I am building a second mill to convert to CNC. The first one I built was kindof built on a whim, and was only used for the occasional manual milling. The table was built so that it was fully supported, with the Y-axis on top, so the rails were at the ends of a 24" long table, but the table itself was only 8" long. Thus, any time you started milling off the center of the table, the table would try to twist, and would bind up against the rails. The way I solved this was to put leadscrews on each end of the Y-table, and connect them with a belt. I made a crude tensioner, and I never once noticed any binding, nor did I ever notice the belt slip. The belt was a crude system put together using pulleys and automotive belts from the local hardware store.

For the new mill:
I really like the idea of a fully supported table like that, especially since my new mill will have a 24"(x) x 12"(y) travel, and one of the things I intend to mill are camshafts, using a vertically mounted rotary table + Tailstock, fastened to the mill's table. Obviously, if I get much angular deflection of the table as it reaches the end of its travel, the cam would be angled during machining, and the lobes would be cut crooked. I havent bothered calculating how much each distance increment of deflection at the end of the 24" table would affect the angle at the spindle, but if I can do a fully supported table, or even a mostly supported table (>50% of the total 48" of travel), then I could avoid that deflection altogether.

My concern about mounting the axes traditionally is that the Y-slide is 6" wide, but would have 48" wide X-Rails sitting on top of it. As you could imagine, a 24"x12" table all the way toward the end of its X travel would be a pretty big cantilever to be supported by the relatively small 6" wide Y-slide. The structural part of the slides are made of .25" think Aluminum and Cold-Rolled Steel bars, with the cold rolled steel bars acting as the slides on the Aluminum rails. The Aluminum rails stick out about an inch into the Steel bars off their mounts.

So, my question is, could I make the Y-axis like I did on my old mill, using two leadscrews and a belt drive? The Belt drive would obviously have to be pretty long, although it might only have to be as short as about 24" (from one leadscrew to the other). Ive seen belt drives used on routers that were pretty accurate, but can they be stiff enough to work with a mill? What are my other options?

Or, am I worrying over nothing?

Here are some images. Other than the screws / Bolts, they are dimensionally proportional.

Here is the Y-slide up close. The light colored parts are 6061 Aluminum, to act as the friction surface. The grey parts are the cold rolled parts, which is the slide that the X-axis would be bolted on top of. The dark grey parts are hot rolled steel, for the actual structure. Both the Aluminum and Cold-Rolled bits are .25" thick, except the thin aluminum sliver mounted vertically to the insides of the hot roll square tubes, which is .125" thick (another friction surface). The cold rolled bars are 6" wide, and go back 12". The bolts allow the cold rolled bars to be adjusted on the aluminum rails by either pushing them apart or pulling them together. They can be locked in place by tightening both sets of bolts against eachother, with one set pushing apart and one set pulling together. The Aluminum rails are bolted to the top of the hot roll tubes, down into the base, with about 1" sticking off the tube to support the Y-slide. The hot roll tubes can be pushed together using the horizontal screws toward the outsides, which will push against the slide to prevent it from twisting in the rails. The blue part is where the X-axis rails would bolt up. The leadscrew dogs (the threaded "nut-like" part that the leadscrew drives) would be mounted on the bottom of the lower cold-rolled plate.
I hope this explains it well enough for you to help.

http://www.lindenwoodracing.com/mill_y.jpg


Here is another, showing a (roughly) proportional view of the X-rails bolted to the Y-slide.
http://www.lindenwoodracing.com/mill_y_full.jpg



Cliffs

Can I use a Belt-drive system on two leadscrews about 24" apart (or further), with accuracy? This would allow me to use a fully- or mostly-supported table, to prevent table deflection from the cantilever effect as it reaches the ends of its travel in the X-diretion.

Thanks!

*edit*


OR, is there a way I can adapt two motors to run on the same signal? I will most likely be using the Xylotex 4-axis kit, so I dont know if something like that would work. However, that would definately make it the most accurate if it could work.

Hey J,
You can set up Mach 3 to use the A,B or C axis as slaves to the X,Y and Z.
Config>Slave Axis
With the Xylotex, You could have the A axis as a slave to the Y.
Use identical motors and make sure the Motor tuning parameters are exactly the same, each driving their own ballscrew.
Hoss

So I could buy an extra motor with the Xylotex kit, and configure Mach3 to send the same signal to two motors? Would i need to buy any extra hardware to go with the extra motor, or could I just splice the motors together into the same harness from the motor drivers?

And, just to clarify, you are saying I could do this with 3 separate axis?

*edit*

Does TurboCNC have this feature?

