LONG Timing Belt application question 3

[tygerdawg]

I am designing a single axis linear drive for a client. For several important design reasons I wish to use timing belts if possible. My fallback position will be rack & pinion, but that causes some other problems.

It's a strange application: 57 ft pulley-to-pulley distance, open-ended belt clamped to a carriage in linear bearings, carriage attached to a 6500 lbs buggy on flanged casters, reciprocating back/forth, +/-0.25in position tolerace, slow speeds and accels. I'll drive it with as big a gearmotor as needed.

I've done an exhaustive analysis of torques, inertia, horsepower, tension loads, etc. Several buggy sizes will be towed by the carriage, so I have a good understanding of the range of torques required and have made system assumptions and adjustments to make this work.

Now I'm at the point of selecting a belt and have hit a brick wall. I've contacted Gates, Goodyear, Carlisle, and Mectrol and have gotten nowhere with these guys except for Mectrol. All except Mectrol ignore me because in their words "no one knows how to calculate the tensile limits on slow moving belts". It seems their entire market is constantly moving endless belts like on your car and my app is so oddball that no one wants to touch it. What I have done is iteratively plug my numbers into each company's design calculator and generate strength, rating values, and recommended part numbers as if my app were endless belts of varying length. From this calculated data and their catalogs, I try to infer working tensile loads, tooth shear strengths, and horsepower capacity from the results. All the sizing exercises in the catalogs are very simplistic and are based on RPM/Torque/HP/length/pitch and constantly rotating systems. The results from all vendor sizing calculators are encouraging, but inconsistent enough with my own calculations to cause me concern. I've narrowed my choice to a certain pitch, tooth style, material, and width. But without explicit confirmation, I'm leary.

My gut tells me that if I put a big enough honkin' belt on this system, it will pull the load. But no vendor will commit to an answer, either "use this one" OR "our belts won't work". And I'm tired of teaching mechanical engineering to application engineers.

Can anybody out there provide any insight?

Thanks in advance

TygerDawg

 

[diamondjim]

Tyger Dawg,
Interesting application. I would think the
long length would be to your advantage. I
assume the strength is based on the webs between
the cogs or teeth on the belt. The pull would
be the same so the length would help to reduce
the strain on the cross section.

 

[Barry1961]

I have used a bunch of the Gates PolyChain in 40'+ applications in the past and it is very hard to get that information. If I remember the rule was that you had to grab at least 4 teeth to break the belt. Less than 4 teeth and the Kevlar cords would just pull out of the ureathane body. I did finally get a Gates engineer to give me some numbers but I was sworn to secrecy, so if I told you I would have to kill you. With this in mind read on at your own risk.

Barry1961

********************************************************

What you have described is just linear belt application, for which we
normally prefer to specify an allowable working tension for normal
loading and then use the UTS to calculate a safety factor when e-stops
are very unusual load conditions. With that said....

The allowable working tension for 14mm Poly chain is dependent on the
sprocket size and speed, but as a rule of thumb, you can use 835 lb for
20mm wide and 1545 lbs for 37mm wide (you can go ahead and just scale
the #'s beyond this). These #'s are extremely conservative as stated
and I would be happy to get more specific with additional information.
Regarding the breaking strength, you can use 5650 lbs per inch of belt
top width.

I hope that this helps, and please feel free to contact me with any
additional questions

Ben Smith
Application Engineer
Gates Corporation
Phone (303) 744 5626
Fax (303) 744 4600

 

[rzrbk]

Several sliding automatic gates use sprocket and timing chain arrangement involving both long length and slow speed. A chain would also seem less prone to stretch.

Also cable and drum is used on short ski lifts although that would not have the timing locked in like you are looking for.

 

[Barry1961]

After a certian width they won't spiral cut belt because the slant of the tooth gets too great. If I remember the max length of a 37MM width spiral cut was around 150'.

Barry1961

 

[Compositepro]

It seems to me that belt stretch and positional accuracy would be a much bigger issue than strength. To get enough stiffness you will have far more strength than you need. 0.25" is 0.036% of 57 feet. You will need a position control that will compensate for the effects of stretch or your system will not be stable and will oscillate.

 

[unclesyd]

What is the travel needed for the carriage?

 

[btrueblood]

I'd agree with compositepro. I worked with an optical comparator that had small timing belts to drive x-y table axes. These belts were endless, but had ~100:1 l/w ratio. The damn thing was nearly impossible to position repeatably to within +/- 1/16" with near zero speed and load; too much stretchiness in the belts. Timing chains may get you closer.

 

[tygerdawg]

Thanx for your comments Y'all.

I guess I should revise my positioning tolerance. It's a sloppy painting process. It could be +/-.75" or maybe even 1" I suppose. This might open things up a bit.

Diamond, the strength of these animals is two-fold: the shear strength of the material across the root of the tooth (and the quantity and width thereof engaged in a pulley) and the tensile capacity of the tensile members. Hence the tooth geometry, pitch, material, and "wrap" on the pulley is important. Tensile members vary among fiberglass, Kevlar/aramid, and steel cords. I kind of lean to Gates with their shock absorbing aramid fibers.

Barry, you've gotten the same "I'm not supposed to give this out, but..." info from that bunch at Gates. Now when it comes to pinning them down, they don't return my emails or phone calls.

CompositePro + btrueblood, good point, hadn't thought of that. Even though they all say "minimal" or "NO" stretch, this IS a long belt with a big load with start/stops...a natural spring. But wouldn't using a big enough Service Factor in the calculations (I'm at 1.8 so far, will make the belt BIG) and a proper tension (and preventative maintenance to keep it tight) minimize this? Hmmmm...will have to ponder.

