[extrapilot]

Hey guys

Im looking at some design options for a 6’x3’ router table. In doing some math on forces involved, I realized- I have no idea what the stepper motors’ ratings are for the bending/shear loads they see in a direct drive approach (in a rack/pinion or “servobelt” implementation). If you assume a couple hundred pounds of linear force from a direct drive, you have that, but also with rack/pinion some cam forces are in play.

Seems like a non-issue, since there are many direct-drive systems. But, it seems possible that this is a source of slop in the drive (i.e. some deformation in a thin NEMA23 axle, or some runout in the bearings, etc). Are there published specs for motor tolerance to these lateral forces?

Regards

[ger21]

Oriental Motor has that info for their motors.
http://www.orientalmotor.com/product...mmon_specs.pdf

[extrapilot]

Hey Gerry

Thanks for the reference. Seems all the manufacturers use different terminology (i.e. max overhung load vs max radial load etc).

Seems this is a real problem for servobelts and R&P. I mean, for a typical NEMA34 with an overhung load rating of 50-70lbs (some as low as 25), you simply cannot use direct drive. I wonder if the spec is a mechanical limit, or just the test limit for the max runout, etc. Obviously, there are tons of direct-drive steppers in service in R&P designs, seeing more than 70lb linear force.

I saw a number of your posts on the servobelt thread- and seems like you sort of gave up on direct drive. In just doing some basic math, it still seems plausible to use direct drive, where a 960oz/in motor driving a 12 tooth 0.2” tooth pitch pinion should be capable of 500-ish oz/in of torque (75 lbs linear force) at 500 in/min. 10 microsteps gets you to about 1.5mil plus a little slop in the belt/pinion- so maybe 2mil. Not great, but you are going to get frame flex on that order unless you are using 3x3 steel frames…

Anyways- thanks for that reference. If you have any additional insight on the servobelt approach, Id be happy to hear it- as it seems an elegant/light/inexpensive approach if done correctly.

Regards

[ger21]

I saw a number of your posts on the servobelt thread- and seems like you sort of gave up on direct drive.

I never thought about using direct drive, as the resolution with a direct drive stepper would be very poor. With your example 12tooth pinion, you'd see the machine move 2.4" per stepper revolution, which is .012" per full step. One of my design goals was to get the full step resolution close to .002". Most will tell you that the microstepping resolution can't be counted on for precise positioning.
From what I've read on the Mechmate forum, a 3:1 reduction provided noticeably better quality cuts than direct drive.

I don't believe that the overhung load on the shaft is equal to the linear force. With the ServoBelt, for example, when accelerating, all the load is on one side of the pinion pulley, with virtually zero load on the belt on the other side. What the shaft is seeing is a torque load, not an overhung load.

The overhung load only needs to be enough to keep the belt meshed with the pulley.

When I built my prototype, I did have a similar concern about the forces on the pinion shaft, but it was mainly due to increasing the belt width from my original design.
I think it's mentioned somewhere in the ServoBelt thread that the shaft that the large "idler" bearings run on receive the majority of the forces. The belt around those bearings is actually loaded from both ends. One end has the motor pulling it, and the other end has the weight that needs to be pulled.

I could be wrong here, so feel free to correct me.

[extrapilot]

Hi Gerry

Its not a pure torque load. Just as with tank treads, there is a tension side and a slack side. Even if you add massive pretension, that is a vector operating in the same direction as the tension side. The slack side does not operate in compression by definition, so it cannot offset the tension vector.

Seems feasible to just bound the pinion and idlers on both sides of their axes, so that the stepper is pure torque. Im just looking for the simplest approach- less to go wrong with slipping couplers, etc.

I don’t have any solid data on if one can reliably microstep at 1/10 for low speed, higher torque. Marris is probably the guy I need to ask about that. But in theory, if a typical 0.1” depth pass in say 6061 with a ¼” bit at say 5krpm is 20-30N- maybe 10lbs force, that is not much for a couple of NEMA34s on the X axis to handle. Y axis may be more limiting, but at least there is less inertia.

With 1/10 microstepping, on a 1” pinion, with 1.8deg steppers, Im showing 1.5mil, which is about the same flex I see in FEA on some 80/20 2x4.5” frame rails under 50lbs load…

[ger21]

I have some images posted either in the ServoBelt thread or my design thread showing some brackets to support the other end of the pinion and bearing shafts. Instead of making the brackets, I just redesigned it to use bigger diameter shafts for everything.

I don’t have any solid data on if one can reliably microstep at 1/10 for low speed, higher torque. Marris is probably the guy I need to ask about that.

I think Mariss is the one guy that will tell you you can.

But in theory, if a typical 0.1” depth pass in say 6061 with a ¼” bit at say 5krpm is 20-30N- maybe 10lbs force, that is not much for a couple of NEMA34s on the X axis to handle.

For my machine, acceleration forces will be far greater than any cutting forces. I designed it to have enough power to accelerate my 200lb gantry at 1G, but will probably only be able to run at 50-75% of that, to have a safety margin to prevent lost steps. At least I hope I can accelerate that fast.