Are taper lock pulleys a good choice for ball screws?

[JeffZ]

I'm currently trying to figure out what the best way to install a HTD 5m pulley onto a 2010 ball screw is.
It will be driven by a 4.5Nm stepper motor through a closed loop timing belt.
This is what I have in mind:
29859

One option would be to have the fixed end of the ball screw machined, either flat (into a d shaft), or into a keyway. This would accept standard AF/BF style pulleys with grub screws. However, I imagine that the loads of a Nema 34 motor are too high for a d shaft setup, and I'm not sure if keyway machining is a reasonable request.

During my research I've also come across taper lock pulleys. At first glance, they seem to be exactly what I need and don't require special ball screw end machining.
From what I've read, their disadvantage is that taper pulleys are large (start at 32 teeth) and heavy (steel).
My design can easily accommodate for the size, and I can't really imagine the weight of the pulley being an issue.

298602986129862

Unfortunately, I cannot find much information on people using them with ball screws.
Do you think that in my use case, taper lock pulleys are a good choice?

Thanks,
Jeff

[Muzzer]

I've used taperlock pulleys wherever possible, as they are much more secure and concentric than keys and grub screws and are often seen on professional machines. As you say, the main disadvantage is the size. All 3 of my machines have them where space permits and my Shizuoka came with them from the factory. I've used keys and grubs where space is tight (lathe cross slide and Bridgeport Z axis) and those don't feel as secure in my mind.

[JeffZ]

That's a relief, taper lock pulleys it is then.
Thank you!

[JAZZCNC]

That's a relief, taper lock pulleys it is then.
Thank you!

Wait there's a big difference between a router and a milling machine so it's not that simple.!

The milling machine is often rotating the ball-screw much slower with a lower pitch ball-screw. A router is using a high pitch and spinning the screw upto twice the speed and often much longer length (heavier) screws so the inertia becomes more of an issue. Also, servo's or steppers change the outcome. I've got a feeling muzzer is using servo's which have much higher torque than steppers.!

The weight of the pulley and the effect of inertia DOES affect the machine negatively in some circumstances ie: 3D where you are accelerating and de-accelerating quickly with lots of direction changes. Rapid positioning like when drilling lots of holes over a large area where you need to accelerate hard and then slow down between holes.!

All these kinds of operations are affected by the diameter and length of the screw which play a big part in the effect inertia has on the system, so any extra weight you add will only affect this negatively. If you plan to operate at high feed rates or do lots of 2.5/3D work then think carefully about any extra weight you add.

If the screws are long and the gantry heavy then 4.5Nm isn't a massive amount of torque if rotating the screws at higher RPM's.!

[JeffZ]

Wait there's a big difference between a router and a milling machine so it's not that simple.!

The milling machine is often rotating the ball-screw much slower with a lower pitch ball-screw. A router is using a high pitch and spinning the screw upto twice the speed and often much longer length (heavier) screws so the inertia becomes more of an issue. Also, servo's or steppers change the outcome. I've got a feeling muzzer is using servo's which have much higher torque than steppers.!

The weight of the pulley and the effect of inertia DOES affect the machine negatively in some circumstances ie: 3D where you are accelerating and de-accelerating quickly with lots of direction changes. Rapid positioning like when drilling lots of holes over a large area where you need to accelerate hard and then slow down between holes.!

All these kinds of operations are affected by the diameter and length of the screw which play a big part in the effect inertia has on the system, so any extra weight you add will only affect this negatively. If you plan to operate at high feed rates or do lots of 2.5/3D work then think carefully about any extra weight you add.

If the screws are long and the gantry heavy then 4.5Nm isn't a massive amount of torque if rotating the screws at higher RPM's.!

Thank you, definitely appreciate all the information.
This is for a diy router milling machine. I'm still in the initial design phase, so I have not ordered any parts yet.
Right now I'm planning for Nema 34 size motors, but that 4.5Nm figure (or motor type) is not final.
I'll figure this out once the dimensions of the machine are locked in.

I got curious and checked the mass of a 36T htd 5m taper lock pulley, and they are listed at 450g.
Definitely more than I was expecting.

I'll be sure to post my design and ask for critique and comments before ordering any parts.

[dazp1976]

Wait there's a big difference between a router and a milling machine so it's not that simple.!

The milling machine is often rotating the ball-screw much slower with a lower pitch ball-screw. A router is using a high pitch and spinning the screw upto twice the speed and often much longer length (heavier) screws so the inertia becomes more of an issue. Also, servo's or steppers change the outcome. I've got a feeling muzzer is using servo's which have much higher torque than steppers.!

