Ball screw size/pitch?
What dia and pitch ball screw size is suggested for a cnc router 60”x40”x8”?
Ball screw or precision ACME lead screw?
I want to be able to hold +/-.001.
Any suggestions would be appreciated.
Thanks,
Jakk
Ball screw size/pitch?
What dia and pitch ball screw size is suggested for a cnc router 60”x40”x8”?
Ball screw or precision ACME lead screw?
I want to be able to hold +/-.001.
Any suggestions would be appreciated.
Thanks,
Jakk
Faster is better...
Size will depend on the weight you are trying to move and how fast you want to move it.
I found this formula to estimate linear force developed using a screw / motor combination:
1. Convert ballscrew lead to equivalent pully radius:
example:
Lead =0.200
0.200 / (2pi) = 0.200 / 6.283 = 0.0318" (equliv. radius pully)
2. Convert motor torque into linear force:
example:
Torque = 100 Oz.-In.
100oz-in / .0318 in = 3144 oz. (linear force)
3144 / 16 = 196 lb (linear force)
196 * 0.8 (efficency for ballscrew) = 157 lb linear force.
If this formula is correct, a 100oz-in motor and a 0.200 screw can push 157 pounds force.
The force you need to push depends on weight of the gantry, friction of the bearings, and cutting forces. (not to mention acceleration).
Hope this helps.
Bill
Go to
http://www.nookindustries.com
Download the appropriate .pdf catalog (s), for ballscrews or acme screws. They list the force required to move a 1 lb load for each of their screws. For example, with my 1/2-4 acme, to lift 1 lb I need .064 in-lbs. Multiply by 16 = 1.024 oz-in to move 1 lb. So my 250 oz steppers should be able to safely give me 175 - 200 lbs even at a few hundred rpm as the torque falls off. And 300 rpm should be 75 inches per minute. We'll see when I'm done, but I'm hoping it works out like this.
Gerry
Gerry
UCCNC 2017 Screenset
http://www.thecncwoodworker.com/2017.html
Mach3 2010 Screenset
http://www.thecncwoodworker.com/2010.html
JointCAM - CNC Dovetails & Box Joints
http://www.g-forcecnc.com/jointcam.html
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
The screw is going to be very expensive if you are wanting to hold .001 tolerance over that long of distance.
Thanks
Jeff Davis (HomeCNC)
http://www.homecnc.info
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Accuracy and repeatability are not exactly the same thing. You might be able to broaden your requirements, if you can settle for the machine being less accurate, but more repeatable. If your screws are mounted in "tight mounts", so that you never get unexpected slop from reversals, if your slideways are properly positioned relative to the driving screw, if your slides are of sufficient bearing length to prevent jamming and stiction, then the machine should be fairly repeatable no matter what screws you use, providing that you always machine in the same direction.
Linear scale feedback is the best way to get high accuracy from low grade screws, however, low grade screws usually have a lot of backlash, which can make your machine "hunt for position" quite a lot.
Scales also cost a lot.
So, this boils down to spending the cash to get high grade ballscrews if you want fewer headaches down the road, IMO.
First you get good, then you get fast. Then grouchiness sets in.
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Thanks to all for their replies and patients with my questions. There will be more to follow I’m sure!
So using Bills formula, same values he used except the servos I am considering have 600 oz peak torque.
I would get 943 lbs linear force. Is that correct?
As far as holding tolerance, wont the software be able to dial most backlash and ball screw lead error out?
And wont that improve accuracy and repeatability? ( I’m guessing)
I would like to rapid at 600 IPM.
CNC router 60”x40”x8”. I think I’ll use 1” ball screw with 5 TPI on all aixs.
Will use encoders with 250 CPR.
I agree with HuFlungDung, “…spending the cash to get high grade ballscrews if you want fewer headaches down the road, IMO.”
Your thoughts would be greatly appreciated.
Thanks,
Jakk
Faster is better...
