force, as I've learned from my mini machine tools.

I do know about the normal methods of doing things in CNC with ballscrews and linear ball slides rather than doveways and acme leadscrews. I've been studying this stuff for a few years now, and the stuff I'm showing you here is my way of reducing the cost by a factor of 10 or more. I know that even with a terribly designed, flexible, cheaply made lathe like my Sherline lathe, you can still cut steel rods to tolerances of about 0.001". So by staying cheap but fixing all of the design flaws, I should be able to make a small lathe cheaply without increasing the cost that much.


I'll show you my spindle design and carriage stuff later today when I have time again!
Check out the Tri hotend!

JFettig
Printmaster!
Posts: 821
Joined: Tue Nov 18, 2014 4:39 pm
Location: Minnesota

Re: DIY Mill Turn Center

Postby JFettig » Tue Jan 05, 2016 5:29 pm

I hear that but you don't really have that problem unless you're trying to hold onto something significantly smaller than the collet - I was just milling some parts that were .740 diameter in a .750 collet the other day and one did pull up but only because the part deformed(it was slotted hollowed out part), I plugged the thru hole with a pin and it held great. The nice part about 5C is there is a wide range of collets, every 1/64" is available and are used in so many applications. Not trying to push it on you but its something to consider. Are you making your spindles? I only bring it up because you can use these things as accurate, ready made spindles. Its difficult to get an accurate spindle turning it yourself. There is a place in UK that makes these spin indexers that take both 5C and ER32 collets, they're kind of neat as well.

User avatar
Generic Default
Printmaster!
Posts: 558
Joined: Mon Jun 03, 2013 6:56 pm
Contact:

Re: DIY Mill Turn Center

Postby Generic Default » Tue Jan 05, 2016 10:32 pm

I am planning on making my own spindle, but I know it's going to be difficult to get the tolerances close enough on the bearing fit. It will probably be the hardest part for me to make on my tiny machine, even though it's simple. I'll keep your recommendation of using the indexer spindle as a back up plan.

5-C collets would work in this machine no doubt, but ER-40 has the advantage of being able to hold parts in a larger size range for each collet. They have a wider gripping range for nominal sizes, which may be useful for oversized stock or undersized stock or values where metric and imperial are close, like 25.4mm in a 1 inch bar. I think 5-C collets would be easy to integrate into this design, there would only be a 2 parts that has to be changed out on the end of the spindle.



I've been trying to come up with a solution for limit switches until now. On a 3d printer, microswitches work fine because they're repeatable to around 0.04 mm. That's not accurate enough on a lathe, since you control the radius and the diameter is twice as big. So 0.04mm repeatability on a turning tool will make 0.08mm diameter tolerances.

This is what I came up with, I haven't made it yet but I think it will work nicely;
Ball Limit Switch.jpg


It's a small non-conductive tube with a spring loaded chrome steel ball held in by a metal screw plug. The metal screw plug will be the positive terminal that the limit switch wire connects to. It is electrically connected to the steel ball. An anode plate, perhaps a simple brass flat surface, is rigidly mounted on the linear carriage. When the carriage gets to the end of its travel, the steel ball contacts the brass plate and closes the circuit.

The ball has a small amount of travel, around 0.1 inch, so the carriage can decelerate. Otherwise it crashes into the non-conductive tube and stops, but the limit switch is still closed.


What do you think of it? I figure a constrained ball-in-tube design with hard surfaces and rigid mounting will give repeatability to within a few ten thousandths at most. It should be quite easy to make on either a lathe or a 3d printer.
Check out the Tri hotend!

User avatar
Captain Starfish
Printmaster!
Posts: 960
Joined: Tue Dec 10, 2013 5:24 am

Re: DIY Mill Turn Center

Postby Captain Starfish » Tue Jan 05, 2016 11:59 pm

GD, my hat off to you for giving this a crack and I hope it works. You've obviously thought it out and done your research, and are NOT the usual numpty that says "Harrr harr I can do this for heaps less by using der cheap bits and it will be super accurate and faster and will cut inconel no worries I dun noe why manufactuers doan doit this way they is so dumb durr hurr". Apologies if I implied this might be the case, it was certainly not my intent.

Like I said, I'll be watching updates with interest because even if you can get it repeatable and usable down to a thou, at the kinda price points you're talking it (or something like it) becomes an interesting and affordable project for me to try myself.

I like the switch idea, looks very similar in concept to the three ball touch probes which can be damned accurate and repeatable. Only thing might be to keep an eye on ways of deflecting swarf away from the ball contact face (and matching plate) or, worse, getting stuck between the tube and ball.

User avatar
Eaglezsoar
ULTIMATE 3D JEDDI
Posts: 7185
Joined: Sun Apr 01, 2012 5:26 pm

Re: DIY Mill Turn Center

Postby Eaglezsoar » Wed Jan 06, 2016 11:27 am

I think it is an excellent idea that will work well and cost little to create.
Great thinking above the board!


Return to “What are YOU making?”

Who is online

Users browsing this forum: No registered users and 2 guests