Indexing Headstock Spindle

[digits]

digits, Is the lathe headstock axis horizontal and the rotary table vertical? (you said on top of). Might be a rigidity problem without some heavy duty (expensive) bearings and couplings etc. I did see some interesting roller gear stuff at the AGMA gear show in Detroit. www.nexengroup.com Might work for some kind of indexing reducer.

Sorry - I didn't make myself clear, the lathe headstock and the rotary table are co-axial - ideally the lathe spindle would go right through the centre of the rotary table and its bearings would be within the rotary table, but that's not practical, so I'd imagined a small diameter drive shaft from the lathe motor behind the rotary table, to the lathe headstock infront of it. The workpiece would be in the lathe chuck, which would move with the rotary table for indexing or independently for turning. I'll whip up some CAD if I get a chance...

Thanks for the link - that looks interesting, but probably a bit hard to DIY!

[RICHARD ZASTROW]

eric, I hope we're not hijacking your thread, just seems to be growing.

digits, I believe your concept can be done. Might be a bit long though. You could have a hollow shaft gear box that would give you lathe speed and a second gearbox to provide indexing. A central shaft running through both, as you mentioned. The central shaft (aka spindle) suspended by its own bearings, and engaged by hydraulic clamping sleeves (Spieth makes them). Would have to be safely set up so both could not be engaged at the same time.
Yes, very Rube Goldberg and impractical. Also, probably expensive as well. Might even qualify for something for a Kostner sequel to "Water World". lol

[digits]

eric, I hope we're not hijacking your thread, just seems to be growing.

digits, I believe your concept can be done. Might be a bit long though. You could have a hollow shaft gear box that would give you lathe speed and a second gearbox to provide indexing. A central shaft running through both, as you mentioned. The central shaft (aka spindle) suspended by its own bearings, and engaged by hydraulic clamping sleeves (Spieth makes them). Would have to be safely set up so both could not be engaged at the same time.
Yes, very Rube Goldberg and impractical. Also, probably expensive as well. Might even qualify for something for a Kostner sequel to "Water World". lol

I had to Google Rube Goldberg - he translates to 'Heath Robinson' in the UK

I stumbled across a fairly simple gearbox design called a perpertual wedge, which allows pretty big step-down ratios in a reasonable area. If I can build one using synchronous pulleys/belts, I guess it'd also be pretty low backlash. My hunch is that it would be possible to lock the input and output shafts together to get a 1:1 drive ratio when 'lathe mode' is required. I'm going to look into this some more.

[BobWarfield]

Late to the game, but this is an interesting post.

I'd love to be able to index my lathe spindle to facilitate C-axis operations. For example, put an air drill on your gang slide and you can now do bolt circles on a flange under CNC control.

It's also important to be able to index a mill spindle if you want to do an automatic tool changer for a taper that requires drive dogs.

A couple of thoughts I've had to this general problem. First, I am envisioning a system that does not be able to continuously machine as the axis rotates. It is a pure indexing system that indexes a position, locks the spindle, and then performs the machining task.

Second, this implies to me a spindle brake. That makes me think of a disc brake. So I envision a rotor with a disk brake caliper to look the assembly. Probably use a motorcycle-sized rotor/caliper to keep the weight from getting too crazy. There's another thread here where a fella did just that and controls the brake with an air solenoid. I would think if you can find a scrapped bike the disc rotor and caliper would be very cheap to buy.

Third, how then to index? Again, it seems to me the easiest thing is to arrange a second motor to do the indexing. I've looked at vector drives with sensors that effectively turn the spindle motor into a servo, but the cheap ones can't seem to position like a servo--they just use the encoder to broaden the full torque speed range.

So why not take a big stepper motor, or better, a servo motor, and use that to do the indexing? Seems to me servos are designed to run at the kinds of speeds these spindles turn at. Just bought a nice one from Homeshopcnc that'll do 4500 rpm, for example. So we make sure that we can run one motor or the other BUT NOT BOTH electrically. We freewheel the servo when the spindle motor is going, and freewheel the spindle motor during indexing.

Why a servo? Well because we can put the encoder onto the machine's spindle (not the spindle motor!) and let it sense the true motion and soak up any backlash to increase accuracy. There won't be much anyway if we use a toothed belt drive. If you prefer, a decent sized stepper is cheaper still. You'll want to groove the toothed belt else when the spindle motor runs it'll scream like a banshee, but that's easy to do too. So, we need to add a toothed belt pulley to the existing spindle motor pulley stack and drive it with a suitably mounted servo or stepper.

I'll have to think about how to interlock the two motors so they never run at the same time, but sure seems like this does the trick relatively cheaply and easily.

Make sense?

Cheers,

BW

[digits]

Hi Bob - if you don't want to be able to use this as a rotary table and cut with reasonable force while the spinlde slowly rotates, I guess a brake and some sort of encoder is all you need.

I do have to ask though, why you need 2 motors if you are not planning mega torque at low speed and high-speed, low-torque operation? Couldn't you just use a beefy servo motor and a brake?

What you have to remember is that motors and generators are the same things. If you have a drive motor set up at say 1:2 (motor:spindle) and a stepper/servo set up as an indexer at say 5:1 ratio for big-torque, when you run as a lathe at 2000 spindle-rpm, your poor old stepper will be forced to run at 10,000 rpm. This will generate huge voltages across the stepper - you would explode its driver unless it was totally electrically isolated, and you would have to take great care when switching motors not to explode drivers by connecting/disconnecting anything unless everything is stationary. If you can work out a way of making the indexing backlash-free, I would have thought that it would be best to mechanically disengage the motors in a twin-motor system.

