There has been a lot of discussion of belts lately. I found there is a ton of information at the Gates site here:
http://www.gates.com/facts/index.cfm?show=Engineering&location_id=3258

(note you can see more info by selecting from the four choices: maintenance, installation, engineering, Belts101 near the top of the page).

One thing I stumbled on that was particularly interesting to me:

http://www.gates.com/facts/documents/Gf000205.pdf

It provides a means for checking for groove wear without a sheave groove wear gage. The measurements are cumbersome to take, will require some calculations, and in the end will probably be much less accurate than using a sheave groove wear gage.

So, it's not exactly practical. Rusty has mentioned before that groove wear gages are practically free (although I'm embarassed to say that we still don't have any .... we just eyeball the grooves and check how the new belts fit in them).

But even if not exactly practical ...who cares. Forum posts are free so I'll post it anyway on the off-chance that it might be useful to someone at sometime for some reason.

The method described at the link above involves getting two spheres (for example from ball bearings) of different sizes that both fit in the groove without touching the bottom. You take a meassurement of the difference in height of those two balls when placed in the groove. That measurement could be a little tricky... they suggest building a little jig to sit on the grooves and provide a surface for depth gage.. I wonder if you could just use a dial indicator to compare the difference when the two balls are placed in there... or is it too inaccurate to try to get a dial indicator to ride exactly on top of a ball ?

Once you have the measurement, you plug it into the equations to determine "alpha" = sheave groove angle and "b" = sheave groove opening width (assuming the edges were straight). Then you can compare those values to the standard values for your groove style which can be obtained in various sources such as Machinery's Handbook, 27th edition, page 2395, Table 2A. (for example 5V groove, 8" sheave would have b=0.6" and alpha = 40 degrees). Also the notes at the end of 2A give some tolerances on these measurements for new sheaves, which could be a starting point for judging what is bad for a worn groove.

I drew the geometry out on the attached powerpoint because I wanted to verify them for myself (and maybe because I have too much time on my hands?). The result shows that the formula's are correct. One thing about the derivation is that it assumes the groove sides are perfectly straight... which of course is not true if your sides are dished out... which is what we are looking for to begin with. It may seem a little hokey to use that formula based on straight groove walls as a basis for a check when the grooves have become curved, but after thinking about it for awhile, I'm pretty sure that wear including curved walls are going to show up as a change in at least one of those calculated parameters (alpha or b).

So what do you guys think? Is that approach absolutely useless? Or might be useful in a pinch? Would a dial indicator work?

And... do you have any other tricks for eyeballing sheave wear if you don't happen to have a wear gage?

This message has been edited. Last edited by: electricpete,

MeasureSheaveGrooveWithTwoSpheresR1.ppt (76 Kb, 21 downloads)

[QUOTE] ]we just eyeball the grooves and check how the new belts fit in them).

all sounds a bit complex to me, 30 years on the tools for me, so i'll go the old school way as quoted

Any flat metal straight edge, like an Exacto Knife, combined with a flashlight can be used to see a worn or uneven sheave groove surface. Only a template tool can verify the groove angle and detect subtle wear. If you own machines with belt sheaves, then buy the groove templates.

Walt

Assuming your belt alignment is reasonable (i.e., you can't actually "see" any misalignment with the naked eye), and drive arrangement is adequate (powerwise)....

If you install new belts and properly "retension" them, only to find them slipping after a few weeks.....

If you are constantly replacing belts on a machine because they glaze over or tear up in a matter of months.....

If you have noticeable "belt-induced vibration" (harmonics of the belt rotational speed), even with new belts installed.....

...then you don't need to measure the sheave wear.... the sheave wear is excessive... replace them. It really IS that simple.

thanks everyone.

We periodically replace belts but not sheaves. We haven't seen unusual problems with squealing or glazing. But various vibration higher than we'd like, including belt speed/ belt speed harmonics even with new belts.

That pattern (belt speed) can indicate worn sheave? I never would have made that connection.

