Tandem Cycling - Carbon Timing Belt Part 2 (oh snap!)

We now have the distinction of being the first team to break a Gates Carbon Timing belt. Last saturday's ride hammering out of the saddle on a short, but steep incline (around 15%) in the big ring, the belt lets go. Fortunately we were going slow enough at the time that we just came to a stop with no problem (except I banged the heck out of my knee on the handlebars, suddenly having zero resistence when your standing on the pedal with full force is not a good thing.)


Apparently, the belt is fairly sensitive to proper installation. The instructions make it clear that you should not "roll" the belt on the rings like you would normally install a chain. Rather you need to move the eccentric to where you can place it over the ring, and then tighten the eccentric. I did this as instructed on the first installation.

However, on the way to a group ride, in the process of getting the bike out of the car, the belt came off as I was getting the bike out of the car. (combination of the pedals going backward, and the belt being pushed by stuff in the car, as I pulled the bike out.

With limited time, and not wanting to loosen and retighten the eccentric, I carefully rolled the belt back on. It didn't take any force to do so. Thus I though the caution in the instruction had to be overstated.
I was wrong.

At the point where the belt broke ou can see very small line of diagonal scores across the teeth where the belt rolled on. Apparently, the twisting of the belt where it rolls on can lead to failure, and in our case did.

In spite of confessing my sin, Co-Motion is replacing the belt under warranty, which I take to be a standup gesture, given that they could have used my user error to deny any warranty defect.

The incident does give me a bit of pause about the whole belt idea. (It pains me to admit this knowing the rath of Mtbke I am sure to have heaped upon me.) Obviously I won't be rollong the new one on, and we'll see how the new one goes.

Back off on the steroids. Good thing it was only your knee that got banged!

How did your stoker fair? If you bonked your knee, did she come out unscathed?

^ not as big a problem for the stoker. When the timing belt (or chain) lets go, the captain's pedals are in free fall. The stoker's pedals are still connected to the drivetrain, so she still has resistence on the pedals.

^ not as big a problem for the stoker. When the timing belt (or chain) lets go, the captain's pedals are in free fall. The stoker's pedals are still connected to the drivetrain, so she still has resistence on the pedals.

Had I thought about it...even a little, I would have figured that out.

It seems as though the belt is non-field repairable. Did your stoker power you home?

merlinextraligh, Thanks for being the official / unofficial carbon belt tester for this forum. Hopefully the second belt will last longer. I hope the usual critic has refilled his medication. If not, get braced for a scolding!

You rolled it on one time and that caused it to fail? I had no idea these things were that fragile.

^ there appears to be a bit of a disconnect, given that the belts supposedly last 4-10 times longer than a chain.

Yet, rolling them on is a known cause of failure. I'm told we're the first to report breaking one on a tandem, but the rolling on issue has been a problem with single speed belt drives.

Belt drives are starting to become more common, with Trek specing them on a new commuter, among others. So I can't believe they are that prone to failure in normal use.

We've done our share of testing new products; some failed some made the grade . . .
Among some of the products we tested: tires (failures/blowoffs), a new-type fork (broke after 'only' 15,000 miles), saddles, rack (bad design), a carbon timing belt (prototype), a new sub-100gram eccentic (failed), etc.
Have broken a couple timing chains and we know the feeling: W-h-o-a!!! . . . Captain's feet do a super fast spin! In each instance captain steered/coasted while stoker kicked in the afterburner to get us home.
Thanx for posting the results and glad no-one was really hurt.
As usual, Co-Motion comes through with super-service
Pedal on TWOgether!
Rudy and Kay/zonatandem

Bummer about the belt, glad you'll survive. Some of the belts I used when prototyping belt drives in model cars back in the 80's were very efficient but fragile. The best belts, performance wise, had short lives. Not so much snapping but rather going from a cog belt to a smooth belt right now.

We popped the drive rings in a Hugi hub on the MTB tandem, big effort, slight climb, slow speed. It was not fun having both sets of cranks go into neutral. The noise and feel through the frame was bad too.

