Belts versus Direct Drive

[bsmith]

What are the advantages/disadvantages of belts versus direct drive systems, especially for long travels like an 8' on CNC routers? It seems that it would be relatively easy to calculate the innaccuracies of a direct driven system, but not so easy for a belt-driven one.

Belt driven systems include:

1) Belts that spin a nut along a fixed lead screw. Example: Tooltech at http://www.cyber-nc.com/hdprices.htm (see the photo at the bottom of the page) shows a nut driven design.

2) Belts that spin a lead screw. Example: http://www.data-cut.com/page5mv.html fourth picture down is a clear example.

3) Belts that take the place of a lead screw. (i.e. a long belt)

The intuitive problems with "1," belts that spin a nut along a fixed lead screw, have been pointed out earlier (moving mass of system, coupling etc). Though little or no data has been posted regarding the advantage of using the driven nut over the screw with respect to screw vibration/resonances and system inertia issues.

The problem with "3," belts that take the place of a lead screw, is that there's no decent information that I can find out there on long belts. Nothing close to the scope of what you can find on using a leadscrew for a precision application. In addition, this approach could possibly be high-cost, though I don't know for sure.

It seems that "2," belts that spin the lead screw, is possibly the least problematic in certain respects. It's perhaps the closest to the "direct drive" system layout and there's a good deal of information that exists on direct drive setups. However, what do we know about dynamic accuracy and precision of the belt and coupling parts of the drive system (leaving out for the moment innaccuracies of the screw itself)? It would seem like the tension, elasticity, coupling etc. analysis would be difficult, but some hard data is needed. Also, good sources of belts (with stated performance specs) would also be beneficial.

Another question: How much precision/speed is lost/gained when choosing these belt drives over direct drives? The above belt systems need to be weighed not only against eachother, but against other direct drive precision lead screws and rack and pinion drives. It would be appreciated if anyone could add informative data to the discussion, or to the FAQ section.

Also, related to this discussion, do stepper motor torque and performance specs change when directly coupled versus indirectly coupled as in a belt system? Is it preferred to "gear" the belt system to maximize the rpm/torque performance?

[HuFlungDung]

Hi Bsmith,

If you are really concerned about ultra-precision over long distances, I would recommend a system that uses linear scales for feedback. This sidesteps all the concerns that you raised above.

As far as I can tell, there is no guarantee of absolute position with belt drives of any sort. They tend to be quite repeatable, I would say, if the load is constant, because this means the stretch factor will always be the same. The accuracy of a common timing pulley drive is dependant on the precision of the outside diameter (actually called the pitch circle) of the toothed wheels. Contrary to appearance, the toothed belt does not provide the accuracy, it merely keeps the belt from creeping on the surface of the pulley. Whatever degree of freedom that exists because the belt tooth is narrower than the space cut in the wheel is the potential backlash amount. Belt tension tends to confuse this issue of timing belt backlash, because there may be sufficient tension for the belt to drive from the pitch circle, instead of "pulling hard" against the teeth of the wheel.

One advantage of the belt system, is that more often, better means are used to fix the individual driving members (the pulleys) to the shafts. I am thinking of taperlock bushings in particular as being the ideal fastening method.

Compare this to a direct drive, which in the cheapest instance, uses a straight solid coupling: but the keys and keyways are a slide fit! Potential for hidden system backlash is quite high, should the setscrews ever work loose, and they often will, on a system that is working hard.

So if you use a two piece direct coupling, then you have to be concerned about the compressability of the joint used in the coupling.

[bsmith]

Thanks for your excellent reply. So based on your experience, do you think one can achieve better performance from a belt drive or direct drive, if one is constrained by the use of readily availabile, moderately priced components? Are there readily available coupling mechanisms that effectively eliminate backlash and small errors in alignment when using a direct drive system?

And further, if I may paraphrase, you are saying that additional accuracy issues aren't raised if belt components are selected properly, and the main issue is the coupling mechanism between the motor and the screw and the potential problems raised by direct coupling in this regard. Are there any other advantages of either system from the motor torque, rpm, or reversing direction standpoints?

[ger21]

1) Our router at work uses a spinning nut on the X-axis. It's a high lead screw (probably around 1 tpi), so it doesn't spin very fast. But it is about 15' long and 2"D. I'm not sure if the spinning nut is to reduce inertia, or whip.

2) This seems to be the preferably way when using servos, due to the high rpm nature of servos. A benefit here is you can use a very high lead screw, and gear down the motor, so the screw spins slower and takes screw whip out of the equation. There is a lot of good belt info here, https://sdp-si.com/eStore/ . They'll send you free catalogs with a lot of technical info about belts regarding backlash, etc. I think the same info is also on the website somewhere. After reading, you'll probably want to go with GT2 belts and pulleys, which have virtually no backlash.

3) Have you seen this http://www.axxustech.com/Axxus4x8/axxus_4x8.htm Basically a rack and pinion, but I like this a little better. It depend on the loads involved if belt stretching would be an issue. But, if I recall correctly, most belts do virtually all there stretching very quickly when first used, and basically remain very constant after that.

[bsmith]

Thanks again, Gerry. This is excellent information -- much appreciated!

[Al_The_Man]

I have used timing belt and pulleys with very good accuracy, some modern timing belt use kevlar and are belted, just like the radial tires on your car, To decrease stretch .
The only thing I have found is that the teeth on aluminum timing pulleys wear over time and introduce backlash, If you want the supreme accuracy and reliability there is a product by GoodYear Called their Eagle PD lineof synchronous belts and sprockets and they are constructed of a Helical Offset Tooth design.
For long lengths as in a CNC gantry I prefer the rack and pinion, If the gantry has not a very large mass then I use belt reduction, But if the machine has a high mass gantry then for fast feed rates I use a precision gear box for reduction.
Al

[HuFlungDung]

Direct drive is best, I would say, if you are working in a setup environment where all the components are mounted on accurately machined surfaces. In such a situation, then it is okay to use a solid rigid coupling, and hardened keys accurately fitted to the shafts and coupling. It only takes a few seconds of maintenance time to check that the coupling is staying tightened properly to the shafts.

Under the above mounting conditions, it is quite a simple matter to use the coupling as the "alignment device", simply by fastening it tightly to the shafts, then loosening the motor mount screws a little bit, give the shaft a turn or two to let everything float to its natural position, and then retighten the bolts.

Flexible couplings are built to allow misalignment to a small degree. You just have to consider how any particular coupling design is built, to know whether it could introduce its own backlash to a system. Most couplings are built for single direction power transmission, with shock load considerations, so these would not make a good choice for ballscrew drives.

There is a type of helical coupling made which is essentially a solid coupling, with helical grooves cut through it, so that it looks like a spring. These have zero backlash, and will operate through a small angle of misalignment. I am not certain of "wind-up" with these units, though, so I can only recommend them for things like driving encoders and such. You'd have to find out elsewhere if large enough models are made to handle the torque of the machine you might attempt to drive with them.

Belt drives cannot be avoided sometimes when the space is limited, or a torque increase is desirable. Belt drives are quite a bit more forgiving in the shaft alignment, although you never want gross errors, It is often quite simple to make a common flat mounting plate to join both the motor and the ballscrew. Pop the pulleys on and the belt, and away you go.

If your machine is quite stiff to move, belt drives will stretch or slip a little upon reversal. The trick here is to use larger belt drive components to minimize the stretch factor. If your planned machine must move accurately back and forth in amounts less than .0005 inch, using rotary encoder feedback, then belt drives might be questionable. In such cases, using a linear scale for feedback will eliminate all forms of system backlash

If you were thinking about the long "rack belt" system accuracy, well, I know a local manufacturer who uses such a system on a "back and forth" process machine. He says it is highly repeatable, but in his application, absolute accuracy at any and all positions was not really a design factor.

[ynneb]

I have belt drive on my machine and to tell you the truth I wouldnt know how accurate it is.
But the question to ask is what are you goinf to cut, and does its accuracy really matter. If I cut a 6 foot frame and it measures 6 foot on my tape measure then that is all that matters. Those few thous of an inch dont matter. If you are only going to cut timber and its prescision is not important then go the cheaper belt option.

Sometimes I wonder, what is it that people are cutting that needs such prescision. Could someone give me an example?

[HuFlungDung]

Benny, stop raining on our parade here with those practical questions, wouldya?

The only place where I get really fussy, is the X axis of lathe. I like a precision of .00005" (or better), so that I can coax .0001 out of the diameter tweaks. This, BTW, is the cool thing about cnc lathe projects: accuracy that you can't even see is often the norm, and it matters for bearing fits, etc.

[Al_The_Man]

Benny, I agree, most gantry's of 6ft and over are used in wood and plasma, gas cutting etc where a few thou is not regarded as an error.
Larken (see web site) routers here in Canada have an excellent machine and they achieve very good accuracy using steppers, rack and pinion with belt reduction.
For very large machining centers then accuracy is usually an issue and the often helical gear rack and pinion is used, the size of the machine justifies the cost.
Al

[bsmith]

Are there sources for long, precision helical racks (about 8' feet or so)? This would seem to be better than spur racks from the standpoint of improved contact and backlash. And.. less complicated than a belt system.

Also, I found this source for rotating nut assemblies (part of Thomson, I believe): http://www.apimotion.com/PDFs/Techni...s_Tech_Bul.pdf

Also, I agree that precision is usually not required for the most typical uses of large-bed CNCs. Furniture, architectural work, cabinetry and large paneling does not need .001" accuracy. On the other hand, if one is fabricated relatively small parts cut out of large panels, and the parts require precision in their assembly, then even .003"/ft isn't all that extreme of a requirement.

[Al_The_Man]

Usually if you can keep tension on the pinion, there should be no backlash on a good spur rack, I would think the helical would be pretty expensive.
(see this thread) http://cnczone.com/forums/showthread...d=4596#newpost
Al