The weight of the pulley and the effect of inertia DOES affect the machine negatively in some circumstances ie: 3D where you are accelerating and de-accelerating quickly with lots of direction changes. Rapid positioning like when drilling lots of holes over a large area where you need to accelerate hard and then slow down between holes.!

All these kinds of operations are affected by the diameter and length of the screw which play a big part in the effect inertia has on the system, so any extra weight you add will only affect this negatively. If you plan to operate at high feed rates or do lots of 2.5/3D work then think carefully about any extra weight you add.

If the screws are long and the gantry heavy then 4.5Nm isn't a massive amount of torque if rotating the screws at higher RPM's.!


I've got a 112mm taper lock and it is weighty.
It's for the 1.8kw servo that's going to be for a mill spindle.
Wouldn't fancy that on a ball screw!!!.
Going to have a go at lightening it with pockets (sort of trivial persuit segments) not cutting all the way through.

[JeffZ]

I did some more research and came across these types of clamping pulleys:
29868

What are your thoughts on these?

[JAZZCNC]

This is for a diy router milling machine. I'm still in the initial design phase, so I have not ordered any parts yet.

Well, that's a good thing you haven't ordered anything yet because you need to decide if it's a Router or a Milling machine because when it comes to component choice's they are very different, and getting it wrong will impact either machine negatively.


Right now I'm planning for Nema 34 size motors, but that 4.5Nm figure (or motor type) is not final.
I'll figure this out once the dimensions of the machine are locked in.

Again be careful because bigger isn't always better.! . . For instance, Nema 34's, even smaller Nm ratings are often wrong for a Router unless they have high voltage drives because they cannot reach the RPMs before torque drops away. 99% of the time you'll find the same Nm rating or lower NEMA 23 or 24 will outperform the NEMA 34 given the same voltage. This is mostly down to the inductance of the smaller motor being much lower which allows higher torque/rpm's.

Also, look at Closed-loop motors/drives for both machine types as they are much better than standard stepper systems.



I'll be sure to post my design and ask for critique and comments before ordering any parts.

A wise choice and the right approach.:thumsup:

[Muzzer]

I've got a feeling muzzer is using servo's which have much higher torque than steppers.!


Yes, my machines are a 3.5 tonne Shizuoka, a 1 tonne Bridgeport milling machines and a Colchester lathe. The ballscrews are mostly 32mm and 25mm and they all have servos, so as Jazz says, the application is a key part of the decision.

Personally I wouldn't consider the additional moment of inertia of a taperlock pulley to be a massive concern if you have a long 20 or 25mm ballscrew but if you have steppers, I guess you may struggle with speed and response as it is. Without a high drive (power supply) voltage, the torque output of steppers quickly tails off, so the headline torque figure may be misleading. If you are still deciding what to use on the motor front, this may be a good time to consider something better and Jazz has good experience you can learn from on that front.

[dazp1976]

Nema 34's, even smaller Nm ratings are often wrong for a Router unless they have high voltage drives because they cannot reach the RPMs before torque drops away. 99% of the time you'll find the same Nm rating or lower NEMA 23 or 24 will outperform the NEMA 34 given the same voltage. This is mostly down to the inductance of the smaller motor being much lower which allows higher torque/rpm's.

A wise choice and the right approach.:thumsup:

Even with motors of near the same inductance is proven.

Results using 60V on a G0704 sized mill, all driver settings identical, X & Y axis:

Nema34, 1080oz (7nm and change) 3.5mh inductance:
Loud, bad resonance, max velocity 2000mm/min, stalls thereafter, Inertia causes 0.2mm overshoots at stop.

Nema24, 566oz (4nm) 3mh inductance:
Near silent, max velocity can exceed 3000mm/min, stops dead, accels like a rocket.

Still have a 1080oz on Z, (2000 is quick enough there). Might need it for my spindles new servo motor (7kg weight)
Upping the power to 72V, might change driver to a stepperonline 100vdc input one (psu can go to 80).


Nema 34 1600oz (12nm) YES I was one of those newbies :whistle:
Thrown straight in the bin. Too embarassed to mug it off onto someone else!

[JeffZ]

Even with motors of near the same inductance is proven.
Nema34, 1080oz (7nm and change) 3.5mh inductance:
Loud, bad resonance, max velocity 2000mm/min, stalls thereafter, Inertia causes 0.2mm overshoots at stop.

Nema24, 566oz (4nm) 3mh inductance:
Near silent, max velocity can exceed 3000mm/min, stops dead, accels like a rocket.
And here I thought it was all down to inductance.


