[JMcDonald]

I have a question:

I am in the process of building a CNC mill, and was thinking about building a rotary table to go along with it. I would build it as direct drive (basically like a lathe head assembly), using a stepper motor mounted parallel to the spindle axis. I would do it likely this strictly because I dont know how Id be able to get rid of all the backlash using a worm drive. I would probably gear it down a reasonable mount - maybe like 3:1 or 4:1 (maybe higher, depending on the sizes of belt pulleys I can find). That way, with 1.8 degree stepper motors, I could get accuracy down to .6 degrees or less (with a 3:1 drive at minimum), which I think is acceptable for what I need to do. Not to mention I would just get that much more torque out of them, which would be important for trying to do lathing operations.

Here are my questions:

1) I was thinking that if i mounted the rotary table vertically, and put stock into it, that I could rotate the stock at some given constant speed (probably something slow, like 5-10 RPM?), and use an end mill in the head (with the spindle on, just like if I were milling) to mill the surface of the stock as it is rotating. How well would this work as a method of "turning?" I could still get a pretty smooth finish with it, right?

2.1a) Using the same rotary table, could I use a regular tap or die to thread a piece of stock? Like, imagine that piece of stock turning at again some slow speed, and being fed through the die (or onto the tap)? I suppose the tap / die would either have to be mounted to the head, or to something else not connected to the table, so that the table could move the stock through / over it as it rotates. The stepper would frequently backturn to clear the shavings from the die to prevent binding. Would this work?

2.1b) Doing the same thing as above, except reversing the placements, so that the rotary table is mounted OFF the mill table, and then the tap / die is mounted ON the table. That way the motion of the tool is controlled. This simply gives more options for mounting the stock and tool in relation to eachother (for example, it might work better one way than another).

2.2) What about if I used the same method as lathes used, to cut the thread? I dont know how fast steppers can spin, but Id imagine its not very fast? Would that be too slow to successfully cut threads?


Here are my predictions:
1: It should be possible and pretty strait-forward. Especially if I use a tailstock or a steady-rest (which I plan doing).

2.1: Again, it seems possible as long as the stepper is strong enough (???) and I can find sturdy mounting points for the off-table components.

2.2: Doubtful, as the stepper is likely either not strong enough or not fast enough.


*edit*

After a few minutes looking through sdp-si.com's online catalog, it looks like I should be able to do a ~6:1 drive ratio. I might find something better locally or somewhere else, also.

[S_J_H]

Having a rotary table and experimenting with a real lathe headstock on the mill I can offer a little advice.

First a rotary table doing a lathe op with a spinning endmill for the cutter will not get you a lathe type finish.
It would be better to just mount the stock in the spindle and mount a lathe tool in a vise or holder on the table. You are limited in size but it works very well.

You can cut threads with a rotary table and a stationary single point lathe tool mounted on the z axis. You adjust the rotary table to a constant speed and then do the math and advance the table at the needed rate into the tool bit mounted on the z axis. Takes a little trickery to get the rotary table running constantly.
But you can also thread mill to make threads using helical interpolation.

If you plan to make a direct drive rotary table it will need a brake to hold it in place for indexing. Unless you are talking about very tiny threads a stepper will have a rough time tapping. Tapping takes a lot of power.
Best of luck!
Steve

[JMcDonald]

Thanks!

I most of the stuff Id need to turn will be too long to put in the head spindle (and most things too large also).

However, I think I will try option one and see how the finish comes out. If it comes out inadequate for the application, then Ill look into building / buying a lathe headstock and tailstock.

BTW, in option one, lets say I was using a 1/2" end mill. For a "turning" operation, I would slowly rotate the stock, and push the end mill down into the stock until I get it down to the required diameter. Then, the end mill would lift up, and the table would move 1/2" over (or maybe just a little less), and repeat. I dont know how smooth the finish would be doing it this way, but I do know that if I tried to pull the end mill along the length of stock for each step of rotation, Id end up with a bunch of thin, flat surfaces (well, 600 to be exact, with a 3:1 stepper reduction and .6 degrees per step) parallel to the center axis of the stock. At least using the first method (basically cutting many 1/2" strips off the surface, making my way down the stock), I know that at least each 1/2" section down the stock would be almost perfectly circular.

That is just to clarify the way I am thinking of this.