Jakk100,
Yes, a 600 oz-in motor is a pretty good size, common on a lot of knee mill retrofits. You should be able to break off a 3/4" cutter with it
Your controller software typically can be set up to make a backlash compensation. If it is really fancy, it may have a table of adjustments for so and so much adjustment at such and such a position.
Having said that, there are some limitations to backlash compensation, such as "You can't fool a climb milling cutter", and believe me, you will want to climb mill.
Also, backlash comp is tricky to make work really smoothly during arc cuts through the quadrant lines, where a reversal must occur, and the comp must be applied. This interrupts the cut with a slight pause and leaves a mark.
First you get good, then you get fast. Then grouchiness sets in.
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Dear Sir:
You ought to consider using rack and pinion drives for that size machine. As the screw length increases, you approach the "critical speed" for a given diameter screw where whip of the screw becomes a factor. Larger diameters (more $$$$) have a higher critical speed.
I have used rack and pinion drives on CNC plasma cutting machines which have high speeds, low backlash, and tolerate dirty conditions. Backlash can be held low, dual side synchronization can be accomplished mechanically, and sufficient drive force can be obtained, all at reasonable price. Most large commercial routers use rack and pinion drives.
Regards,
Jack C.
1. Convert ballscrew lead to equivalent pully radius:
example:
Lead =0.200
0.200 / (2pi) = 0.200 / 6.283 = 0.0318" (equliv. radius pully)
Hello Bill, I was trying to figure this formula out.. where did the 6.238 come from? and what does it mean to have the radius pulley equivalent ?
Thanks?!
Hi Jakk,
I have uploaded an Excel spreadsheet that might be able to help you with your ball screw selection. It's here:
http://www.cnczone.com/forums/linear...ml#post1671584
jcc3inc is right regarding critical speed for the ball screws. Basically, if you rotate a ball screw too fast it will whip. Think of a skipping rope. A 60" ball screw that's 1" in diameter can only be rotated at around 1600 RPM. With 5 TPI ball screws that'll only give you around 320 IPM speed.
Another thing to consider with backlash compensation is that the backlash has to be measured physically in order to be entered into software for compensation. I think it would be pretty hard to measure with 0.001" accuracy over 60" without some serious measuring instruments.
All this being said, there are machines out there that can perform to all of your desired specifications so keep going with your project!
Hi RasXYZ. Here is a formula for ball screw torque from the Thomson ball screw catalogue:
T = (F x P)/(2 x Pi x e)
Where:
T = Output Torque (lb/in)
F = Linear Force (lb)
P = Ball screw lead (in)
Pi = 3.14
e = Ball screw efficiency (which generally = 0.9)
Rearranging the above equation, you can find the linear force a motor will output through a ball screw like this:
F = (T x 2 x Pi x e) / P
For the example used in this thread:
F = (6.25 x 2 x Pi x 0.9) / 0.2
F = 176 lb Linear Force
Note: 100 oz/in = 6.25 lb/in.
Thanks,
Toby
Thanks Toby!
What I understood from this is that the finer the pitch the greater the linear force output. Giving me finer resolution.
But this would mean my steppers have to spin faster since the pitch is less.
I was going to be getting Nema34 motors. 800-1000 oz-in stepper motors. My machine would not see loads greater than 200 lb I would think.
I guess what I need to look at is the torque-speed graphs and figure out my desired speed range. See if the motor can handle at those RPM.
Are Nema34 motors over kill?? Is there any advantage.My machine is extruded aluminum 15 series. I want to move as fast as possible with out sacrificing quality.
Any comments welcome! I am still doing my research, I am still sourcing parts for the frame.
Thanks again, and ahead of time!
That's right RasXYZ, the finer the pitch of the screws the greater the linear force output. It's the mechanical advantage of a screw mechanism. With all mechanical advantage, however, comes speed losses. That's the trade off unfortunately.
