Simple Solution- Drive 1" Ball Screw Shaft with NEMA-23-Size Stepper Motor Shaft

I have some 18-inch long 1-inch diameter ball screw shafts with dual-path ball screw "nuts." I also have some NEMA size 23 stepper motors I'd like to use to drive these ball screw shafts. About 0.65-inch end length of these long shafts is cut to a concentric 0.75-inch diameter, apparently for insertion into a support bearing. NEMA 23 size stepper shafts seem to be either 6mm or ~0.24-inch diameter. Shaft connector blocks add unwanted length, unwanted expense, and reduce heat rejection from the motor's shaft into the driven shaft, another unwanted effect. Heat rejection from stepper motors ultimately determines their duty-cycle electrical current input limit. Colder environments and better air flows enable higher driving currents. Similarly, well thermally-connected aluminum motor mounts with lots of surface area are able to reject higher heat flow rates which enable higher motor-driving currents than comparatively thermally-isolating plastic motor mounts, which would be the opposite polar extreme. Typical steel motor mounts are thermally somewhere between the best and worst. Most motor heat rejection discussions fail to consider dumping heat through the motor's shaft, yet that's a valid but minor consideration.

I've seen motor shaft couplings which include flexible synthetic rubber links. If both shafts are not concentric, flexible links allow much longer bearing service lives. But synthetic rubber is a lousy thermal conductor. Solid steel, brass and aluminum shaft couplers generally limit thermal conduction by their limited pressure contact areas onto both shafts more than they limit thermal conduction within the coupler.

Here's my simple, minimalist plan. Bore and tap a hole 90 degrees to the long screw's axis located within the 0.75-inch diameter end section. Then bore another hole about 1-inch deep into the long screw's axis starting from the 0.75-inch diameter end. Diameter of that concentric hole will be a light press fit to accept the stepper motor's shaft, which will have two flats ground on opposite sides against which two short Allen screws will press. These Allen screws will bind the two shafts together
yet be short enough to allow a flange-type 0.75 inch internal diameter bearing to support the long ball screw shaft axially and pick up single-direction thrust loads into the bearing support frame.

This configuration will be shorter, less expensive, and provide better motor heat rejection than traditional end-to-end shaft coupling configurations.

I keep learning by reading how others have found clever ways to design systems. I don't know if this suggestion rises to that standard. But it seems like it should be considered when coupling a small diameter shaft to a much larger diameter shaft.
John

Unless everything is completely in line and mountings are square to the face then you could introduce binding or radial loads on the stepper.
Thermal conduction will not be a problem on a well designed system but interference will be.

You would be better off inserting a short stub into the end of the ball crew and use an Oldham coupling to join the screw to the motor. An Oldham coupling is the only coupling that can accept errors in two planes without imposing a radial force.

The easiest way to mount a stepper is simply to bolt it on to shaft. Add a tie bar to stop the motor turning and let it wobble.

hang it by the wires

I have at least 50 different tools which clamp into 30,000 rpm hand grinder collets with such concentric accuracy that side-to-side "run-out" inaccuracy is probably less than usual ball-screw bearing clearances. Why would we cut the shaft-retaining bore hole so inaccurately that run-out would cause interference problems at ballscrew rotation speeds? I have an air-driven dental drill that spins over 100,000 rpm, and its removable tools seem to stay pretty concentric with the motor's shaft. If they didn't, it would vibrate like crazy or self-destruct. Bore it accurately and forget it for these ballscrew spin rates.

Oldham couplings do impose radial forces on non-concentrically-connected shafts. That radial force is zero only at zero rpm, but it becomes positive as rpm increases from accelerating the non-concentric sliding coupler. If you're saying that at ballscrew spinning speeds, that radial force is trivial, I agree. But if you tried to couple two non-concentric shafts at hand grinder speed, that force would become objectionable.

