Hi all,

I would like to make a 4-axis CNC milling machine for machining small aluminium parts, and I'm trying to decide which spindle and axis motors to use.

The budget is less then £10,000 ($20,000). For that money I would like to achieve:

200mm of travel in the XYZ axes, 2m/s speed.
Spindle speed of 20,000rpm.
Auto tool changing
(plus the usual coolant flow, PC, drive electronics etc, enclosure)

Can anyone suggest a suitable spindle and axis motors to use? So far, the best looking ones are just outside my budget. Spindle FS45 from Hofer Precision: http://www.hoferprecision.com/0311/hfmotorized.htm and linear stepper motors from H2W: http://www.h2wtech.com/stepmottor.htm

Can anyone suggest something as good for less money?

Hugo

You are looking at things well outside the scope of milling aluminum there. The spindle really doesn't need to do 120,000 rpm, and you are going to be paying orders of magnitude more for one that does than a more practical 20,000 RPM spindle. NSK has some nice ones for under US$5000, there are good options all the way down to under $1000.
The linear steppers are a bit impractical as well. I would look at any number of pre-built machines, there are good ones out there already, even if you only use the basic chassis to build up something semi-custom from.
Reinventing that wheel completely from first principles is not only likely very unnecessary, its a hell of a lot more technical and difficult than you might think.

Thanks Stepper, that's very helpful. I have been Googling for spindles for ages, and could only find people offering them in the range of $10,000. I was sure there had to be something less expensive around somewhere.

Yes, 20,000 rpm is exactly the speed I want, not 120,000. Where did you get that number from?

Why are linear steppers impractical? As far as the control is concerned, they are surely identical to steppers on leadscrews, but they require more current.

Normally I'd think about buying a maching, we have three already, but a) they cost a packet, b) they're much larger than they deserve to be considering the size of the parts they can make, and c) they're never quite what we want. It would be great to have a machine which is designed specifically for manufacturing the parts we need. Also, I am very comfortable re-inventing the wheel in this case. I am a professional robotic engineer, so I'm right at home here.

Thanks
Hugo

I will state the obvious then. Since you have experience and machines already, then you know milling creates a dirty environment. It is extremely difficult to seal ways off from just swarf even if you don't intend to use coolant. The motors you linked specifically say for light loads, so you are looking for some semi high speed machining? Very light repeat cuts?

These motors certainly appear fast enough. If you take very light cuts, they may do, but at a premium price. I don't have experience with linear's but I am curious how well these might work on the Z, which would carry some weight on it with the spindle.

Here are some relatively cheap spindles.
http://www.homeshopcnc.com/page7.html
I imagine they are for light duty cutting like what you intend. As opposed to hogging off metal.

I was finally able to design and built a mill myself. Just a standard benchtop size mill. The entire machine was likely around $5000.
I did manage to get 6600 RPM on my spindle. Just a standard X2 Chinese made mill head and tread mill motor with timing gear and belt drive. The entire spindle setup was well under $500.

I normally run the machine @ 200 IPM rapids, but set it to 300 yesterday. It would do 350 for certain, but this is pretty dang fast for such a small machine.
I have 14" x 9.5" x 9" travels. You want to move yours much faster. I am using preloaded precision ground ball screws and nice beefy THK linear ways.
My motors are 495 oz running off 72 VDC @ 12 amp PS and Gecko 203V's set for just 3 amps.
Bigger motors and more amps might push this machine up to a speed you could live with. Not sure though.

At any rate, I look forward to seeing what you come up with. The wheel has been reinvented a lot on this forum and most still roll well. :)

milling creates a dirty environment
Yes. I have been thinking a lot about how to seal off the slide ways from swarf and coolant. I think I've thought of a really nice way to do it. I'll post some CAD images later.

I am curious how well these might work on the Z, which would carry some weight on it with the spindle.
They wouldn't have enough strength to lift the weight of the spindle quickly, so I'm thinking of using a spring to support the weight.