Still have a 1080oz on Z, (2000 is quick enough there). Might need it for my spindles new servo motor (7kg weight)
Upping the power to 72V, might change driver to a stepperonline 100vdc input one (psu can go to 80).

Which drivers are you considering?
I have the CL86Y or CLA86T from stepperonline on my list of options for this project.
I could run them directly off 60VAC

[johnsattuk]

Taperlocks are used extensively in industry, I have made my own using ER collets, pic of one on my bench mill and one on my router, both driven by servos

[JAZZCNC]

Personally I wouldn't consider the additional moment of inertia of a taperlock pulley to be a massive concern if you have a long 20 or 25mm ballscrew but if you have steppers, I guess you may struggle with speed and response as it is. Without a high drive (power supply) voltage, the torque output of steppers quickly tails off, so the headline torque figure may be misleading.

Surprisingly it does Muzzer.! . . Esp with a stepper system on long screws and smaller steppers feel it the more and really struggle with it if the pulleys are heavy. This is why when building machines with long screw lengths and steppers, (over 4ft) I select a higher pitch screw and use a ratio to lower screw speed, this serves 2 purposes, lowers the chance of screw whip and lowers the inertia.
I also only fit Nema 34 motors which use high voltage drives, often these are at mains voltage depending on motor size 8nm + but for the smaller 34's I still wouldn't use drives less than 80Vac/100Vdc. The 4.5Nm and 5.5Nm Nema 24 Closed loop Lichuan motors I use now are so powerful running on 60Vac that I rarely fit Nema 34's below 8nm as the difference they allow in performance is much better.

[dazp1976]

And here I thought it was all down to inductance.



Which drivers are you considering?
I have the CL86Y or CLA86T from stepperonline on my list of options for this project.
I could run them directly off 60VAC

Need to do more research but the CL86Y atm.
I'm still pondering about going closed loop later down the line. If this can run in open loop as well (for the time being), then I'll get it right now.

Otherwise it'll be the DM860Y as the machine sits, and then I'll look again at a later date if I change to CL.

[JeffZ]

The 4.5Nm and 5.5Nm Nema 24 Closed loop Lichuan motors I use now are so powerful running on 60Vac that I rarely fit Nema 34's below 8nm as the difference they allow in performance is much better.

Are you referring to the LC60H2102, LC60H2112, LC60H2127 series motors?
They were my preferred choice initially, and I only removed them from my list of options because shipping was close to 200€ for 4 motors (from the Lichuan aliexpress store).
Are you driving them using the Lichuan LCDA86H drivers?
My original plan was to run 4xLC60H2127 + LCDA86H at 60VAC.

[JAZZCNC]

Are you referring to the LC60H2102, LC60H2112, LC60H2127 series motors?
They were my preferred choice initially, and I only removed them from my list of options because shipping was close to 200€ for 4 motors (from the Lichuan aliexpress store).
Are you driving them using the Lichuan LCDA86H drivers?
My original plan was to run 4xLC60H2127 + LCDA86H at 60VAC.

Yes, the LC60H2102 and LC60H2127 with LCDA86H drives for the Nema 24. For the NEMA 34's I use the 10Nm LC86H3156ZB 3 phase motor with LCDA2260E 220V drive.

I buy them in bulk (100sets) and have them custom built with 5mtr cables and also have the annoying 15pin Dsub connectors removed. I ship them via Sea so the cost is lower but yes shipping via Air freight is expensive these days.
For UK builders I offer to supply them, along with UCCNC controllers, WC-spindles, etc but given Brexit and the EU situation I don't really want to get into sending outside of uk due to export hassles.

[routercnc]

The other thing to think about is for an axis with double ballscrews you need to be able to square the gantry. This might mean being able to rotate one ballscrew very slightly by hand (without the pulley secured) until the gantry is square, so a key way or D flat would not work as this puts the pulley at a fixed location on the ballscrew. A simple pair of grub screws allows the pulley to sit at any angle on the ballscrew.

Also on the inertia side I did the calculations for machine design years ago and the mass of the gantry was a pretty small part of the motor requirements. The acceleration of the ballscrew was a large part of the motor requirement, and this in turn related to the ballscrew inertia. Inertia relates to having mass a long way from the axis of rotation, so heavy pulleys will have an effect. If you want to pocket them out then removing mass furthest out makes the biggest difference. And larger pulleys are significantly worse as inertia relates to the square of the radius so goes up quickly.