On the MTB tandem, each time we drag our timing chain or chainring on logs or rocks, I keep thinking how it would be nice to have a driveshaft and gear setup like a car rear axle. Use a splined BB spindle to couple the gears, and stuff the shaft inside the frames lower tube. Seal the tube and run the thing wet.

PK

However, on the way to a group ride, in the process of getting the bike out of the car, the belt came off as I was getting the bike out of the car. (combination of the pedals going backward, and the belt being pushed by stuff in the car, as I pulled the bike out.

With limited time, and not wanting to loosen and retighten the eccentric, I carefully rolled the belt back on. It didn't take any force to do so. Thus I though the caution in the instruction had to be overstated.
I was wrong.

At the point where the belt broke ou can see very small line of diagonal scores across the teeth where the belt rolled on. Apparently, the twisting of the belt where it rolls on can lead to failure, and in our case did.

From your description it seems that the belt came off fairly easily. Wouldn't you say that the belt "rolled-off"? If that is the case then the belt could have been damaged when it rolled off or when you rolled it on. Either way it does not seem like that alone should have damaged a belt... but it did.

I'm glad you are ok and that you posted this.
I was considering going with the belt drive on my next tandem.
Now I am thinking not. Safety is always the most important thing when cycling.
Oh and why were you in the big ring on a 15% hill?
I would not be doing that even on my single.

^ 1. I wouldn't necessarily rule the belt out. the failure apparently occurred as a result of me failing to follow directions. Also in hindsight I don't think it was completely without warning. I was hearing a noise, sort of a clunking sound, that I thought was coming from the captain's bb. In hindsight I think it was where the belt was already starting to break. (its not there now with the timing chain.)

2. We've broken a timing chain before, so it's not totally unique to the belt.

3. As for being in the big ring, this is like a 100 yard climb (not 100 yards vertical, 100 yards total) We bogged down more than I would have liked, and were muscling it out when it broke, but we've done that little section in the big ring a bunch of times before.

To me, the verdict is still out on the belt. If the next one fails or wears out following the instructions, then there's a problem. If it turns out it truely was my installation, then its just a cautionary tale that you do need to follow the specific instructions.

Other than the belt snapping, do you prefer it over a chain and was it worth the extra cost?
How much does a replacement belt cost?

Well, somebody had to be first (and I'd bet money that you're really not the first). If nothing else, thanks for being the first team whom I'm aware of that was willing or able to share your experience, as I'm sure it will go a long way towards driving home the need by others who own tandems with the Gates Carbon Drive to use care when handling the belts.

Just some miscellaneous thoughts that come to mind...

- A Gates carbon belt is probably best described as inelastic, far more-so that even a chain

- The carbon cords that are at the core of the Gates carbon belts are analogous to carbon seat posts and handlebars: if you handle them properly they will remain incredibly strong.

- However, if you bend, twist or stretch the fibers beyond their design limits they will crack -- in much the same way that a carbon seat post will crack when a clamp is over-torqued -- and once the carbon and resins that hold them together debond the cords lose most all of their strength.

- Like all carbon bits and uber-light or exotic materials, buyers really do need to read and heed all of the manufacturers handling instructions. Note that most of the high-end stem, seat post clamp, seat post and handlebar manufacturers have taken to putting maximum torque values on their components for this very reason.

- Gates seems to be well aware of what will damage the carbon cords in its belts:
http://www.carbondrivesystems.com/images/uploads/installation_1239728288.pdf

This makes me a bit worried, because I'm pretty sure that I've also not been following the instructions for our Gates carbon belt perfectly.

Even when the eccentric BB is in the most rear-ward position, there is still not enough slack to install the belt without turning the cranks a little to help it on, and so I have done this a few times. It requires very minimal force with the cranks, but it is certainly not possible to install it without this assistance. The BB spacing should be perfect for the belt because we have a stock-sized 2008 Co-Motion Speedster.

Can anyone who has this kind of set-up confirm that it is possible to get the belt on by following the method described in the instructions, or does anyone have any idea how I could do something different to make it possible to do so? Thanks.

Sorry to hear about your incident! My wife and I just upgraded from our Cannondale to a Comotion Machiatto w/the belt drive. The belt drive feels more responsive on accelerations than the timing chain, and certainly quieter. The bike is pretty amazing overall.

