[aubreyj]

Hello all:
I am trying to choose a stepper motor and reduction gearing setup for a 9x49 BP clone. I am attempting to approach this from an engineering perspective as opposed to a guess-and-check or see-what-others-did standpoint, then see how my results compare to what works for others.

Design Constraints:
My goal is 20ipm minimum with a 30ipm optimum.
My power supply will be dedicated for the z-axis and will be matched to the motor.
Total weight of knee/saddle/table/vise/workpiece ~ 1100lbs.
I would like to keep the setup under $1k (I'm designing and building the brackets and guards).

Before I start:
1) Yes I understand that using the knee is not optimal for drilling and certain milling/contouring operations.
2) Yes I understand that the cast iron knee and cast iron column way/gib interface is not designed for high speed repeated movement and can accelerate wear and/or seize if driven too hard too often.
3) I am designing for use without gas springs deliberately as a sort of insurance policy. If I find that I have under-designed the system, I will add the springs to assist.

My understanding thus far:
-It is more important to match a stepper's power output (wattage) to the application as opposed to matching the torque.
-The maximum power output (wattage) for most steppers is just past the corner speed.
-Bipolar parallel wiring extends the corner speed out farther along the RPM range, effectively creating a longer flatter torque curve and more usable high-torque RPM.
-Increasing the power supply voltage in multiples of the stepper's rated voltage will extend out the torque curve with greater multiples providing diminishing returns.
-It is optimal to use a constant wattage/variable voltage power source as opposed to a fixed voltage power source.
-High amperage low impedance motors are designed for higher RPM use.
-Higher RPM at power (wattage) peak = higher gearing = more usable torque.

My approach to motor selection and gearing:
1) Find the PPR at the stepper's power peak just past the corner speed.
2) Calculate the RPM at peak, and determine the pulley ratio necessary to drive the gearshaft at 200RPM (20ipm at 10:1 shaft ratio).
3) Multiply torque output by pulley ratio to give approximate effective torque.
4) Calculate upwards force on the table using bevel gear/acme screw drive ratio, effective torque, and approximate drive assembly efficiency (acme screws are 20-40% efficient, bevel gears are 98-99% efficient + some losses in the bearings, friction in the ways, etc). I used 25% as a (hopefully) real world estimate.
5) Use the torque/force results to select motor and gearing.

I have compared NEMA 34s rated from 960-1810oz-in thus far as well as some NEMA 42s up to 2830oz-in. I am somewhat restricted in that not all stepper retailers publish legible torque curves (and even fewer publish power curves).

Thus far my standout system (keeping within my price range) seems to be an Anaheim automation 34Y314S 1700 oz-in NEMA34 at 9.9amps and 3.5mH using their constant power driver MLA10641.
This setup puts out an effective torque of 3965oz-in at 20ipm and 2643oz-in at 30ipm (3.5:1 and 2.4:1 ratios respectively). This provides a lead screw upwards force of 3892lbs for 20ipm and 2594lbs for 30ipm.

This is contrasted to lower amperage higher inductance offerings such as a NEMA 34 1810oz-in Keling setup at 8.8amps/5mH putting out 2098oz-in at 20ipm at 100V (1.6:1 & 2059lbs force).
Also contrast a Keling NEMA 42 2830oz-in 6amps/26mH running at peak power at 20ipm & 100V will only put out 1614oz-in (0.76:1 & 1584lbs force). Direct drive will only net 15ipm at peak power.


So I ask some of you veterans out there:
Am I on the right track?
Am I missing anything major? Elephant in the room perhaps?

Thanks much for any and all constructive input!

[aubreyj]

Hmmm... Lots of views with no replies.
Maybe I need to be more direct:

When selecting stepper motors and pulley ratios to drive a heavy BP type knee at 20-30ipm...

1) Is the statement; "It is optimal to use a constant wattage/variable voltage power source as opposed to a fixed voltage power source", borne out from a real-world perspective?