Thanks for the help! Reading your thread is what got me thinking all this lathe business in the first place, heh.

[DareBee]

You will need a massively powerful stepper (on the rotary) to overcome the milling forces. This is why rotaries tend to have a large gearbox.
The backlash concerns you mentioned is why off-the-shelf rotaries are 10k - 50k dollars.

[JMcDonald]

Hmm. Maybe Ill try a multi-stage reduction setup to get more speed reduction and torque multiplication. That means with two sets of those same gear sizes, I could be looking at a 36:1 reduction. With three sets, I get 216:1... that is probably a little extreme, but it does still show that it would definately be possible.


And, my alternative would be something like this:
http://www.littlemachineshop.com/pro...1810&category=

And btw, do you mean that the rotary table without a large reduction (like those of typical drives) would not easily be able to turn the stock against the end mill?


Hmm... I think I might build two indentical headstocks. One would be heavily reduced (at least 36:1 like I mentioned above, but I will probably try to match the 72:1 that seems to be common for RTs) one, driven by a stepper. The other would be driven by the same type of 2.5HP motor that is going into my mill head. Each would be built so that the belts would be easily disconnectable from the motor (swiveling motor mounts most likely). That way each could be used as a tailstock for the other. We'll see, I suppose!

[Chris64]

I've been thinking about the same idea. I've pretty much given in to the idea that I would need to build it with two motors and a transmission. One to direct drive (via a DC motor, maybe 1/2-3/4 HP) for turning (with a rotation sensor for threading), and the other, either a stepper or a servo on a worm drive for indexing. Switching gears could be done with some type of pressure plate between the two gear options. This would be nice because there wouldn't need to be any type of syncro needed for the gears to mesh.

Backlash with a worm drive is a tough problem. I figured you could even use a planetary gear setup if it was geared low enough. For indexing tasks you would only need around 60 rpm tops...so you could be in the neighborhood of around 50:1 gear reduction. Needless to say that it would have some pretty good holding strength with that type of reduction. Some of the Servo motors I found browsing around appeared to have around 20 lb/in holding power. I assume that this is it's ability to resist turning (@ 1 inch). With a 50:1 that would be around 1000 lb. Of course if you're on a mill table with 8" of Y travel that will definitely change things (up to 4 inches from center) assuming you're holding a large work piece (i don't know the math - I'm not even sure if the math I already figured is right). It still seems like it would be possible.

Ideally I'd like to make it capable of hold a peice strong enough that I wouldn't have to keep turning the parts over manually to make something...and also be able to turn/thread. Basically, make something that fits somewhere between a good 3 in 1 machine and a full blown turning center.

I have no experiences with anything like this. I figured I'd work on collecting the expensive parts where ever I can find them and design something from there. Keep us posted if you try anything.

[cota]

do rotary tables use ball bearings or plain bearings?
can they handle the rpm's to do turning without wearing too fast?
seems they are designed for very slow spinning.

[philbur]

A cost effective solution might be to buy a rotary table with backlash adjustment. From memory you can get one for around USD150 - 200. I converted an 8" Vertex RT to CNC using a 640 oz-in stepper. I've used it to engrave 1/8" numbers on the circumference on a 3" steel disc with no backlash issues at all (see picture). Of course you can't have a the large gear reduction of an RT and still have high rpm, so a separate, simple direct drive solution, with a bloody big steppper or better still a servo, might be the way to go for the "lathe" function.

Phil

[project5k]

didnt hoss pretty well perfect this 2 motor system? one fast motor for "lathing" and a stepper for rotary?

[philbur]

His web site shows the lathe arrangement but I can't see that he has yet solved the rotary table function. His stated plan is to use a stepper so I think he will need a large speed reduction, so his simplest solution might still be to put a stepper on and existing rotary table.

Phil

didnt hoss pretty well perfect this 2 motor system? one fast motor for "lathing" and a stepper for rotary?

[project5k]

i definately remember seeing pictures of him using the lathe head as a 4th axis, using a stepper, infact i just had a pm conversation with him about what motor and gearing he used. it was a 200 step/rev motor 1/8th stepped and 1.8:1 or so gearing...

maby i misunderstood what i read, its not a rotary table, or "c" axis, its an "A" axis..

[JMcDonald]

it was a 200 step/rev motor 1/8th stepped and 1.8:1 or so gearing...

Really? Thats all the reduction he used?

