Oy, what to do, what to do.

[MrWild]

I started researching because I'm new to electronics and want to do a CNC mill. Although I'm not clueless with electronics and hobby CNC, I am inexperienced. I have some "What would you do?" type questions and I'm looking for opinions. After I hear differing views, I'll probably use a salad of them to creat my own. Here is my desires, and my largess.

I want to fab a machine that is capable of milling aluminum and holding +-.003 tolerence with a budget that will make most of you cringe and say no way. Particulars are:

12" X, 8" Y, and 3" Z.

1/4"dia. end mills with not over 3/4" long flutes is how hard I plan to push the machine.

Linear bearings will be of the roller skate bearing variety but with a twist. 1" sq. x 1/8" wall tubing will be the rails. The linear bearing will be 1-1/2" sq. x 3/16" wall tubing with dual row skate bearings or a slightly larger version bearings for the top and sides with a single row on the bottom. A long enough structure will keep the stresses close to the corner weldments (linear rail will be bolted for adjustability). A slot cut in the 1-1/2"sq linear bearing will allow a strengthing rod for increased rigidity.

The column will be 4" sq 1/4" wall tubing. Corner legs will be same tubing as the linear bearings. Welding the base and column as a unit should give me a strong and wobble free mill.

So, what is my largess? A Fairy god father sent me a care package and now I have too many choices of which way to go for drive motors.

I have four or five good Slo-syn 2.5v 4.6a six wire steppers (Unipolar?). supposed to have 300 oz. in of holding power.

I have DC servo motors with encoders. 36v. 1.1a 2450rpm Hytachi DCM-05A02-E1080. Citizen P.O. H8003010.

I have three Slo-Syn six wire steppers with encoders 5.0v 1.0a 60 oz. in.

I have three 30v. 0.7a 4 wire steppers (bipolar?)

I have 9ea. 12v 20a transformers. 1ea 22.5a halving transformer (440-220, 220-110).

I have a 1/4hp DC motor ?rpm and speed controller for the quill motor the little brown men are bringing.

I have a Bridgeport and southbend along with 15 years of T&D experience in a large tool room. (disability retired now-really puts a crimp on expensive hobbies)

I have the odd bits and pieces for a power supply(s). i.e. 600v 35a rectifiers and enough capacitors of large enough size to smooth the high amps.
So with all that back ground,

Would you go with servos or steppers?

If servos, is there an open source drive? I am on a strict budget and I am not able to buy Geckos.

If you say steppers, the capable steppers I have are 2.5v and 4.6a. I've seen a thread with someone using pretty much the same steppers with Alan's great Picstep design. Can I expect enough power using a reduced amp Picstep driver and running the six wire unipolars as bipolars, or should I consider the 7a bipolar driver on Phils site? I'm not looking for speed, speed in metal breaks endmills. I'm retired so I can read a book for my "rapid" traverses.

I'm finally making useable PCBs, but find the process a real chore. Once the mill is finished it can do the drudgery. Doing it all myself though, which open source driver would you choose?

Lately there have been some what might be considered high end drivers from the old days selling on ebay for less than driver board kits. I don't have to do it myself if there is a cheaper and as good or better alternative. Should I steer clear of these drives, or embrace them?

Knowing I'm just looking for slow and steady, are the half step designs all that much yesterdays news compared to a microstep drive especially when none meet my steppers amp capabilities?

Running a unipolar as a bipolar is said to increase torque by 25%. With a reduction in amps if I go with a microstep drive, will I still see a 25% increase (over a unipolar at same under amped drive) Sort of, will the extra torque of biploar help regain the torque lost due to less amps? i.e. a 3a unipolar micro stepping drive will be less efficient than a 3a bipolar microstepping drive?

How far into the smoke zone (with large enough heat sinks) has Picstep been torture tested to?

Lots of questions, lots of area for opinions. My budget looks like this.

Drives and BOB = $150 total.

Metal = $150 (I have an asortment already)

Bearings = $40

Mcmaster lead screws and nuts = $50

Enco lead screw (for z axis) = $15

It sounds tight, but I am industrious and tend to not mind long hours of labor if it will save me a buck. I have a fair hardware assortment from the good years yet, so the nickle and dime stuff shouldn't hit me too hard.