However, our new tandem is a whole lot more fancy than what we were riding, so it's probably hard to compare apples to apples. Good to know about the potential issues with the timing belt - thanks, we'll exercise caution.

The inquiring mind wants to know: how much are these belts? Is the Gates carbon fiber
the only option here? Timing belts are available in fractional to multi horsepower ratings
for constant service, not sure how that translates to a strong team standing on the cranks
on a steep hill. I suspect your team could go upwards of 2 kw for a very short time.

sch, the answer to all of your questions is probably somewhere in this lengthy thread:
Carbon Timing Belt System (http://www.bikeforums.net/showthread.php?t=532680&highlight=gates+carbon+drive)

This makes me a bit worried, because I'm pretty sure that I've also not been following the instructions for our Gates carbon belt perfectly.

Even when the eccentric BB is in the most rear-ward position, there is still not enough slack to install the belt without turning the cranks a little to help it on, and so I have done this a few times. It requires very minimal force with the cranks, but it is certainly not possible to install it without this assistance. The BB spacing should be perfect for the belt because we have a stock-sized 2008 Co-Motion Speedster.

Can anyone who has this kind of set-up confirm that it is possible to get the belt on by following the method described in the instructions, or does anyone have any idea how I could do something different to make it possible to do so? Thanks.

Installed the new belt today. I think I have the answer to your problem. First, put the ecentric as far back as possible. Second install the rear ring, and put the belt on it. Third, slip the captain's ring over the spider of the crank, but don't bolt it on. Fourth, put the belt on the ring while its unbolted. The fact that you can move the ring around will give you the play to put the belt on without rolling it on. Then, push the ring in place to bolt it on,with the belt already on the ring.

I hope that makes sense.

Installed the new belt today. I think I have the answer to your problem. First, put the ecentric as far back as possible. Second install the rear ring, and put the belt on it. Third, slip the captain's ring over the spider of the crank, but don't bolt it on. Fourth, put the belt on the ring while its unbolted. The fact that you can move the ring around will give you the play to put the belt on without rolling it on. Then, push the ring in place to bolt it on,with the belt already on the ring.

I hope that makes sense.

That makes sense, thanks. I have previously tried removing the whole crank, spider, and ring to install the belt onto that before re-mounting the crank. At least this only required messing with one bolt instead of the five that are required to take the ring off and replace it. However, I was unable to re-mount the crank when it had the tension of the belt pulling it out of position (and I did try doing so for several minutes). Hopefully removing just the ring instead of the whole crank will make this work better, but it is certainly not an easy solution.

A far better solution would be for them to make the belt 5-10 mm longer - I have LOTS more room to take up the extra length using the eccentric BB that I don't need with the current belt length. Maybe the belt length is designed to initially be this way because the belt will stretch a little over time, and then I will need to use this extra tensioning and the belt will go on easily without removing the ring. However, in our first 2,500 km of use, I haven't seen any noticeable stretching. Plus, I can't imagine that it would match the teeth on the rings that well if it was to stretch by that much, so I'm really not anticipating much stretching to happen.

Anyway, enough moaning about bad design. I guess the only way for me to avoid a repeat event like yours is to remove and replace the timing ring whenever putting the belt on. Thanks a lot for this tip, I'll let you know how I get on when we pack the bike up again to fly to Canada in two weeks.

A far better solution would be for them to make the belt 5-10 mm longer

Given the tooling costs to change the drum mold they probably use to lay these things up, I'm guessing that ain't gonna happen any time too soon unless they're selling belt drives at a much faster clip than I could imagine. It would be easier for Co-Motion to simply take 2mm out of their boom tubes on newer bikes.

Given the tooling costs to change the drum mold they probably use to lay these things up, I'm guessing that ain't gonna happen any time too soon unless they're selling belt drives at a much faster clip than I could imagine. It would be easier for Co-Motion to simply take 2mm out of their boom tubes on newer bikes.
Seems it would be simpler to make different diameter rings?

Seems it would be simpler to make different diameter rings?

Perhaps... But, to be candid, of all the 'innovations' to come along for tandems the belt drive is right up there with the superlight Topolino wheels: damn expensive, 100% proprietary, of questionable value but really cool nonetheless.