[digits]

What I have been thinking about doing is to combine both a stepper driven rotary table and a lathe headstock into one. I would actually mount a lathe spindle on top of the rotary table, and drive it with a dc motor using a shaft running through the centre of the rotary table. For indexing operations, I would lock the headstock to the rotary table, and then rotate it using a stepper. Backlash would me measured and corrected via an encoder mounted to the lathe drive shaft - that way it'd work for both lathe and RT ops.

[JMcDonald]

So, how much do I REALLY need to gear the stepper down?

Lets say I use 6:1 gearing, and am trying to work a 1" shaft (bad choice of words...)

425 oz-in * (1 / 16) = 26.6 lb-in
26.6 lb-in / 12in = 2.21 lb-ft
2.21 lb-ft * 6:1 = 13.28 lb-ft

> 0.5" radius / 12" = 24

13.28 lb-ft * 24 = ~319 lbs

Is that really not enugh?

[project5k]

that sounds like a bunch of torque to me... i cant see needing that much.. when i did my math i figured a 200 oz in motor, geared 6:1... your not gonna be turning it all that fast, so you'll be pretty close to the holding torque of the motor, (the slower you go the stronger) when i did my math i was more interested in the # of divisions i could get.. 200 steps/rev 1/2 stepped, * 6:1 gearing gives me 2400 divisions per rotation on the work peice. thats .15 deg per division... if thats not small enough divisions, then i really need to rethink what i'm trying to make...

[JMcDonald]

Yeah. Im also wondering if that would be enough to thread (thread a shaft through a die)! If I were going to do that, Id probably use a pretty high gear ratio to try to get at least a 10:1, which would give me somewhere under 500 lbs of force at the surface of a 1" shaft. I dont think it would take near 500lbs of force to thread a 1" shaft... It would take twice as much force as a 1/2" shaft would (twice the circumference), and you can do that by hand pretty easily (which Id guess is like 20-40lbs of force, depending on how much leverage you are putting into it).

[kuhncw]

Steve, you mentioned in your 5 Oct post that:

"Takes a little trickery to get the rotary table running constantly."

Would you elaborate a bit on how you did this?

Thanks,

Chuck Kuhn

[philbur]

Don't calc to lbs force stick with 13.28 lb-ft. Take a 1 inch die/tap and simply measure how many ft-lbs it takes to thread. Rememeber that your calc is based on the holding torque.

In any case if you are mainly interested in threading/tapping the 90:1 rotary table at 5 rpm would work perfectly.

I think if you want "proper lathe" performance you need at least 2,000 rpm and not less than 0.5 hp at the spindle. He presto, you already have a spindle on the mill that will do that, but not with position control.

Phil

So, how much do I REALLY need to gear the stepper down?

Lets say I use 6:1 gearing, and am trying to work a 1" shaft (bad choice of words...)

425 oz-in * (1 / 16) = 26.6 lb-in
26.6 lb-in / 12in = 2.21 lb-ft
2.21 lb-ft * 6:1 = 13.28 lb-ft

> 0.5" radius / 12" = 24

13.28 lb-ft * 24 = ~319 lbs

Is that really not enugh?

[JMcDonald]

In any case if you are mainly interested in threading/tapping the 90:1 rotary table at 5 rpm would work perfectly.

Hmm... sounds like a pain, heh.

I think if you want "proper lathe" performance you need at least 2,000 rpm and not less than 0.5 hp at the spindle. He presto, you already have a spindle on the mill that will do that, but not with position control.

I am building two identical headstocks. This is so they are both the same height. One will use the same 2.5HP motor that my mill uses (for lathing), and the other will use a reduced stepper. I am doing that so when using one, the other can be used as a tailstock. With my table I should have about 18" between them, or more if I try to stick the headstocks off the table a bit.

Although maybe I wont even worry about trying to thread stuff with the rotary table. I could probably more easily use the lathe headstock to thread stuff, or just do it by hand if I really needed to. I don't really have that many threading operations i need to do, so I dont think the extra hassle of compound gearing the rotary table (mostly from trying to keep each pulley set taught to keep backlash down) will be worth it for the slightly greater ease of threading later on. I will probably just gear it down with a single ratio and call it good.

[project5k]

could you just use the stepper setup and some tiny endmills and just cnc cut the threads rather than trying to turn your stock through a 1" die??