Thanks in advance for all of you answers and opinions.

Bill

Yeah, I know I haven't mentioned software. That's a hurdle I'll hit when I have a machine well under way.

[jhwatts]

Based on my experience you will typically go over budget anyway. I have recently started retro fitting a mill. The cost I had originally intended has been greatly exceeded.

[soaringtuna]

Do you guys have a reasonable source for lead screws? I can't seem to find one (other than McMaster;..... which means $$$)

Any ideas?
Marc

[MrWild]

Originally Posted by soaringtuna Do you guys have a reasonable source for lead screws? I can't seem to find one (other than McMaster;..... which means $$$) Any ideas? Marc

I think you have to first decide what your machine is going to do, and then decide how large it will be, and then decide how cheap you want to get. You can use Enco lead screws and deal with the purported lack of tight tolerences, or you can go with a better lead screw. If you have plans for a 4' x 8' table, you'll want fast rapids to get the tool from one point to another, and also fast machining speeds so the wood is cut and not burnt.

Machining metal raises different issues. You won't get the stiffness you need to maintain rgidity with a large platform. Cutting metal means you slow it down and gear it up so you can creat the pressure needed for the mill to work. If you get too cheap with a lead screw by going with a soft material, or too thin a diameter, then you have problems with how the table travels. You can have stiff linear bearings, but the lead screw will allow the table to bounce around if it isn't strong enough for the job.

There are ways to create accuracy with less than optimal hardware if you are willing to take the time to achieve it. You can lap the lead nuts to the screw, use double nuts and either spring load them (whch creats friction and isn't optimal for metal cutting as it allows chatter) or trap them in an adjusted attitude so that they allow the least amount of slop and friction. My machine will have the McMaster lead screws because the problem of ganged inaccuracy is less than what you might find with the cheaper Enco lead screws.

For doing routing I like the Enco lead screw idea that uses a tapped chunk of plastic as the lead nut. You adjust the amount of slop by either wedging the split plastic nut apart or drawing it in. As much as it looks good for the lower stresses of a wood cutting machine, I'm leery of it with a metal cutting machine.

Do you need ball lead screws? Choices here. If your motor has the gonads, you don't need to have the low friction of a ball lead screw. If you need to have super tight precision (I mean how many lf us are going to utilize parts with +-.0003 accuracy?) then ball screws have their place. The trade off I've observed so far is you can put the money into a more powerful actuation system (driver/motor/gearing) or put it into state of the art lead screws and linear bearings. Cutting metal you don't need 150-200 inches per minute rapids. Lord, at that speed you just get to break more end mills. Having a mill capable of running across a 12" table in 1.5 seconds might look good at the drag strip, but you don't need it.

Why am I using quality acme lead screws for my X and Y but not my Z? Even if Enco Acme is out .006 in 12", my Z with just 3" of travel will see less than .002 deviation from actual over it's 3" of travel. I can live with that.

And if you only want to drill those blasted holes in a pcb board, you can get away with +-.010 tolerance and still get the lead to solder to the trace. For a small machine like a 4" x 6" PCB driller, you could get away with 1/4-20 grade 5 threaded rod. Not as cheap as allthread, but worth the extra few dollars. A lead nut can be made by oiling the rod, running two nuts close but not tight/snug, and glueing them with thin super glue. Of course you won't route with this method (trapped glued lead nut) but it will be plenty strong for moving a drawer slide linear bearinged PCB drilling table around.

Bill


These are just my opinions and my opinions sometimes are . Most of my opinions have been earned through trials of fire. Sometimes just because the book says something is impossible due to a math equation doesn't make it so in real life. However most of the time the mathematical answer is the correct one.

[soaringtuna]

An excellent description of the various options...

I have already built my router, and I have been producing parts for a few months now... its 16'x32' working area... I have been machining balsa and light plywood, so the duty is certainly light... I had used 8mm threaded rod for all the axis but have found that its too slow, and flimsy. It works, but i want to upgrade this at this time.