[dazp1976]

The other thing to think about is for an axis with double ballscrews you need to be able to square the gantry. This might mean being able to rotate one ballscrew very slightly by hand (without the pulley secured) until the gantry is square, so a key way or D flat would not work as this puts the pulley at a fixed location on the ballscrew. A simple pair of grub screws allows the pulley to sit at any angle on the ballscrew.

Also on the inertia side I did the calculations for machine design years ago and the mass of the gantry was a pretty small part of the motor requirements. The acceleration of the ballscrew was a large part of the motor requirement, and this in turn related to the ballscrew inertia. Inertia relates to having mass a long way from the axis of rotation, so heavy pulleys will have an effect. If you want to pocket them out then removing mass furthest out makes the biggest difference. And larger pulleys are significantly worse as inertia relates to the square of the radius so goes up quickly.

If I had the knowledge and tools I'd prefer to drill/tap a longish threaded hole into end of the ballscrew and bolt on a pilot bore pulley. Use say 5mm high tensile. Guess it depends on machine size wether it would hold.

[dazp1976]

Yes, the LC60H2127 with LCDA86H

I buy them in bulk (100sets) and have them custom built with 5mtr cables and also have the annoying 15pin Dsub connectors removed. I ship them via Sea so the cost is lower but yes shipping via Air freight is expensive these days.
For UK builders I offer to supply them, along with UCCNC controllers, WC-spindles, etc but given Brexit and the EU situation I don't really want to get into sending outside of uk due to export hassles.

Interesting. How do you feel about supplying just 2 of those 2127 sets?

[JAZZCNC]

Interesting. How do you feel about supplying just 2 of those 2127 sets?

Yes no problem, I'm low at the minute and can't spare the current stock I have but I have another batch arriving in around a week's or so (they are actually arriving in port about now so should clear customs and be with me in about 7 days of landing.)

I'll PM you when they arrive.

[JeffZ]

The other thing to think about is for an axis with double ballscrews you need to be able to square the gantry. This might mean being able to rotate one ballscrew very slightly by hand (without the pulley secured) until the gantry is square, so a key way or D flat would not work as this puts the pulley at a fixed location on the ballscrew. A simple pair of grub screws allows the pulley to sit at any angle on the ballscrew.

Shouldn't it be possible to auto-square using 2 drivers and adjustable endstops?

[johnsattuk]

The other thing to think about is for an axis with double ballscrews you need to be able to square the gantry. This might mean being able to rotate one ballscrew very slightly by hand (without the pulley secured) until the gantry is square,

Which is what I do with my twin screw router:thumsup:

[routercnc]

Shouldn't it be possible to auto-square using 2 drivers and adjustable endstops?
Yes but you likely need to square it when you first build it.

[JAZZCNC]

Shouldn't it be possible to auto-square using 2 drivers and adjustable endstops?

Yes and No.?

The problem with using the switches to get the square is you are applying twisting forces on the gantry/linear bearings/ball screw combination causing friction that will very quickly rob power and lead to stalls or missed steps and excessive wear.!.

You need to set the gantry as close to square as possible mechanically ie: loosening pulleys and plates etc so there are no twisting forces on the linear bearings and ball screws. So having setscrews rather than key ways allows you to turn the ball-screw and have it spin in the pulley for fine-tuning because the belt pitch won't allow fine-tuning.

Then if required only use the switches for the very fine tuning.

[JeffZ]

Yes and No.?

The problem with using the switches to get the square is you are applying twisting forces on the gantry/linear bearings/ball screw combination causing friction that will very quickly rob power and lead to stalls or missed steps and excessive wear.!.

You need to set the gantry as close to square as possible mechanically ie: loosening pulleys and plates etc so there are no twisting forces on the linear bearings and ball screws. So having setscrews rather than key ways allows you to turn the ball-screw and have it spin in the pulley for fine-tuning because the belt pitch won't allow fine-tuning.

Then if required only use the switches for the very fine tuning.

Ah, that makes sense. Thank you for explaining.

[dazp1976]

I did some more research and came across these types of clamping pulleys:
29868

What are your thoughts on these?

I'd give these a go on the ballscrew end.
Needs a bit more research but I've seen them on aliexpress.

https://www.aliexpress.com/item/4000311379356.html

Edit:
Amazon is much more £
https://www.amazon.co.uk/s?k=clamp+Timing+Pulley&rh=n%3A12465285031&ref=nb_sb_noss

[johnsattuk]

IMO The advantages of "taper-lok" and my DIY ER solution is that they are more compact, better concentricity, better grip, no damage to the shaft from grub screws, infinite ajustability and depending on pulley size less mass.