Here are some relatively cheap spindles.
http://www.homeshopcnc.com/page7.html
I imagine they are for light duty cutting like what you intend. As opposed to hogging off metal.
Thanks! These look pretty nice, and not madly expensive. But what I'd really like is one with a pneumatic collet, so the machine can change its own tools. Currently we have a 10-tool changer machine, which is nice, but we do still have to swap out tools all the time. What I really want is at least 20, if not 100. This is feasible for a machine what can just let go of one and pick up the next. (but not for our Emco 105 which has to have the 10 tools installed in tool holders).

I normally run the machine @ 200 IPM rapids, but set it to 300 yesterday....You want to move yours much faster.
Whoops! My bad. I wrote 2m/s! Jeez that would be very fast indeed. It would probably shake itself to pieces. I meant to write 2m/min.

I was finally able to design and built a mill myself.
I'd love to see a pic. Is there one on this forum somewhere?

Hugo

As someone who used to build linear motor (servo) based robotic platforms, I can tell you that 2 m/s is freaking fast. Furthermore, getting up to 2 m/s in a 200 mm work envelope will means you have only 100 mm for ramp up, for an acceleration of more than 2 g's. We did make a pick and place robot that was just capable of this, but starting cost was $80,000. Is there a reason you need to move that fast? At that speed, you'd be able to cover your entire work envelope in 0.2 seconds (assuming a triangular velocity ramp). Seems like these sorts of speeds would be more important for a larger work envelope.

Also, I would counsel against a linear motor for the Z -- it's nice to have this axis hold still when you hit an e-stop. My $0.02.

Ahren
www.cncrouterparts.com

Oops -- saw your reply right after I sent mine. I see 2m/sec isn't required anymore. A factor of 60 less should be completely do-able :).

And a linear stepper machine would be really cool to see!

Ahren
www.cncrouterparts.com

it's nice to have this axis hold still when you hit an e-stop.

That's a very good point. Hmm. How about if the e-stop just cut off the step/direction signals? Or does an e-stop have to cut power too for health and safety reasons?

Hugo

Hi Hugo

Here is the link to Lee's build.
http://www.cnczone.com/forums/showthread.php?t=8632&page=1&pp=40

Regards

Andy

Whoops! My bad. I wrote 2m/s! Jeez that would be very fast indeed. It would probably shake itself to pieces. I meant to write 2m/min.


I'd love to see a pic. Is there one on this forum somewhere?

Hugo



I thought that was a little fast for most things. That would be a good speed for a printer. ;)

Here is a link to my build thread here at the Zone.
http://www.cnczone.com/forums/showthread.php?t=32609

Thanks, Andy. That's my router. ;)


As Ahren warns, you may have issues with this on the Z. I use a gas spring on my Z to keep it from spinning down when power is cut. This keeps the head in place on the screw when power is cut. It holds position really well.
My Z axis probably weighs around 70 pounds or so and the gas spring is rated @ 30 pounds. This doesn't over work the axis on the down path and assists in raising some.

How would a linear motor hold position if power were cut? I think you would have to have a precisely weighted spring in order to pull that off. Percise anything usually costs.

In practice you really want the e-stop to cut power to everything. And then there's always the problem of machine startup and shutdown -- I guarantee you'll forget to apply the brake or park the z at its lowest setting. From experience, (we had a high-lead ball screw that would backdrive under the weight of the Z), I would make this axis screw driven. You can still get high speeds, but without the worry.

Ahren
www.cncrouterparts.com

Leeway, that's a very nice build. Properly stable looking, and nicely machined.

I couldn't get the video to play more than a couple of seconds before freezing :( and the video from the router seems to have been removed.

there's always the problem of machine startup and shutdown
That's a good point. OK, how about the springs are over powered, and when the power cuts, the spindle automatically lifts up? I don't mind about loosing axis position because I'm going to get the machine to auto-zero the tool height automatically anyway.

Hugo

Thanks.
Here is a direct link to my video's on YouTube.
http://www.youtube.com/user/LeewayMan
I fixed the dead link on the router build as well.
Thanks.



If you use an over powered spring, you will need some heavy duty motors to run the Z down all the time. This is why I use less than half strength.