Hey J,
You would need one motor cable hooked up to the Y axis and one motor cable hooked up to the 4th A axis, don't splice them together. If you used another 2 axis Xylotex or (2) 3 axis boards, you can have a slave for the X and Z too if necessary.
I don't know if turbocnc can do this or not.
here's a few threads I found in the cnczone Forums.
http://www.cnczone.com/forums/showthread.php?t=31347&page=2
http://www.cnczone.com/forums/showthread.php?t=31404
http://www.cnczone.com/forums/showthread.php?t=32136&highlight=slave

I was considering using 2 motors on my Z axis each driving the single ballscrew
with belts and pulleys.
This shows someone using 2 motors to drive a spindle.
http://www.chicobritish.org/CNC/4)%20Spindle/slides/4)%20Top%20View%20of%20Pulleys%20in%20Place.html
Doesn't look like i'll need to now, but it's an option.
Hoss

what would be the purpose of 2 motors on the spindle.. i really cant see needing that much torque, and there'd be no speed gain.. now, using something like that arrangement to drive a z screw, alright now i could see it...

I could have the same benefits of using two on the Y-axis. I could mount the columns and rails further apart, so elastic deformation wouldnt play as big a role in accuracy. I could have also done a bridge style mill, which would have saved me lots of space (the total work area would have been about 48" x 30", rather than the current 48" x 40"), and very likely gained a bit in the rigidity department.

*edit*

Question:
What all is actualy needed to drive each motor from the same computer? It doesnt seem that the Xylotex will work with more than 4 axes, so I figured I might as well buy all the parts individually.

Thanks!

well you can drive up to 6 motors in mach i believe, so you could have 3 axis with 2 motors on each.. you need a breakout board, and then a driver for each motor. and a powersupply that could drive all 6 motors. i would have to go back and see, but if i remember correctly there arent enough pins to run 6 outputs on a single printer port, so you would need a second port and breakout board. then in mach slave the motors for each axis together. i would suppose it might be possible to put 2 motors wired in parallel to each other on a single driver, but i would think this dangerous as you would have no way to correct it if one got ahead or behind the other. plus the combined inductance and current to that single driver would be pretty high, so it would take a serious driver. the motors would have to be perfectly matched, and so many other factors....
i would strongly caution against this approach, the pitfalls are just too great.

personally i like the sepperate components idea, that way if there is a problem it's much easier to trouble shoot, and if there is a failure, it will hopefully only affect one rather than taking out a bunch... its just like buying a tv that has a vcr built in vs. buying the 2 sepperately. if one fails, you only have to repair/replace one part, but if thier together, then if one dies the other might work, might not, and if you wanted to repair/replace one, then you have to do both.. yes up front the sepperated parts will cost a little more, but in the long run i think its a much better bet.

just my .02

If you want to add more drivers for say the B and C motors, get another 2 axis board( or (2) 3 axis boards). If it needs a parallel port too, you can add a second parallel port PCI card. i got this one at Dell (http://accessories.us.dell.com/sna/productdetail.aspx?c=us&l=en&s=dhs&cs=19&sku=A0401239) a while back just in case I need it. I still have my old Stepperworld 3 axis driver.
Wanted to have the parts ready in case I needed a Slave for the Z axis or to play with a 5th Axis.
There are cheaper parallel cards out there than Dell for sure.
crap, here's one I just found thru pricewatch.com (http://www.pricewatch.com/) for $9 (https://gogocost.secureserverdot.com/xcart/product.php?productid=275).
My xylotex is a bipolar driver but the stepperworld is a unipolar (why i yanked it)
I would make sure the slave is using the same board as the master.
Hoss

Project5k you type faster than I do:)

You could also just connect your 2 Leadscrews with a Timming Belt and Pulleys to syncronise them.
Drive the Timming Belt with one Motor inbetween the Leadscrews.
Seams simpler to do, and be more efficient Power wise.
The Force would be dirrected into the leadscrew that lags behind.
So you would not need 2 Motor of the same strength to drive independently.
Just a Thought!
Good Luck.

Thanks for those tips!

I was also going to ask how accurate it could be if I drive two leadscrews with one belt? If I could use timing belts rather than multiple motors, it seems it would save a lot of complexity for me. The leadscrews will be about 40" apart (Y) and 25" apart (Z). The Y leadscrews would have to move about 50+ lbs I'd guess, and the head will probably weigh a good 50+ lbs also. While mass might be added in some areas to improve rigidity, travel speeds are a very minor concern, so I am not expecting much belt-stretching acceleration.

Well as with most things mechanical there are different options
on configuring the Timming belt arrangement.

1) good - one timing belt covering the 40" distance driven somewhere inbetween.

2) better - two timming belts comming from a double Timming pulley mounted
motor centered between the leadscrews this would shorten the length of each timming belt because the centerdistances only needed to be 20".

Depends on the accuracy your machine will be capable of, if the overal design is configured to end up with less then 0.001" accuracy I would go with option 2 else option 1 would be simpler and sufficient.