John S said, "Thermal conduction will not be a problem on a well designed system . . ." Following motor manufacturer guidelines to determine limits to which you electrically load motors is one very conservative strategy. Prudent manufacturers rate their devices based on a distribution of usual expected environmental conditions. They leave some head room in those ratings so end users usually won't overheat motors they market. One big issue with permanent magnet motors is gradual flux loss from heating those magnets. That's why permanent-magnet-based position-sensor coil inductors fail so often in automotive applications. They are mounted in differential gearboxes generating pulses with every passing gear tooth, or in engines which swing iron parts past, inducing pulses in their coils with each swing. As these magnet's strength gradually degrades from heat-cool cycling, often before those vehicles reach 100,000 miles, one of those sensor's pulse strength drops below the threshold required to trigger the position sensing circuit. Similarly, permanent magnet based motor's performance can slowly degrade from heat-cool cycling. Any way you can reduce peak motor heating may extend that motor's original performance behavior. I'm not saying that rejecting motor heat through their shafts is a big deal, but it has value and dismissing that value because motor makers don't suggest it is your choice, not mine. I wonder why Emco attached radial fin heat sinks to some of their mill's stepper motors. Those motors didn't need those external heat sinks to comply with the motor maker's requirements. How hard you can safely drive any specific motor depends mainly on balancing heat generation against heat rejection.

Quote:

---

Originally Posted by LoveLearn

Why would we cut the shaft-retaining bore hole so inaccurately that run-out would cause interference problems at ballscrew rotation speeds?

---

We would because the mounting tolerances required for a solid coupling are generally too high. Lack of concentricity places a cyclical load on the bearings and the load rating for that type of load on a bearing is significantly smaller. This is one of the numerous advantages of using a timing belt drive.

Quote:

---

Originally Posted by LoveLearn

I have an air-driven dental drill that spins over 100,000 rpm, and its removable tools seem to stay pretty concentric with the motor's shaft.

---

Tool is only supported on one end there, so different situation entirely.

Quote:

---

Originally Posted by LoveLearn

Oldham couplings do impose radial forces on non-concentrically-connected shafts. That radial force is zero only at zero rpm,

---

Yes they do, but only due to friction between the sliding disk and aluminium end pieces. It will be negligible. Again if you're worried about radial forces use a timing belt as that applies a constant radial load which deep groove bearings tolerate better than cyclical loads. Still, that's analysing it too much . . much better reasons to use pulleys, such as matching the motor torque/rpm characteristic and resonance supression.

Quote:

---

Originally Posted by LoveLearn

One big issue with permanent magnet motors is gradual flux loss from heating those magnets.

---

See here:

http://www.electricmotors.machinedes.../bdeee3_5.aspx

Quote:

---

Ceramic or ferrite magnets lose about 0.13% / °C of their remanence above 25°C, while rare earth and AlNiCo may lose only 0.03% / °C. But this loss is generally reversible if the temperature is kept within the motor rating. Colder temperatures are seldom a problem. Since the coefficient curve is linear, magnets are stronger at lower temperatures.

---

So the temperature rating is a threshold, which if you don't exceed that the magnets will not suffer permanent damage. The figure often quoted for stepper motors is 80°C case temperature, which is below the temperature rating of the magnets to account for the thermal resistance between the case and rotor.

Quote:

---

Originally Posted by LoveLearn

As these magnet's strength gradually degrades from heat-cool cycling

---

That's thermal fatigue, which is a different problem not addressed by shaft heat rejection.

A heat sink on an expensive machine doesnt mean it's necessary, many machines add aesthetic and unnecessary features to appear expensive. The manufacturer's ultimate objective isn't a top notch machine but money. Also the motors to which you're referring are now obsolete - the design and materials have improved significantly since then so if heatsinks were required then, that does not imply they are required now.


As long as the motor is bolted to an aluminium mount I'm sure it will be fine, unless you're running them at a high voltage. Either way just make it, measure the temperature and think about adding heatsinks if you need to. My hottest motor (Z-axis) never goes above 50°C and that's just on 3mm thick mild steel.