I 'took one for the team' wit the Topolino wheels and will continue to use them when conditions warrant but don't feel I've been left out by not being able to fit the belt drive to our tandems.

I assume the timing pulleys are molded. A new belt mold or new pulley molds would be very expensive. They would have to sell a bunch of these things to justify either mold.

We now have a number of rides on the new belt. I tried to follow the instructions as closely as possible in the installation.

After a couple of rides, the belt started slipping under heavy pressure. It first did it on the same hill where the first belt broke. At first I wasn't sure that it was in fact the belt skipping, but when it became more frequent, you could see that the captain's crankset was advancing, and at the end of a ride where it skipped multiple times,the captains pedals were about 60 degrees advanced.

The belt tension was reasonably snug, certainly tighter than you would typically run a timing chain. In talkin with Web Cyclerly they suggested upping the tension, so I tightend the belt pretty much as tight as I could.

The tightening stopped the skipping, but lead to a substantial amount of noise. So I backed off the tension, noise disappeared, but it skipped again. Right now we're back to very tight, noisy, and no skipping.

I'm going to try reversing the belt direction as suggested in another thread.

I'm concerned at this point that the violent skipping under pressure may have damaged the belt (given that they can apparently break from just being rolled on the ring) And the noise level is really not acceptable.

So for us the belt is an ongoing saga. And I'm beginning to see why Co-Motion does not reccomend retrofitting them.

Ever since I tried to put the power down on a belt-drive equipped Strida folding bike, and the thing just skipped like a slipping clutch in a car, I have been skeptical of the belt drive.

I know it can work, AND I think it's a good idea... but man, everything has to be spot on.

The skipping is alarming. I've read that these belts are being used by racing teams, so there must be some trick to it. The Gates website shows a special tensiometer in use, but I don't see it for sale or its cost. It looks very simple and inexpensive, though. By the look of the tensiometer, the distance between centers would effect tension, the greater the distance, the greater the tension. Which wouldn't make any sense if the belts stretched at all. But, since you are skipping teeth, they obviously do stretch! The Gates site instructs to have 1/2" deflection at 10 lbs., but all photos show a belt on a single bike, which is much shorter than a tandem belt. So this is all quite unclear to me.

I'm friends with a very fast local tandem team, I would call the captain extremely strong, riding passes on a 90" fixie. They just got their newest racing tandem, with carbon belt. I'll talk to them about these issues.

I was intrigued by the belt and heard many liked it.
I think this nails it for me. I'm sticking with chains.
What happens if you are a long from home and have a problem?
Are you suppposed to carry an extra belt at all times?
Just doesn't seem very practical.

^^ It's not *supposed* to be practical. Just profitable.

I was intrigued by the belt and heard many liked it.
I think this nails it for me. I'm sticking with chains.
What happens if you are a long from home and have a problem?
Are you suppposed to carry an extra belt at all times?
Just doesn't seem very practical.

I still want the thing to work. Also I don't want to be overly negative and discourage others, particularly if my experience is unique, or the result of my installation, or maintenence.

But personally, I would think twice about retrofitting one again, until there is some more experience out there about their long term results. And also, hopefully allowing the price to drop in the interim.

All hail the early adopters (I'm not one of them). Without them, technology wouldn't advance.

Gates does market a tensionometer. It treats the belt span like a guitar string. The inputs to the device are the unit mass of the belt (like gm/cm) and the belt span. You then pluck the belt like a guitar string and a microphone on the device listens for the fundemental frequency. This translates into tension. The higher the tension, the higher the frequency.

These devices aren't cheap (many hundreds of dollars).

Where I work we use timing belts all the time but mostly for precision motion transfer - not for power transmission. However, Harley Davidson uses timing belts for power transmission on their motorcycles so it is obviously possible. I wonder if the belt was sized correctly for this application.

I would think that if the belt were slipping, the captain would feel a distinct clunk as the belt slipped one tooth & caught on the next - not a clutch like slip. But for the belt to slip, all the teeth on 180 degrees of wrap around the pulley would have to disengage all at once until the belt teeth advanced one pitch. I think someone would have to watch carefully as the captain pushed hard on the pedals with the bike in a stationary mode.