When this is a $ driven industry the cost of sheave guages vs:

The method described at the link above involves getting two spheres (for example from ball bearings) of different sizes that both fit in the groove without touching the bottom. You take a meassurement of the difference in height of those two balls when placed in the groove. That measurement could be a little tricky... they suggest building a little jig to sit on the grooves and provide a surface for depth gage.. I wonder if you could just use a dial indicator to compare the difference when the two balls are placed in there... or is it too inaccurate to try to get a dial indicator to ride exactly on top of a ball ?

Once you have the measurement, you plug it into the equations to determine "alpha" = sheave groove angle and "b" = sheave groove opening width (assuming the edges were straight). Then you can compare those values to the standard values for your groove style which can be obtained in various sources such as Machinery's Handbook, 27th edition, page 2395, Table 2A. (for example 5V groove, 8" sheave would have b=0.6" and alpha = 40 degrees). Also the notes at the end of 2A give some tolerances on these measurements for new sheaves, which could be a starting point for judging what is bad for a worn groove.

I drew the geometry out on the attached powerpoint because I wanted to verify them for myself (and maybe because I have too much time on my hands?). The result shows that the formula's are correct. One thing about the derivation is that it assumes the groove sides are perfectly straight... which of course is not true if your sides are dished out... which is what we are looking for to begin with. It may seem a little hokey to use that formula based on straight groove walls as a basis for a check when the grooves have become curved, but after thinking about it for awhile, I'm pretty sure that wear including curved walls are going to show up as a change in at least one of those calculated parameters (alpha or b).

I probably couldn't walk out to my truck for those tools in the time it would take to pull the guages out of my back pocket

With that said.
How much wear is too much? That's what I would really like to know.

Robert U.,:"How much wear is too much? That's what I would really like to know."
First, obtain the proper wear guages for the type of V-belt sheaves you will inspect.
I was told to recommend sheave replacement if you can see a 1/32" gap anywhere between the walls of the sheave and the guage. Therefore, if any part of the sheave groove is worn 1/32 inch, replace the sheave. Does this amount hold true for any size belt? I would think not. If in doubt...Rustythevibeguy (vibguy?) is right, use everything to judge.

quote:

But various vibration higher than we'd like, including belt speed/ belt speed harmonics even with new belts.

That pattern (belt speed) can indicate worn sheave? I never would have made that connection.

Whenever I found 2X belt frequency I was fond of a quick test with my second favorite tool, a dial indicator. I'd MOunt the indictor base to ground (Not the local bearing or motor support), with the indicator stem first on the motor and then on the fan bearing inline with the belt pull. The I'd Roll the motor by hand and look for sudden low spot(s), and mark the belt back in both the motor oand fan sheave grooves with chalk.

If the belt has a fat or thick spot (quite common with wrapped belts) then the motor and fan will be pulled towards each other by higher belt tension when the thick spot is in a sheave groove groove. If I can see a few thousandts (MILs) of motion at 2 rpm, I figure there will be at least that much at 300 or 400 cpm (or whatever 2X belt rpm is).
Fan structure designers and adjustable motor base designers may be pretty far removed from the service guys who get to battle to correct the "excessive vibration" warrantee claims. Similarly, they all may be pretty far removed from the 1 or 2 people in the organization that understand bearing loads, and what happens when an eccentric lock bearing mounted on "industry standard" sized commercial shafting gets a shot of extra radial load, and belt tension stiffness.
The toothed belts with cut sides run a bit cooler (especially around smallish sheaves), and were much less likely to have thick spots, so were my favorite for sensitive equipment.

However, the Gates heavy duty green wrapped belts are very rugged in applications that use idlers to apply belt tension for clutch action, like riding lawn mowers. I never took vibration measurements on a lawn tractor.

What about powerband belts?

All of these methods which have been discussed may be great on a motor or component that is not running. But, as a vibration guy, how do you check a running sheave on a piece of equipment that may only be down once or twice a year?

Just a rule of thumb, but what I look for is belt depth...if the top surface of the v-belt is running below the top of the ribs on the sheave then the sheave is most likely worn to the point of needing to be replaced. Also if the bottom of the groove is shining brightly then most likely the sheave is worn to the point where the v-belt is riding too deep in the groove. The only cause of which could be worn belts or a worn sheave.

This is just some thoughts on what works for me. This is taking into consideration that the guards allow for visual inspection.

Travis