Tracking strips like grinding belts

[Kozzy]

Someone requested that I post part of a report I'm writing for work on tracking belts.  That report grew to be a little too complicated for a forum post but I decided to toss out some of the basics which might help people understand some of what's going on with their belt grinders.

First, assuming the belt is straight, all the pulleys are perfectly straight and square in the same plane, and the pressures are even across the whole belt, the darned thing automatically tracks.  There are no forces causing a belt to "un-track" so the belt would run like a dream.  However, this is the real world and nothing is straight, aligned perfectly, or evenly loaded.  That means what you are actually doing when "tracking" is compensating for a multitude of errors.  The first goal is to remove those errors where possible and reduce the affect of errors which they can't be eliminated.  As basic as that sounds, it really is important--for instance, a poor quality belt has a LOT more built-in error than a good quality belt so there is an obvious place one could apply the "reduce error in the first place" dictum.  Same kind of thing goes with poor quality or worn bearings, framework out of square (welding distortion is a big culprit on this), etc.  Chase things you can improve before you chase the issue of tracking or you'll be running in circles with "tweaks" to cover-up other problems.

When a belt passes over a very slightly skewed pulley, it will actually track to the HIGH side. Your brain might be saying it'll chase the path of least resistance on the low side but that's not what happens with slight misalignments.  If you look at the exaggerated drawing of a tapered pulley and the path of the "ribbon", you can see that the ribbon tries to tilt to match the skew of the pulley's face.  This causes the belt to push at an angle as it leaves the pulley and is what actually causes belt tracking.  Over several complete revolutions, the belt will try and climb the hill.

As the belt "climbs the hill", tension increases because it's path gets longer due to the skew of the pulley face.  When the belt is inelastic like a sanding belt, it will eventually reach a point where the tension is so high it has to slide back down at the same time the pulley is still trying to make it climb.  This is bad for the whole system.  It's a little like driving a car with one tire out of alignment against the road...which wears out the tire quickly and makes steering a bit wonky.  Because of this factor, the goal is to run with a little tension as possible.  People tend to want to run the opposite direction by applying as much tension as the system will tolerate and that actually interferes with the goal of tracking.

On a tapered pulley as shown, you actually have belt on both a left and right taper.  If there is more belt to the left, there will be more belt coming off the pulley skewed toward the right.  The belt will then enter the other pulleys in the system with this force and skew. The result of this is that the belt will try and track to the right over time (climb toward the center).  You can see by the drawing that the belt will eventually seek a center point where there are the same forces on the left as there are on the right. 

So why not used a simple tapered pulley?  Imagine the steering on your car having only 2 settings--fully turning left and fully turning right.  No in-between.  You could still effectively steer straight down the road but it'd be a constant battle of correcting.  Turn left and you immediately have to turn back right to keep the car going down the road straight.  When a belt is on a simple tapered pulley, much the same happens.  The result is that the belt will tend to dither back and forth, constantly trying to seek that center point and constantly over-correcting.

That's why crowned pulleys are used instead of tapered.  You get the exactly same tracking "skew" of the belt on the pulley surface that you would from a tapered pulley but it is variable rather than simple full-blast left/right.  That allows the belt to settle into a place where the forces are least at battle...which is the center of the crown.  If the belt happens to start wandering to one side, those tracking forces kick in very slightly to bring it back.  If it wanders more, the "taper" angle of the crown influencing the belt increases and tracking forces back to the center also increase.

Notice one other important aspect of the path of the ribbons in the drawing.  As each ribbon passes over the pulley, they skew a hair outward before coming back to the center and slightly angled as they leave the pulley.  Sanding belts don't get wider--so something has to happen to the belt in order to make that "effective" wider happen.  That something is called an "approach bulge".  The belt entering the pulley actually forms a bit of it's own crown and creates a bulge rather than laying flat--which allows it to effectively get a little wider as it passes around the pulley.  Too much approach bulge by having too much crown causes the belt to "work" the fibers which causes the paper or cloth backing to break down much more quickly.  