My interest in this thread is more along the lines of lathe-spindle/rotary table in one as I really want to reduce the number of setups I need to make parts - indicating inside the crush/mince zone isn't much fun! I think though that machines such as the Cublex 25 just have scary 30HP servos with 200Nm of torque from the motor alone! This must be a cooling nightmare at low rpm with the huge currents required to generate all that holding torque...

[BobWarfield]

Digits, take a look at the cost of a 3 HP servo motor and appropriate drive and I think you'll see why we don't just use a big servo motor. A relatively cheap import set is $2000. Blow it up somehow and you get to spend another $2000 fixing it. If you really truly need to machine a portion of the spindle arc rather than moving the part over to a 4 axis mill, I guess I could see you problem. But, to actually build something like this is going to be a massive project. You'll wind up having to build a mill's axes on your lathe or close to it. Getting all that working rigidly and without backlash makes me think you're better off buying a used lathe where it's already been taken care of.

RE this business of 10,000 rpm generator steppers, you're not envisioning the right linkage. There's no reason to run the servo/stepper as anything but 1:1 to the same drive we're applying the spindle motor too. So it'll run same speed as the spindle motor. I don't need huge torque for indexing, and the brake is so I don't need holding torque to machine the part. Also, part of the reason to build an interlock from contactors is to keep an open circuit on the indexer motor rather than have it complete a circuit that we then have to dispose of the current from.

RE your cooling nightmare, just use a separate electrically driven fan activated by a temperature sensor so the fan rpm is independent of the motor's rpm.

See more on this stuff on my blog:

www.cnccookbook.com

Search for "vector drive".

Cheers!

BW

[digits]

I suppose the first question I need to ask is how much spindle power do you actually need - and better still, if you know it, what sort of torque at what sort of speed? Those homeshopcnc servos are good for 600W continuous (~0.8HP) and can be driven cheaply with a G320.

I'm guessing though that you'd rather have a 3-phase induction motor for your spindle. If you want to simply index without cutting, unless you have very irregular parts , I can't see that you'd need much torque at all for the indexing. So, why not just drive your main spindle with a drive capable of indexing and add an encoder for feedback? If you don't need 3HP to index, couldn't you poweroff your VFD, disconnect it from the motor and then use a cheap AC-servodrive (e.g. Granite Devices) to index at only 100-200W of power?

Two motors can be made to work, but IMHO it is a bad idea to keep them both mechanically connected at the same time - couldn't you just physically disengage the indexing motor from its belt when you change modes?

[BobWarfield]

I suppose the first question I need to ask is how much spindle power do you actually need - and better still, if you know it, what sort of torque at what sort of speed? Those homeshopcnc servos are good for 600W continuous (~0.8HP) and can be driven cheaply with a G320.

I'm guessing though that you'd rather have a 3-phase induction motor for your spindle. If you want to simply index without cutting, unless you have very irregular parts , I can't see that you'd need much torque at all for the indexing. So, why not just drive your main spindle with a drive capable of indexing and add an encoder for feedback? If you don't need 3HP to index, couldn't you poweroff your VFD, disconnect it from the motor and then use a cheap AC-servodrive (e.g. Granite Devices) to index at only 100-200W of power?

Two motors can be made to work, but IMHO it is a bad idea to keep them both mechanically connected at the same time - couldn't you just physically disengage the indexing motor from its belt when you change modes?

1. 3HP, as I mention above would be a good number. Take a look at the spec on most CNC lathes--that's still light. 7.5 HP is a more likely number.

2. The Granite Devices approach may have potential. I'll have to give it some thought.

3. Where's the harm in driving with 2 motors? As I've mentioned, the circuit will be opened, so we're not generating any current. What design do you propose to economically disengage the motor?

[digits]

1. 3HP, as I mention above would be a good number. Take a look at the spec on most CNC lathes--that's still light. 7.5 HP is a more likely number.

OK, so that's going to be a pretty chunky main motor and drive - what's a ball park cost for a 7.5HP motor and vector drive?

3. Where's the harm in driving with 2 motors? As I've mentioned, the circuit will be opened, so we're not generating any current. What design do you propose to economically disengage the motor?

Even if the 2nd motor is open circuit, it will still generate a large EMF when spun at turning speeds. You would have to be sure that this can't arc across your disconnection circuit, or reverse bias any semiconductors in the drive if you're disconnecting electronically. What circuit do you have in mind for swapping the motors over, and what happens if you hit e-stop or have a power failure under high-speed/high load operation? If your circuit disconnects a motor while it is still spinning, I would expect really, really bad things to happen to the driver board. Switching a multi-kilowatt inductive load is not trivial!

I'd also imagine that a 7.5HP spindle motor is going to have a shed load of rotational inertia - your indexing motor will have to overcome that if both motors are always physically coupled to the work. If your drive train uses gears, couldn't you just rotate one motor so that it comes out of mesh, or use a mechanical clutch? Similarly, a belt drive could move a motor pulley out of the way and push an idler up against the belt in its place to maintain the same tension?

Just my 2p's worth...

[BobWarfield]

digits, thank you for your thoughts. I don't see anything here that isn't pretty straightforward to deal with.

RE your thoughts disengaging, I have a hard time imagining the gear linkage that isn't full of backlash, which is a no-no for an indexing rig. It's possible a belt and idler arrangement would work better, but even there, we would typically want a toothed belt arrangement for reasons of backlash.

Doesn't sound very practical to me.

Cheers,

BW