Let's talk belt grinders

[HennieL]

In your search of those millions of videos, did you ever find a process that works for repairing belts that are brand new but which have non-functional splices (i.e. that break)? I've tried hide glue, super strong fiber tape, layers of this and that, etc... But everything seems to break after not all that much time. Unfortunately, I bought a bunch of surplus belts, so I would really like to use those.

That's probably why the belts were sold as surplus in the first place - long past their "use-by" date

In my experience, it's not advisable to re-join belts purely due to safety - unless you are using them on a belt grinder with speed control turned all the way down... and then you might as well just hand sand. A belt that's running at optimum grinding speed can cause quite bad injuries when it breaks, and can even result in blindness if the tip hits your eye.

Another reason for not re-joining belts is that the "bump" created by the DIY joint is invariably thicker than that of the original belt, and the vibration caused by this leads to inaccurate grinding, and even faster wearing-out of the newly made splice.

Best to use them as hand-sanding belts as stated by BaronJ above.

 

[BaronJ]

Hi Hennie, Guys,

I've not had any real issues with the belts that I've glued, yes there is a slight bump as the joint comes round, but you get that to some degree with a new one.

I find that if I clamp the glued ends of them in a vice overnight they don't easily come unstuck. I use a scalpel to shave off the dried glue that oozes out of the edges of the join on the sides of the belt.

I think people use too much glue thinking that a thick layer will hold better !

 

[awake]

Time for the plans for a couple of rather simple parts ... which may be the only unique part of my build - the tensioner bracket and tensioner pivot:

What makes this unique? The provision for setting the angle of the bracket on tension arm using two set screws. This turns out to be a very simple solution to a problem that I have not seen addressed on any of the many videos and articles I have perused. I can't say for sure that no one else came up with this idea first; it is such a simple idea that it seems hard to believe it hasn't already been used. But not only have I not seen this idea in any of my research; I haven't seen any real solution to enable precisely setting the angle of the tensioner mechanism. Of course, every design includes a bolt or some such that applies against the pivot plate to set the angle in the vertical plane, but the problem is, any rotation of the tensioning mechanism on the horizontal plane (where it attaches to the tension arm) has a VERY large effect on tracking - far more effect for a tiny change than adjusting the pivot angle does.

The typical solutions are either to "be very sure to attach the tension mechanism straight with regard to the tension arm," or "spend time very carefully nudging the bracket back and forth until you get it just right." (See Brian House's video on tuning the Revolution belt grinder as an example of the latter approach.)

Given how sensitive the tracking is to this adjustment, I wanted a precise and controllable way to adjust the angle. What I came up with above is super simple - as shown in the picture on Sheet 7, the tension mechanism spans the tension arm with around 1/8" of space on each side; on one side this allows room for the tensioner plate to swivel on its hinge (vertical plane), allowing adjustment of tracking using the 3/8-24 bolt that goes through the tension arm and bears on the tensioner plate. On the other side, this 1/8" space allows room for the 1/4" set screws (grub screw) to be adjusted in or out, creating a rotation of the bracket in the horizontal plate, swiveling on the 3/8-16 bolt that attaches it to the tension arm.

In practice, this design decision has worked exceptionally well. If nothing else from my design is useful to anyone, I highly recommend using this idea, or coming up with something else that will allow for similar adjustment!

 

[awake]

Now for the tracking wheel and its spindle:

There is nothing unusual about the tracking wheel itself, and many will prefer to buy one rather than make one. The main design decision here is how much crown to use on the wheel. I have no idea what the optimal crown may be; I simply started out with the crown as shown, in which the diameter at the center is approximately .060" (1.5mm) larger than the diameter at the ends. (If someone prefers, I can put in the radius of that curve ... but in terms of my own manufacturing, that dimension was not or will not be particularly useful.) I figured if that did not seem to work well, I would experiment with other dimensions for the crown ... but it does seem to work just fine, so I have not tried anything else.

In terms of the spindle arrangement, I have gone a route that is rather different from most that I have seen. Rather than simply using a bolt of the appropriate diameter, and using a ny-lock nut to secure the wheel on the bolt, I have made a spindle that first attaches firmly to the pivot plate via a 1/2-20 thread, and then holds the wheel snugly in place using a "cap" and a 1/4-20 button head screw. This approach would not make much sense for anyone who does not have a lathe to size the length of the spindle and cap combination to just the right length - but with a metal lathe, it is surprisingly easy to get this just right.