Really, I the biggest difference, are my drivers. I have built a 3 axis driver (i am an electrical engineer) for under 15 bucks, but it isn't a chopper. Now that I have obtained sample ICs for some chopper drives, I am building a new driver (3 axis based on this). So, my 3A steppers are going to haul A$$... and I want to upgrade the lead screws to be able to handle it, instead of destrying themselves... Which is why I wanted a source for this stuff...

BUT: cost is a big issue. ENCO, I have never heard of that, but will check it out. I don't need super tight tolerances, I mean, most of the balsa parts are ok within 0.05 and I would be happy... (RC planes), but I know I can do better with better screws.

Is there a place we can get this sutff used?

Cheers all!
Marc

[MrWild]

This guy has the right idea of make it work Mephisto engineering.


http://www.cnczone.com/forums/showthread.php?t=14239

[paulC]

You may find these interesting. They came up in an earlier post.
It is a cheap thread screw that looks pretty good.
http://www.hobbycnc.hu/CNC/Otletek/Otletek.htm
http://www.hobbycnc.hu/CNC/Mech5/Mech5.htm
http://www.hobbycnc.hu/CNC/Mech8/Mech8.htm
You would need access to a lathe.
If you used stock thread the pitch could be quite unpredictable as it would be die cut. As long as it is consistant you should be able to compensate with software.
Should be able to use the idea on all your axis's as they are quite small.
Also has the advantage of having a fine pitch which would effectively gear your machine down, giving more touque for your aluminium cutting.
Paul

[soaringtuna]

That is quite the elegant solution ... I have access to a lathe, so that is manageable for me, so long as I can find someone up here in Canada who can sell me a piece of ACME screw of any size (I am desperate at this point and would take anything bigger than 3/8".

Ideas?

[MrWild]

Originally Posted by soaringtuna That is quite the elegant solution ... I have access to a lathe, so that is manageable for me, so long as I can find someone up here in Canada who can sell me a piece of ACME screw of any size (I am desperate at this point and would take anything bigger than 3/8". Ideas?

Yes it is. I had to sit and think of why it moves. When you roll two peices of threaded rod together, nothing happens. The cloclwise of one counteracts the counter clockwise of the other. Zero movement. Thing is, the rollers are larger. The larger rollers with the same pitch have to go twice the distance for one thread to progress. Because of this diferential, it acts like a gear reduction on top of smooth and zero backlash. He has too many slots though for someone with just a drill press to copy it.

The top two legs can be held in simple accurately placed holes. Placement becomes an issue then though. I think all the slots may be to make lining it up easier. If the screw is off an eigth inch, just adjust the rollers to compensate.

You built your drivers for only $15? Good Job.

[paulC]

No, I don't think there is any gearing.
As the planetry threads are captive the whole thing has to move at the rate of the primary thread rotation. The advantage is in more contact points on the thread, less friction and the ability to remove backlash.
Paul

[MrWild]

But they're not captive completely. If they were held solidly in position there would be movement of one thread forward or back as the lead screw turns a full revolution. The rollers are allowed to roll however (why they have bearings). If the rollers were 1-1 dia. of the lead screw, and they had enough contact pressure so that they rolled at the same speed as the lead screw, there would be no movement at all (see my mind experiment above). With a larger diameter roller, the lead screw turns them at a reduced speed. As the lead screw ramps up, the counter rotating rollers ramp down. Because of their larger diameters they ramp down more slowly (due to their larger circumference) so that there is less than a full thread of table movement for every rotation of the lead screw.

Great concept. I'm ready to go buy 1/4-20 and 1/2-20 threaded rod and see if I can bore the 1/2-20 precisely enough to make roller 1/4-20 lead screws. It'd have to be extremely precise or there'd be binding.

Now if I could just make some 4a. microstep drivers for $15ea. I'd be in heaven.

Bill

[paulC]

To provide a gearing, the threads would have to move relative to the primary thread.
As they can't, they simply spin with the motion being transfered to the axle and then to the frame.

When selecting the thread diameter remember you need clearance for the bearing. The bearing has to be smaller than the planatery thread with this design.

1/4 inch threads can whip about a lot. I think you will need something bigger unless your axis travel is extreamly small. There is no requirement for the diameters to be different. You could make the whole thing from the same diameter thread.

Paul