2) Does 25% make sense as a reasonable estimate for knee mechanism power transmission efficiency? (See original post) Should I be according more to friction forces in the ways?

3) I have calculated the time it should take from rest to achieve maximum speeds of 20ips and 30ips given the table weight and estimated forces applied to it from 3 different approaches (using inertia&momentum, using a free body diagram with F=ma, and using Newton's motion equations).
I keep getting the same answers of around less than 1/10th of a second!?
I'm not entirely sure that makes any kind of intuitive sense?
I would be much obliged if one of you mechanical engineers could check my calculations.

Thanks for any and all help!

[79TigerPilot]

Hi,
I converted a 1966 step pulley J head BP 9x 42 to CNC. Here is a list of my drive components. After updating with ball screws I can do 100"/minute but have kept it down to 75"/min max for safety.

The power supply is 65V DC 20A and my drivers are Gecko 203V's. Limit and home switches are inductive proximity type.

The knee lead screw is driven directly through a window cut in the side of the knee. Currently all axis have ball screws.

Current Mill Drive Setup:
X,Y axis
Tormach 1558 in-oz Stepper motor Nema 34
Drive configuration: (2:1 28 tooth and 56 tooth pulley
SPD/SI A 6A25M028DF1508 HTD 5mm pitch (28 tooth)
SPD/SI A 6A25M056NF1512 HTD 5mm pitch (56 tooth)
SPD/SI A 6R25M133150 15mm wide 133 teeth timing belt
SPD/SI A 6R25M117150 15mm wide 117 teeth timing belt

Knee:
Kelinginc.net KL342160-62-8A (1810 in-oz) Stepper
Drive configuration: (3:1 24 tooth and 72 tooth pulley)
SPD/SI A 6A25-024NF1508 HTD 5mm pitch
SPD/SI A 6A25-072NF1512 HTD 5mm pitch
SPD/SI A 6R25M133150 133 teeth 15 mm wide belt

Bridgeport Knee screw is 20-5/8" long (I think)
For Ball screws a little longer may be better to keep from running the nut
off when moving knee up to max Z. But that mayrequire raising the BP base
a little so the screw doesn't bottom with the knee down.

Quill:
kelinginc.net KL34H295-43-8B (906 in-oz) Stepper Motor


Picasa Web Albums - CNC Mill Project

[aubreyj]

79TigerPilot:

Great information - Thanks!

What lead pitch did you use on your ball screw knee conversion?

Also, do you ever have any regrets/drawbacks about choosing steppers over servos? Have you had any lost step and/or accuracy issues that you feel may have been resolved with servos?

At 75 & 100ipm with a 2:1 ratio, assuming 5mm/0.2" X&Y leads, that would put you right at 750 & 1000rpm with the steppers.

[79TigerPilot]

The knee is a 5 TPI lead screw
Steppers were easer and lower cost vs servos. Resolution is about the same. On my x and Y the pulse scaling is 20,000 pulses per inch when using the Gecko 203V that does 10x microstepping. One can't really rely on the microstepping to give you any real accuracy so I just assume my minimum resoulution is 1/2 thou (2000 steps per inch). But on the BP you are never really going to get that accuracy anyway, too many things betweent he stepper and a 750 pound table.

Without an external motion controller the pulse rate from Mach will limit the rapid speeds to 75IPM without upping the kernel rate. Servo's would be better in this case. I never loose steps but my cutting rates have been mostly below 35 ipm. Rapids at 75 don't loose steps.

Craig

79TigerPilot:

Great information - Thanks!

What lead pitch did you use on your ball screw knee conversion?

Also, do you ever have any regrets/drawbacks about choosing steppers over servos? Have you had any lost step and/or accuracy issues that you feel may have been resolved with servos?

At 75 & 100ipm with a 2:1 ratio, assuming 5mm/0.2" X&Y leads, that would put you right at 750 & 1000rpm with the steppers.