Benefits of Belt Drives for Steppers

Hi All, first post here. Have been CNC for a few years now and decided to build myself a new machine from scratch. Have been educating myself and in the area of motors and drives, had an unexpected observation. Like many others, I too am attracted to the import lead screws for their economy. In the speeds for me, 5mm and 10mm lead screws need to turn at 500-1000rpm. So I start studying stepper torque curves and see the dramatic reductions in torque and then see that many DIY builders incorporate timing belt drive in an attempt to operate the stepper in a range of higher torque. So I started designing a drive to incorporate such but as I was comparing the (torque) benefit belt vs direct drive, I was surprised to find none.

I considered 3:1 timing pulley gain from stepper to screw. The 3:1 ratio means the torque at the screw is now 1/3 the stepper shaft just due to pulley ratio. It turns out, my stepper torque was about 1/3 at 450rpm, so there was absolutely no (torque) benefit in drive screw motive force compared to direct drive. In some cases, the belt driven screw still offers some packaging convenience, but not performance, unless possibly you have more dramatic gain/reduction through the transmission. For grins I checked at 6:1 and there was still little to no gain with either approach. In fact, the reduction instepper torque seemed to track the inverse gain/reduction through the timing pully ratio throughout the speed range. The same held true if you are stepping down the stepper speed. Although you now get an increase in torque delivered at the slower turning screw, the stepper suffers a similar reduction in torque at the higher operating speed, so little to nothing is gained.

Now, if you have a (more) constant torque motor/servo, all bets are off, and this of course is not the case. This doesn't consider any benefits (or detriments) for resloution, but with microstepping there are many options and the affects seem somewhat subtle as far as torque goes. I've also been cautioned about the potential for lead screw whip at higher speeds but it's hard for me to see that as much of a concern with 20mm screw andmax unsupported length of say 750mm.

I guess the moral of the story is: Power is the rate at which you make torque and once you set the speed of the lead screw you are not going to alter the motor power with a transmission, and in the case of steppers, my observation is neither the delivered the torque, yet I see all these DIY CNC routers equipped with belt drives for this reason. What am I missing?

It was sort of a deflating revelation for me. I've opted for 10mm lead screws and at the moment am planning to direct drive everything except possibly one axis, and that would be for packaging, not drive performance.

Best,
Kelly

Like a lot of things, people use something just because others have used something, but as you've realised, that doesn't mean it's the best something.

However it looks like you've only considered torque, which is only one part of the equation.
You also need to consider inertia and acceleration.

Ultimately what is best, depends on balancing acceleration, maximum speed, accuracy, and any physical constraints (i.e. where the motors are going to mount).


Another thing is don't rely on microstepping for accuracy. When microstepping, the motor position is essentially balanced between two electromagnet springs, so load can push it off the microstep position. I'd personally base accuracy around full steps, but half step accuracy is still a reasonable assumption.

Quote:

---

Originally Posted by m_c

.....However it looks like you've only considered torque, which is only one part of the equation.You also need to consider inertia and acceleration.

---

I did consider acceleration for the build and have went to lengths to make the trades to minimize moving mass while maintaining rigidity for my work envelop. I just didn't delve into it in this post because I viewed any differences in total moving mass due to motor selection to be a small factor in total system mass/inertia, so for the most part, system mass decisions were independent of motor selection other than selecting motor size/power of course. Value for money was another trade that landed me on closed loop NEMA34 steppers, as opposed to the next step to servos.

These things aside, for any given speed, the reliably acheivable acceleration is a very strong function of motor torque. I found that most motor manufacture's rule of thumb was keeping system inertia <10x that of motor rotor inertia, and in some cases for very high speed system the recommendation was <4x. In the end, it's an easily tunable parameter and unless your average tool path lengths are very short, you don't pay much of a price in run time for backing off a little on acceleration.

I studied a lot builds here and elsewhere as they relate to my machne build and was going to start a build thread but I dont think there is anything particularly novel about the design I settled on, and chose to focus more on the construction methods to optimize the mass/rigidity trades.

Thanks for the post,

Best,
Kelly

Well thought out! Most constructors seem to be fixated on the headline (stall) torque and/or don't consider the fall off of torque with speed. Of course, if you gear your stepper down 2:1 or more, your ballscrews etc need to be able to withstand the resulting force when you inevitably run against the stops. To me, if you go stepper, you might a well go direct drive. Conversely, if you go servo, you might as well use a reduction ratio, as most ballscrews aren't happy at 3000rpm+ and you generally get less torque from a servo to start with.

One moment of inertia mistake I made was to retain the large cast iron handwheels on my first (Bridgeport) conversion. I thought it would be clever and sensible to retain the possibility of manual operation but once I had got used to an MPG, I soon dropped that idea. And once I'd calculated the MOI of the system (mass of table, saddle, vise and workpiece, referred back to an equivalent MOI at the motor) and compared it with that of the handwheel, I could see how much they messed up the response. So if you insist on retaining handwheels, go for small and/or plastic ones.

Quote:

---

Originally Posted by Tool-n-Around

I studied a lot builds here and elsewhere as they relate to my machne build and was going to start a build thread but I dont think there is anything particularly novel about the design I settled on, and chose to focus more on the construction methods to optimize the mass/rigidity trades.

---

Hi Kelly welcome to the forum, this is exactly why you should share/start a build log, your design doesn't have to be novel but it will be interesting to allot of us, we can focus on the construction aspects together and maybe learn from each other too :)

Are you building a moving gantry type machine or mill?

MACHINES
Discussion related to machines, working with them and building your own.

Quote:

---

Originally Posted by Lee Roberts

Are you building a moving gantry type machine or mill?.

---

Gantry. Considered Moving and Fixed Gantry, and Moving Table, but moving gantry seemed to best fit my goals.

Quote:

---

Originally Posted by Lee Roberts

Hi Kelly welcome to the forum, this is exactly why you should share/start a build log, your design doesn't have to be novel but it will be interesting to allot of us, we can focus on the construction aspects together and maybe learn from each other too :).

---

I'll see what I can do but am pretty far along. The Z is built. Most of the linear motion components are in hand or on order and design fairly well committed. Motors, drives, machine controller not yet committed but fairly far along yet I must confess, still learning.and plenty of room for learning and input there. Here's teaser on the Z.

https://youtu.be/v1X02ZWP2mE

Best,
Kelly

I always use belt reductions to get 5um per half step, tuck the motors out of the way and keep the hand wheels. I round every movement out to 0.005mm. Works for me.
Thing is my latest mill conversion has steppers and DRO's. Not a problem unless the stepper controller can read the DRO's. I couldn't resist...
What will I have it do when the motor disagrees with the DRO by >0.01? :confusion: :frog: