27 x 64 Belt Drive Machine
|
A
larger faster
machine |
 This
belt drive table was built to address the speed and
size
limitations of cheap leadscrews.Attributes include:
 The machine is made of lumber and 1/8 and 1/4 inch
aluminum stock. This metal does not require special tools to work. Hack
saws or carbide blades will cut it, but the metal is not as rigid as
steel or thicker aluminum. Because of this strength limitation, this machine is designed to use a trim router. The machine will flex or chatter before a full sized router is pushed to its potential. However, the machine can support a full sized router so that larger bits can be used. The work on the left was carved with a 2.5 hp Makita router with a 1/2 x 2-1/4 inch double fluted bit. The stock is SPF framing lumber and the pattern is 16 x 6 x 1 inch deep. The cuts were made at 100 ipm at 1/4 inch deep per pass. This depth of cut worked well; deeper cutting at this speed, with this bit, caused chatter. The pattern was completely cut in less than 5 minutes with four roughing passes. A photo of the final product is at the bottom of this page; here is a 40 second video. The HobbyCNC 200 oz.in. system was also tested. With slightly different gearing on the X axis, it was able to rapid at 200 rather than 300 ipm while supplying adequate cutting forces. The different gearing is addressed in the plans. The HobbyCNC 305 oz.in. system performs similarly to the Xylotex 269 oz.in. system. ..............................
Belts are forgiving of misalignment and are less likely to ratchet or
skip teeth as compared to some rack systems. However, they can rebound
under
high accelerations, which can cause the
router bit to leave a wiggly trail as it bounces back into line. This is resolved by slowing the stepper's acceleration in the software. The feed-rate can remain high, so the stock can still be cut quickly. The longer the belt, the more rebound there will be. Long belts are costly, and all belts wear with use. The heavier the use, the sooner the belts will need to be replaced. Rack and pinion, and leadscrew drive machines are not as efficient as belt drive machines, but the components usually last longer. This machine, like the others in these plans, was designed to be home-shop friendly. However, it is the most complex machine in the 5 plans set, and building would be difficult without a dedicated shop space. The machines
were all designed to use the stepper and drive systems from Xylotex and
HobbyCNC. The recommended software is Mach3 or TurboCNC.
These suppliers work with the Do It Yourself market, and supply documentation and help forums for their products. .....................
|
| A
background of the
design considerations To achieve the size and strength necessary it was decided to use aluminum for all of the gantry components. Thin stock of 1/8 and 1/4 inch was used because of the cost and ease of cutting with home shop equipment. Thicker stock increases the price and is more challenging to work. This initial choice of thin aluminum was the guiding factor around which other design decisions were based. From the 4 x 8 foot metal table shared on this site, it was clear that a full sized router could not be pushed to its potential with 1/4 and 1/8 inch aluminum components; therefore a trim router was chosen as the spindle. The rail systems are pipe, conduit and skate bearings.These are less costly than commercial rails while still giving the solidity and accuracy required for most wood projects.  The ideas for tensioning, clamping and routing the belts
are simple but commercial systems are costly and/or difficult for the DIY builder to purchase. Therefore, these parts had to be fabricated from off-the-shelf components. The clamps and tensioners were straightforward to create, but the idler pulleys had to use unrelated but commonly found components. Right photo. It is possible to simply use bearings rather than flanged idlers to route the belts, but derailments do happen for no apparent reason during long cuts, so an ounce of prevention is worth the effort. Also smaller diameter bearing idlers cause premature belt wear. Belts require gearing between the stepper and drive pulley to make the axis move in concert with the stepper and the router-spindle. These extra pulleys add complexity and cost, but direct drive does not work with the sizes of steppers, belts and pulleys that are readily available at lower prices. This gearing was a trial and error process. The ratios have to be a compromise that balances resolution and speed against cutting forces. Deriving the results required a fair amount of time and money. The belts were also somewhat of a challenge. The manufacturer's specifications imply that the belts are quite strong, but buried in the fine print it is noted that designs should incorporate a 15 to 1 safety margin. Therefore, it was necessary to build a prototype and abuse it until something broke, and to then upgrade the failed components to the next largest easily found size, and repeat the process. There is a relatively narrow range of tension where the belt allows minimal backlash while not overloading the bearings, or causing premature belt failure from too much tension. The adjustment systems on this table are simple to set, so this tension can be found by trial and error on each axis. A tensioning meter is unnecessary to find a "good enough" tension. Chains in place of belts were tested on this machine. The cost of comparable sprockets and chains is less than pulleys and belts, but the results were so unsatisfactory that the idea was abandoned soon after the testing began. Chains proved to be less efficient than belts; the table's rapids decreased significantly, the backlash did not improve. Tensioning the chains enough to improve accuracy slowed the table even further. Others have warned of chain drive; my experience is the same. There are better methods such as belt drive or rack and pinion for the same effort and just a little more cost. The plans for this machine are included in the 5 Plan set. |
 |