How much crown?  That question is one of the mysteries of the universe because there is no single right answer.  In theory, the accepted amount of crown to track the belt itself is quite small-only 1/5 the thickness of the belt difference in diameter from the center to the outer edge.  However, there is one more factor in play:  When sanding, the pressure on the belt is not even across the whole width of the belt.  If you sand on the left half of the belt, pressures there cause higher belt tensions on the left side and the belt will try to pull that way to compensate.  Because of this you need a crown which compensates not only for belt tracking but also for uneven belt tensions in use.  The greater the uneven pressures you intend to apply, the more crown that is needed to compensate.  It becomes a judgement call as to how much crown rather than a "rule".  I may get some argument here but I'd shoot for a 1/16" (1.58mm) diameter difference at the maximum in most cases of a 2" wide (about 50mm) belt.  That'd be effectively a 1/32" (about .79mm) total crown from center to edge of the belt at the contact surface. Narrow belts should get less crown but wide belts shouldn't get much more or that approach bulge problem starts to get out of hand.

Cheaper machines often have much more crown because it hides other alignment problems in manufacture.  It might be good to have the extra brute tracking force but it's not great on belts and actually can track a little less efficiently.

The ideal shape of the crown is roughly the center 60% of an ellipse. This sort of oddball shape gives much more balanced tracking as the belt goes off center with it's very light affect in the middle and increasing effect as the belt wanders.  Elliptical is a pretty tough shape to machine so most people go with a simple circular arc for a crown or some iteration of that which they can easily produce.  In some cases, what is called a "narrow-bodied roll" is used where the center 60% (roughly) is perfectly flat and there is a simple and slight taper at the edges to induce some tracking force should the belt wander.  Narrow-bodied rolls are easy on belts although they don't track quite as well as a crown.

Because there is a correlation between the tracking forces and the friction between a belt and pulley, the crowned pulley us usually located at the main drive.  That is the area of the highest dynamic tensions and is should be the perfectly square reference to which all the other pulleys are squared and aligned.

To actually affect tracking on the belt you use a secondary flat pulley.  You don't want the crowned pulley to be adjustable because it's already doing the tracking job it is designed to:  Making it adjustable (skewed off plane) only introduces some really weird forces on the belt as you've added skew AND crown together and those tend to fight.  That secondary pulley for tracking, by being able to be very slightly tilted, has much the same effect as a taper would against the belt in the drawing above--except you get to adjust whether that taper leans left or right.  Small changes can have large effects on a fast moving belt so you want to be able to finely tweak the skew of the tracking pulley when adjusting.  That generally means remembering to use a FINE thread on the adjuster rather than coarse.  Better yet, use a custom (non-standard) extra-fine-thread if you have the ability to machine your own in order to get the fine adjustments that are often needed.  Small tweaks, not large changes.  

The precentage of wrap of the belt on this tracking pulley will also affect how it influences tracking.  That gets a bit complicated in the theory realm so it's easier simply to do your adjustment and watch the result.  

That's a rough overview of why crowns track and a little about how to make the system work together but there is one more point to remember.  Sometimes things don't seem to be following the rules.  Life is like that.  You may have to break the rules a little to force the machine you are working with actually track a belt.  That's not "wrong", it's just adapting to the reality of your situation.  

 

 

[JHCC]

Very interesting. Thanks for spelling all of this out for us.

[Andy98]

Fantastic post. Thank you very much.

[SLAG]

Glenn,

I suggest that this post be pinned.

Kozzy thank you for sharing this information with us.

SLAG.

[Anachronist58]

On 12/15/2016 at 2:42 PM, Kozzy said:

Someone requested that I post...........

Wow. Four pages and three years back in this topic - what a gem.  Feel compelled to move it a bit closer to daylight......