If you use an over powered spring, you will need some heavy duty motors to run the Z down all the time

What I meant was that they would be well balanced, and the lifting force was only a couple of newtons more than the weight. Of course this would mean making sure the spring didn't travel much compared to its length so that the force remains fairly constant.

Hugo

Hi swiss
These speeds are doable if you use AC motors or linear motors I have some doing 1.5G
& able to position at +/- 1micron at these speeds you also need the control to do this as well & theres only one control that I know of that can achieve this & it can rapid up to 300m per min that is over 11,000 inch per min & cut at 60m per min that is cutting at 2,300 inch per min
Look at softservo systems for a control as for a spindle the faster you want the machine
to move the higher the RPMs you will need so a 20,000 rpm spindle you are looking at 200 to 400 inch per min cutting You still need a good control with got lookahead to make this work

Hi mactec,

That control package looks sweeeet! I bet it costs a packet. Any idea how much?

Actually, I would rather use brushless linear motors than steppers. But I decided not to because a) The Mach3 software we use only seems to support steppers, and b) I'd have to spend ages designing and building the PID control modules.

Sadly, I reckon this would be a bit outside my budget :(

hugo

Ahren wrote: And a linear stepper machine would be really cool to see!
Yeah, I think so. I was hoping that someone on here would have made one already, but I can't find anyone.

Hugo

So, reinvent the wheel. :)
You can be the first.

So, reinvent the wheel.
You can be the first.

Stop, stop. You're getting me all excited. I couldn't sleep last night thinking about milling machines, and I was late for work.

If anyone's interested, here's my idea for the machine. Rather than being a massive beast, I want to make a fairly small footprint one. It would be 2' x 2' x 7' tall. It would be sturdy by virtue of being bolted to a supporting wall of our workshop.

The frame (green) would be made from that nice aluminium extrusion. The bottom of the machine would contain a coolant tank, the middle would house the milling stuff, and the top would house the electronics.

The rails etc would be mounted on a granite slab (blue-grey). Then some rollers (blue) mounted on the top and bottom of the slab.

Then a very thin sheet of stainless steel (light grey) would be wrapped around the rollers. At the y-axis moves, the sheet would run round the rollers. This should protect all the Y-axis slides from a great deal of swarf etc. I would like to do something similar for the other two axes yet.

Anyone think this is a good or bad idea?

hugo

Hugo, this machine sounds really fascinating - can't wait to see the finished product! I hadn't ever seen the linear steppers before, would they suffer from the possibility of lost steps?

I thought Mach3 would work fine with linear servos, as long as your controller will accept step/direction input? Wouldn't the servo drive/software take care of the PID? I have a Trilogy linear motor sitting on the shelf I was rather hoping to use this way. Okay, I'm over my head here :confused:

I had the same problem you're having with choosing a spindle for the machine I'm halfway through building too - reading your specs you might as well be reading mine! - except for your fancy linear steppers. Looking at your posts, it seems you would prefer a spindle with auto tool change capability that doesn't use separate tool holders, but rather will grip the shank of the tool? I thought the Hofer FS45 used toolholders, though I might be wrong. It looks very similar in appearance and spec to the Kavo 4041 that I settled on for my new machine, but the kavo auto chuck uses collets which may suit you better.

I really wanted to comment regarding the NSK spindles stepper monkey suggested too - I used one of their E3000 series spindle motors with their NR-5041 auto tool change spindles in my first build, and I've pushed it hard enough to find out the weaknesses! It will take up to a 1/4" shank max. though I've found it hard to avoid chatter with 1/4" cutters over 2" in length, perhaps because the collet won't allow the shank to go especially deep. If the cutter does chatter badly it can jam the air chuck, then it doesn't drop the tool which is a pain. 1/8" cutters are no problem, they will go in as deep as you'd want them to. The Kavo is still sitting on the shelf while I build #2, but it appears not to have the same collet depth limitation. It also feels like a more solid unit. So it should be too, they cost a lot more.