Uncle, the carriage is tentatively designed at ~20" long, must travel the full 57' run end-to-end.

TygerDawg

 

[MikeHalloran]

Small timing belts, at least, are made by wrapping facing, matrix, cord or cable, more matrix, and cover, on a toothed drum, curing the assembly, and then slicing the product to width as the drum is removed from the cured, internally toothed tube of belting. It hadn't occurred to me that you could get longer belts by slicing the tube helically ... but chances are you're going to end up with at least one splice, probably more, and I'd speculate that's why the application engineers are reticent about performance claims.

Let's say you can make a full- strength ~114' belt of whatever width you need. Now apply a load, and you've got a pair of 57' bowstrings alternately humming, maybe enough to snag each other's teeth if the pulleys are small enough or the tension is low enough. Probably not a good thing to happen.

Or, suppose you get it tight enough to not sag appreciably. Now, if someone sits on the middle of the span while the machine is idle, they may be able to develop enough extra tension to break the belt.

At this point, I'd take another look at roller chain, and especially at bolting conveyor chain to a long bar to make a rack, and putting the gearmotor on the carriage.

I might even glue and screw 57' of timing belt to a bar, put the gearmotor on the carriage, and engage the rack with a short belt turned teeth out or a double belt as a 'caterpillar pinion'.





Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[tygerdawg]

Thanks for the feedback Mike.
The belts are made as you say, all vendors confirm that fabricating my length requirement is not an issue. Goodyear said "we ship'em out in spools of up to 300 feet". And certainly the least-torque requirement is mounting the gearmotor directly on the carriage and using some version of rack & pinion. That would require designing the system to meet explosive environment specs, and I don't want to do it unless I have to. But I might have to.

TygerDawg

 

[MikeHalloran]

It doesn't have to be an _electric_ gearmotor; it could be a compressed air gearmotor, or a hydraulic lsht motor, but now I understand why you'd prefer to mount the motor remotely from the carriage.

Have you considered a cable drive? Comprising a gearmotor with two spools of cable at one end of the carriage travel, and an idler pulley (and tensioner) at the other end of the carriage travel. One spool unwinds as the other winds, and a rotary encoder on their common drive shaft tells you the carriage position.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[tygerdawg]

Mike
Yup, one of the first ideas: put a winch on either end of the spray booth. Nice cheap solution, but discarded the concept because of positioning requirements and safety of popping a cable. Must index the buggy through the booth a little bit at a time, maybe even go back & forth for paint coverage. Looked at trying to apply "standard" industrial winches to this problem. COULD certainly be done, would require a lot of fancy programming of inverter drives to get the positions/speeds right because of changing spool diameters as I wrap & unwrap cable. I decided it would be beyond the capability of the client to support the system and they'd be calling me saying "yore durn machine don't work!"

TygerDawg

 

[GregLocock]

Why not separate the position reading function from the moving function? It should be trivial to set up a position sensing system on the trolley, which then commands the drive function remotely.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Warpspeed]

Nothing wrong with cable drive, wrap the cable several times around a large drum mounted on the buggy, and there should be no slip. Maybe even a commercial lift motor ?

I also agree that a totally independent and unloaded position sensing system will be required to give the long term accuracy you require. Perhaps a light duty toothed belt, roller chain, or a rack and pinion.

 

[unclesyd]

Does the centerline of the carriage have to travel the full 57' or does it only travel 37 ft total, 10' from each end?

 

[macmil]

How about a long lead or jack screw with say acme threads. with the carriage riding on suitable "nuts" or carriers. No stretch and I would imagine the control would be ok, either with limit switches or optical sensors.
mac

 

[unclesyd]

macmil,
That's what my question was leading up to. I was looking for the actual distance the carriage had to travel to see if an Acme screw was viable. Was also thinking optical or sound distance /location measurement.

 

[tygerdawg]

Thanks for all the comments, it's good to have many inputs. I think I need to move away from belts based on comments and the **deafening silence** from belt vendors.

I've considered a capstan & winch drive, but can't find enough info on it to guarantee the positioning. I've excluded ballscrew or similar due to the length (required stroke is ~55 feet). Over the weekend I have been looking into rack & pinion, but again it creates other challenges I'd rather not address unless I have to. Will look into chains and maybe, with suitable tensioning, these can be applied to minimize slack and provide suitable positioning tolerance. I'll start talking with vendors.

A sticky issue is the need for reciprocating motion: index a 24'-long buggy to a stop every 36" or so, process it, then retract and repeat the 36" segment until done. Hence my need for some positioning tolerance. Sensing position is not going to be an issue, I'll use encoders coupled with static prox switches to give double verification of position.

The ultimate issue is that this is a "retrofit" kind of engineering project with all the floor space constraints, low existing technical level of the personnel, and other issues that constrain the design. The client has very low expectations of what things should cost. I'm trying to develop a concept with which I am confident. From that I can get a reasonable BOM, then project costs and submit for approval to proceed. It might get cancelled immediately based on the projected cost.

TygerDawg

 

[unclesyd]

You can get Acme threaded rod up to 50'. That is the reason for my previous post. I have to admit that a screw that long presents some problems. The longest I’ve seen was about 25' and I can’t conjure the bearing and support arrangement.

http://www.keystonethreaded.com/acme.htm

If the travel increment is the same each time could some type stop arrangement be accomplished without the resort to a precision drive and positioning arrangement.