I made the wheel first and pressed in the bearings (more about this below). Then I made the spindle out of some .750" hex stock; I made it to the dimensions shown on the drawing, tweaking the diameter for a snug sliding fit on the inner race of the 6201 bearings that I chose to use. (Of course, you could use a different bearing, and adjust the dimensions of spindle, cap, and wheel appropriately.) Finally, I made the cap, leaving the "shaft" portion a little bit long. Before parting off, I put the assembled wheel on cap and slid in the spindle; then while holding the spindle firmly against the end of the cap, I could slide the wheel assembly back and forth to see how much play I had. I carefully skimmed off the excess until I achieved a fit without any play - actually, I aimed for a couple of thousands of an inch of very light pre-loading of the bearings. Then part off the cap, and voila! A lovely spindle.

A word about the tracking wheel that I am actually using for the moment - I plan to make the tracking wheel out of aluminum, just as soon as I am able to cast one up (or as soon as I find a suitable chunk of aluminum to machine to size). To get started, however, I thought I'd try 3d printing a prototype tracking wheel, in part to give me a chance to try different settings for the crown. I didn't really expect it to last very long, but thus far, the printed tracking wheel has held up just fine in mostly light duty use - all that I have done so far on this belt grinder. (I also printed the idler wheels for the D-plate arrangement ... and I've worn out one set of those, when the bearings got hot enough to allow the bearing seats to distort. Not sure if that indicates too much pre-loading, or just the expected heating of the bearings, but I'll come back to this discussion when I get to the D-plate.)

I've attached the .stl file that I generated and used to 3d print the tracking wheel. I've also attached the OpenSCAD program used to generate this .stl file (contained in the roller.zip attachment, since the forum does not allow uploading a file with .scad suffix). Note that this program requires some OpenSCAD libraries that I have developed; these are contained in the libraries.zip attachment. (You will need to unzip the libraries into your OpenSCAD libraries folder.) The roller.scad program can generate a variety of wheels, with crown or without, for any size of wheel (diameter, length, bevels), bearing, crown, and so on - just change the appropriate parameters, which hopefully will be obvious in the program if you are at all familiar with OpenSCAD. Note that all measurements in this program are intended as mm. FYI, on my printer, printing this in PLA using 20% infill and .25mm layer height, setting the clearance to .1mm radially (.2mm or .008" for the diameter) gave me a nice light press fit for the 6201 bearings - this setting will no doubt need to be tweaked depending on the printer, filament, and layer height.

 

[awake]

Here at last is the next installment in the belt grinder build. I apologize for the long delay - it has been almost exactly 3 months since my last post on this! Too much work, too little time for hobbies ... obviously, the world is very messed up.

Here is the motor mount:

Nothing very exciting here, except that this is the part most likely to have to be customized to suit the particular treadmill motor that you are using. The motor I am using has a u-shaped bracket that creates two "ears" that stick out from the motor, 3.5" apart:

Each "ear" is approximately 5.5" wide, with mounting holes that are 4.5" apart. The motor is not centered over the mounting holes, but rather is offset .5" to one side. To say it another way, with the motor mounted sideways as shown above, the mounting holes at the top of each ear are 2" above the centerline of the motor, but the mounting holes at the bottom of each ear (not visible in the picture above) are 2.5" below the centerline of the motor.

The fixed mounting plate shown in the plan is welded so as to be centered on the main tube assembly vertically and flush with one edge (note the edge indicate in the plans). The offsets in the placement of the slots in the adjustable mounting plate shown in the plan, plus the off-center placement of the mounting holes on the motor bracket, allow me to choose whether the motor is mounted above, on, or below the centerline of the main tube assembly, simply by flipping either or both the U-bracket on the motor or the adjustable mounting plate upside down.

This was the key design decision that I wrestled with - reviewing various other designs, it seemed that the motor was sometimes mounted above or below the centerline, and I wasn't sure how much the placement would affect the balance when the grinder is rotated from vertical to horizontal. I wound up situating the motor centered on the main tube assembly, and the balance seems to be perfect.

At the risk of stating the obvious, the slots in the adjustable mount let me adjust the motor left or right, and allow enough clearance to twist just a bit either way, so that on assembly the drive wheel mounted on the motor spindle can be adjusted to precisely the same plane as the platen wheels - this is important to achieve smooth tracking.