Having said that though, for light work the NSK is very nice, quiet and accurate. Max. DOC in aluminium with the 1/4" cutters seems to be about 0.02", so it is light duty only. For 1/8" cutters or smaller, it seems to have ample power. With 1/4" cutters I cut at around 2.5m/min, .015 DOC, ~35,000 rpm mostly, which is pretty close to your design speed. It cost about $3500 all up with the controller, cables, motor, spindle and air line kit. I think the real bonus has to be the separate spindle and spindle motor though - you can replace just the spindle for half the cost of rebuilding the Kavo, really good for for a novice like me who is likely to crash the machine at any moment... I wouldn't hesitate to get another one for my current project, except that circumstance steered me towards the Kavo, I already had a controller, and the spindle was offered at a price too good to refuse).

Anyway, I think I've gone on long enough. All the best with your machine, I'll be watching this thread with interest.

Regards,

Jason

Hi Jason,

Thanks for your reply. Your spindle suggestions have set off a whole load more research. While googling for "Kavo 4041", I found www.5bears.com which was very interesting (maybe that's your site?). I definitely want to use collets, as tool holders cost more and just take up lots of room. I'm fairly sure the Hoffer uses them too.

1/4" is the largest cutter we'd ever use too (actually 6mm). But 0.02" seems to be a very shallow cut. The tools we use can cut as deep as their diametre, at about 37in/min, though they are less than 1 inch long.

$3500 is pretty good value. I was quoted something like $10,000 for the Hoffer spindle! I'm going to spend some time looking at these more closely.

Hugo

Jason,

I was wondering too about using the step/direction from mach 3 to somehow control a linear motor (take the step position and subtract the real encoder position, then feed that into a PID controller) but I think it's probably not a good way. I believe that the best servo motor controllers make use of the desired speed an acceleration of the motor, which is fed to them from the CNC software. However, a controller that's being fed from only step / direction pulses would have to infer the speed and acceleration, and would therefore have this information just a little too late.

I'm sure the linear steppers could suffer from a loss of steps just the same as normal stepper motors could. However, I have a feeling that the steps are a bit larger, so it might be harder to miss them.

hugo

www.5bears.com isn't my site - I would be proud if it were! It was inspirational to me, I read and reread the CNC mill project on that site before I even started thinking about getting into this addictive field. The gentleman responsible is a member on CNCzone (Swede) from memory, though I don't think he has been active for a while.

The Kavo might actually be somewhat cheaper than the Hofer if you're buying new? I'd be interested, I got mine as a second hand refurb. The power supply and frequency converter were priced to me at about 1600 euro, but I didn't get a price for the 4041 spindle.

Would the cutting force of the cutters you refer to above cause missed steps with the linear steppers? A deeper cut is certainly possible with a smaller diameter cutter (paradoxically...) with the NSK, as the low power output is the limiting factor with the 1/4" cutters. I've been increasing RPM and feed to compensate. Lots of light cuts.

I see you're looking to machine Aluminium - I'd be happy to run a test with the NSK spindle with your cutting parameters and a similar tool if that would help? It might give you an idea whether the NSK would fill your needs - I'd hate to be responsible for raising hopes unjustifiably, or encouraging you to buy a spindle that wouldn't fit your needs...

Jason

I suspect you know considerably more than me here - I don't know much about servo behaviour at all to be honest.

Zone member Xerxes built his own linear motor I think, there was mention of it in a thread here somewhere. He also has a company making very interesting servo drives - I plan to buy some for my machine as soon as I'm ready for them. Perhaps he might comment on the issues you mention regarding using Mach3 and linear servos? (I'm particularly interested in this since I already have a linear motor I must use :)

I must say, the linear steppers sound much more straighforward...

Jason

I think you are right to say they may be less complicated than linear ways and screws, but unless highly powered, I can see them useful in much but non contact cutting. That is unless you are talking very light high speed stuff.
I would suspect this would also be limited to aluminum and plastics cutting. The stronger mag field would attract too much swarf in steel milling.
Something like plasma, laser, printing etc.. would likely work fine with linear motors.
Milling with them seems to just be unwise given the cost, especially when they state for light duty.