I would think that if the belt were slipping, the captain would feel a distinct clunk as the belt slipped one tooth & caught on the next - not a clutch like slip. But for the belt to slip, all the teeth on 180 degrees of wrap around the pulley would have to disengage all at once until the belt teeth advanced one pitch. I think someone would have to watch carefully as the captain pushed hard on the pedals with the bike in a stationary mode.

It's a hell of a clunk, not a smooth slipping. You can hear it and feel it. Loud enough that people around us go WTF?

At first I questioned whether something else in the drivetrain was making the noise. However, it doesn't do it when we put the timing chain back on, and you can see where the captain's cranks have advanced after it does it, so I'm pretty certain it was in fact skipping.

To put this in context, it only does it with a lot of torque being put on the pedals. It happened doing stomp drills (accelerating from 10mph in a 53/12 as hard as you can.) and climbing out of the saddle in the big ring.

If it were just under the load of doing stomps, I wouldn't be concerned, but racing and competitive group rides, there are going to be times where we're going to be doing short power climbs in the big ring, and skipping in that situation isn't acceptable.

All hail the early adopters (I'm not one of them). Without them, technology wouldn't advance.

Gates does market a tensionometer. It treats the belt span like a guitar string. The inputs to the device are the unit mass of the belt (like gm/cm) and the belt span. You then pluck the belt like a guitar string and a microphone on the device listens for the fundemental frequency. This translates into tension. The higher the tension, the higher the frequency.

These devices aren't cheap (many hundreds of dollars).
<snip>Then what is this?
http://www.carbondrivesystems.com/images/uploads/installation_1231907551.pdf

I don't see any other tensiometers on their website. This one looks very simple and cheap, and as you say of the other device, it also takes span into account.

And I used to run OS/2 Warp on all my machines.

This is the one we use at work:

http://www.gates.com/brochure.cfm?brochure=2787&location_id=3539

Again - it treats the belt span as a guitar string that will oscillate at different frequencies based on the belt mass, span & tension. May not be practical for bike owners.

Increasing the belt tension to prevent it from skipping is probably not a good solution. This will likely increase the friction in the system, cause premature belt wear and/or failure, accelerate BB bearing wear and put a lot of bending force on the bike frame.

Some belt tooth profiles (curved -vs- trapazoidal) are better for transmitting power. I would assume that Gates and the bike designers would have optimized this already. Since you already have ~20 teeth of engagement on each pulley, it is hard to imagine how one could reduce the skipping problem some other way.

My tandem racer friend got back to me with the following:

"We have about 1100 miles of trouble free use so far. Much of it has been under extreme racing conditions. Our experience so far: quiet, clean, maintainance free, and light weight. It is essentially a smaller version of the kevlar belts they use on Harleys and Buells successfully on their motorcycles. Generally on the motorcycles the belt can last the life of the motorcycle.

If you are considering the upgrade I would certainly go for it . . ."

I was intrigued by the belt and heard many liked it.
I think this nails it for me. I'm sticking with chains.
What happens if you are a long from home and have a problem?
Are you suppposed to carry an extra belt at all times?
Just doesn't seem very practical.



My tandem racer friend got back to me with the following:

"We have about 1100 miles of trouble free use so far. Much of it has been under extreme racing conditions. Our experience so far: quiet, clean, maintainance free, and light weight. It is essentially a smaller version of the kevlar belts they use on Harleys and Buells successfully on their motorcycles. Generally on the motorcycles the belt can last the life of the motorcycle.

If you are considering the upgrade I would certainly go for it . . ."

Maybe they are not designed to break unless damage and we are too used to having chains that break every so often. I don't think people on motorcycles carry an extra belt.

However, if the belts last the life of a motorcycle, why does Gates say they "only" last twice the life of a bike chain?

You may be surprised by the amount of torque a heavy bike rider can exert on a drive train while standing and putting all the force they can into the pedals. It's more than a Harley, and may rival a V8 car. Don't ask me to prove it, but this was argued out on rec bikes tech years ago and the concensus was stunning. Small chain rings increase the burden.