Yes, you need to pay close attention to the torque/speed graph of any motors you're looking to use. Just because a stepper has a high holding torque it doesn't mean that it will output the torque you require at the speed you require. Nema 34 sized motors aren't overkill if that's what's needed to output your required torque.
7xCNC.com - CNC info for the minilathe (7x10, 7x12, 7x14, 7x16)
Lets back up a step.
The forces being applied are too much for 15 series extrusions. See: http://www.cnczone.com/forums/diy-cn...cad-posts.html thread for Dave Malicky's excellent analysis of various gantry sizes and the deflection you can expect. The unsupported lengths and overhang are a killer for deflection.
My machine is roughly a the same size you are looking for but uses 90 series Bosch Heavy extrusions for the Y and a 6x6x1/4" steel tube for the X. Still early in the testing but a few cuts in it looks good.
If you doubt my doubts, take the rails and mock it up. Clamp it together and then attach a dial indicator to a solid under surface aimed at the carriage. Now yank on it as hard as you can (look at the forces in the calculations which are huge)) and see what the deflection is. This will surely swamp the linearity of the drive system. I sort of thought his deflection numbers were high until I clamped the 6x6 tube on the mill and started machining the end to make it pretty. Amazing amount of deflection with even light end mill cuts before I put the diagonal braces in.
Now comes the hard part. Justifying how stiff you want it / need it vs the ever increasing size and rapidly increasing cost. This is a never ending battle between cost and expectations.
As far as whip, I see nothing noticeable on my chinese 20mm 10 mm pitch ball screws even at 250 ipm. This is with my 150-170 lb loaded gantry and double screws on the Y axis.
Agree with mbronkalla.
Vastly increased section sizes are needed.
Also, you will need to use servos if you want the accuracy to 0.001, and speed.
Also, you will NOT get 0.001 accuracy, without ground class 0 screws.
You can get local accuracy, ie feature size, of 0.001 inches, easily.
Just need a servo step resolution of half that, or better, or a full-step stepper resolution of half that, or better.
Ie you steppers full-step resolution needs to be 0.01 mm == 0.0005".
So the microstep at 1:10 needs to be 0.001 mm = 0.0005".
My machines are very accurate, and high resolution.
I use step sizes of about 0.2-0.3 microns, on 12" lathe, and a scratch built VMC mill.
All use ac brushless servos, now.
Speeds are 125 kHz (Pokeys) and 300 kHz (CSMIO-IPS).
Top speed is totally unimportant to me (although its 5x faster than steppers can do).
I use 25 mm (very short screws under 400 mm) to 32 mm diameter (1200 mm length) ballscrews.
To get good preformance on a 60" size, a 40-50 mm screw is the way to go, unless you make a rolling nut design.
A rolling nut is at least 4 to 8 times more rigid (half the free length = 8x more rigid).
A router does not have the requirements of the machines you use. We have a $100,000 router at work. The Y axis uses a 72" long 32mm screw, and it moves about 1000lbs, and cuts upwards of 1200ipm, with a 2Kw AC servo.To get good performance on a 60" size, a 40-50 mm screw is the way to go,
For home built routers like we're discussing here, a 25mm ballscrew should be more than sufficient.
Gerry
UCCNC 2017 Screenset
http://www.thecncwoodworker.com/2017.html
Mach3 2010 Screenset
http://www.thecncwoodworker.com/2010.html
JointCAM - CNC Dovetails & Box Joints
http://www.g-forcecnc.com/jointcam.html
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
Having high resolution servo encoder gives you more resolution, but not necessarily more accuracy, because you're relying on the motor's position not the actual linear position of the ballnut.
Also I highly doubt you'll find C0 screws long enough for the router needed. The gantry on a DIY router would be way lighter than any axis on a metalworking machine, and the cutting forces are very light comparatively as well. You can't spend your way to more ridgidity and accuracy.
BTW, is it me, or is this the longest running thread ever?! Holy thread resurrection, Batman!