I do like that design though.
A caged machine. That is cool. I see how swarf deflection would work to some degree. I would suggest that you use rubber mounts on the walls when securing it. I don't yet have any resonance in my machine, but that could be a problem and better to forecast it now than have to redo later.

The stronger mag field would attract too much swarf in steel milling

That's fine for us. We only ever cut plastics and Alu.


for light duty
Yes, their power is my biggest concern. I'm keen to work out how much power I need for milling. My very rough guess, pulled out of my a** is that I'd want at least 100N (40lb) force for cutting with our 6mm dia tool. Looking at the specs for one of the lower powered motors, I can see that a pair of motors per axis will give me 130N continuous force, and 520N peak!

Someone else reckoned that the acceleration would be the biggest problem. For a 10kg table, a peak force of 520N would give me 5Gs acceleration, which is plenty for me.

Even if that's not enough, the most powerful ones in the range give 955N continuous force, and 3820N peak! Surely that's enough? The thing I really like about these motors is the almost total lack of moving parts. Nothing touches, nothing rubs, nothing wears out. Attaching the coils to the base, and the magnets to the moving axis means that even the wires don't have to move. They could be extremely reliable.

I would suggest that you use rubber mounts on the walls when securing it.
That's an interesting idea. I was thinking about the sanity of screwing it to the wall, but mostly I wondered if it might hard to do it without deforming the machine. Can you tell me more about the resonance? What specifically would resonate? Wouldn't a secure connection to the wall hold it tight enough that it wouldn't resonate?

Hugo

Hugo, this machine sounds really fascinating - can't wait to see the finished product!

Don't hold your breath. It's still a total fantasy. I need to convince my company that it's worth building rather than buying. I'm pretty sure they'll see the point of building when they look at the prices of similarly speced machines.

hugo

Thanks for mentioning the NSK E3000 and tool changer, I think I'd like to check that out with the manufacturer!

Where do I look for bits for this thing? Any suggestions? I'm thinking of carving wood, plastics, plaster, and light metals.

Thank you!
DIY-Guy "Just a computer guy making the switch from soft-ware to hardware...."

Where do I look for bits for this thing?

Doesn't it just take normal milling bits?

Hugo - any success with convincing your company to go ahead with this project? Whenever I've looked into buying a machine rather than building I come to the same conclusion, that I can build what I need for a fraction of the price, and get exactly what I want that way too.

DIY-Guy - I should mention that I'm becoming wary of using the E3000 spindle with the 1/4" cutters in aluminium - I would consider this to be the limit of this spindle, proceed with caution. Probably would be fine in plastic, wood or plaster.

I get carbide 1/8" shank bits here:

http://stores.ebay.com/CARBIDE-PLUS_W0QQcolZ4QQdirZ1QQfsubZ0QQftidZ2QQpZ2QQtZkm

They come in a nice box, and most have the depth stop rings. I like the 1/8" 2 flute end mills from Kemmer precision. They work really well in plastic, with a nice smooth finish, and pretty well in aluminium too. I haven't tried them in wood or plaster.

For 1/4" shank cutters, I've been using p/n RV323-5504 from Enco (Atrax short 2 flute uncoated carbide) lately. Nice finish in 6061 aluminium, and cheap!

Jason

EDIT - yes, it just takes normal bits...

Doesn't it just take normal milling bits?

Sorry. I'm just a computer guy, a newbie in the world of CNC. So I asked where to find bits of that size. I don't (yet) know a "normal milling bit" if it bites me. :)

Do you have any suggestions for where to look online? Maybe a few URLs please?

Thanks again!

DIY-Guy
"Transitioning from software to the real world, where machines rule!"

DIY-Guy,
I live in the UK, so I wouldn't know where to go in the USA.

If you're cuttingg Aluminium, be sure to get a tool which is specifically designed for aluminium. It's a sticky metal, which will build up on the tooth, making it blunt, and eventually causing it to fail. The right tools will have a coating which prevents this build up.

Hugo