I don't see how even the strongest rider could develop as much torque as a V8 car engine.

Conservatively, a V8 can develop 300 lb*ft of torque. A 175mm crank arm on a bike is .57 ft long.

300 lb*ft/.57 ft = 526 lbs. I don't see how any rider could apply that much force to the pedals and if they could, many things on the bike would fail.

If a bike were designed to withstand that much torque, you wouldn't want to ride it - it would be far too heavy.

I'm at 195lbs. and if you drop me from .57 ft. and I land on one foot how much weight is the impact point?

FWIW,

I went back and looked at a couple of old Powertap files from races, on my single bike. The highest tourque I saw was 1633 lb/in. Converting that to foot pounds yields 136 foot pounds of torque.

I'm betting that when we broke the belt we were putting out more torque than that because we were in a very big gear, on a steep grade, and pretty much giving it 100% to keep the pedals turning over.

Nonetheless, I don't think it would have broken without some preexisting damge in the installation.

dvs - torque = force x distance from the axis of rotation - not weight x the height of the drop.

The only way you could increase the force put on the pedals above your weight would be to pull down on the bars. Then maybe you could briefly apply 150% or so of your weight.

The number I remember was about 200 ft lbs on a mtb crank with a 22 tooth ring. I suspect a 200 pound rider could exert over 400 pounds of force on a pedal by collapsing wieght, extending the driving leg, pulling up on the other side, and pulling up on the bars.

In any event, its more than a Harley can develop. Too bad we are limited to such low RPM's.

EDIT I just googled and I saw the torque rating on a Harley is 81 ft/lb's.

Another edit- I had a 5.0 high performance Mustang, which was a very fast car. It was rated at 300 ft/lbs of torque.

This whole discussion on rec bikes tech came about over a discussion of chains breaking more often on mtb's compared to road bikes. The small chainrings allow cyclists to test the limits of chains

The Harley belt is considerably wider than the Gates tandem timing belt and I would think the alignment of the motorcycle belt to be essentially perfect so a comparison of the torque doesn't necessarily go to the heart of the problem, but does emphasize the load on these belts.

Does anyone have information as to Gates intended market for this product? Someone had earlier mentioned that a version of belt drive was appearing on lower end, cruiser bicycles. Tandem applications will range from cruiser types to racers and it's possible the stress on the belt over this range of applications would vary by a significant factor. Has Gates really designed this product (belt plus hardware) for the severe application represented by really strong go-fast tandem teams?

^ They were being raced in the Co-Motion Classic Tandem Race last year, apparently without any problems.

You armchair physicists are confused if you are comparing crankshaft torque on a motorcycle with crank torque on a bike.

IF you wanted to make a comparison (engineering by analogy is an awfully imprecise science!), you should compare belt tension on the bicycle vs. belt tension on the moto. Ignoring preload, belt tension on the motorcycle will come from crankshaft torque multiplied by transmission gear reduction and divided by output pulley radius. But it seems like a totally pointless comparison.

Just to update how the belt is working, after playing around with the tension, we've gotten to the point where it is basically quiet, with occassionally a little noise, and we can't make it skip.

Now, my question is whether when it was skipping it damaged the belt. Gate's literature says that skipping can break the carbon fibers, so I'm pretty sure I know what they're going to say when I talk to them.

I'm pretty sure I know what they're going to say when I talk to them.

No Belt For You!!! Come back one year...

http://jenkinsfamilyblog.files.wordpress.com/2008/12/soup-nazi2.jpg

The smaller the pulley (or chainring), the higher the tensile load on the belt (or chain) for a given force input by the rider.

If a rider can exert 200# (at 3 o'clock) on a 175mm crank arm, the belt (or chain) will see about 577 pounds on a 30 tooth (~61mm radius) chainring.

If the noise were caused by the belt momentarily coming out of engagement with the pulley teeth - either by overload or by friction with the flanges or something else, the belt path would also momentarily (and periodically) be increased - stretching the belt. This could very easily break the tensile cords - espeacially if the stretch were not uniform along the length of the belt.

It would make me very wary of standing up. Sounds like you were lucky that you weren't badly hurt the 1st time.