TandemGeek

For those of you who are using a Gates Carbon Drive belt & sprocket system on your tandem, or even on a single bike, I have a couple questions about your belt.

1. Have had to install or re-install your belt since taking ownership?

2. If, so, did you check the tension of your belt after the installation to ensure it was per Gates' specification?
a. If yes, do you remember what that specification was?
b. What method or tool did you use to measure the tension? Finger pressure, a Gates Tension Tester / Krikit Gauge, or...???
c. If you used your finger, how did you "calibrate' your finger such that you'd know what the specified tension would feel like?

3. What tool or tools did you use to tighten your eccentric to achieve the proper belt tension and how much effort did it take to get the specified tension?

4. If you have a belt, but have never had to remove or re-install it; have you ever checked the tension to see if it's per Gates' specification?

5. If the answer to all 4 of the previous questions is No, did you even know that there was a specific tension requirement and that belt tension needed to be checked?

Brad Bedell

1. yes. I installed it. Did not need to re-install it.
2. Yes, I checked it.
a. by feel.
b. finger, deflected about 3/4" (without pressure on any of the pedals)
c. years of calibration checking belts on cars. *IF* the Gates belt was adjusted too tight, it did not spin smoothly, seems the cranks or pulleys were out of round slightly. One can feel the variation in pressure when spinning the cranks. (The drive chain was removed when doing this setup)

3. Allen key, see supra.

4. I check the bike over for anything at least once a week. Belt never seemed to change tension.

5. Yes, I knew the tension requirement is there. After dealing with belts, band saw blades and what not long enough, one learns to feel when something isn't quite right. I do own an automotive belt tension checking tool, I suspect it would work on this.

Same thing really for a torque wrench. On carbon fiber, I always check torque. On metal to metal bolts, just by feel. I'm usually pretty close when I go back through to verify things.

Bent In El Paso

I am glad you posted this. I have been wondering the same thing. Here are my answers.....

1) Yes

2) The spec on the CoMo web sit is when pressing down on top of the belt half way between the pulleys with 10 lbs of force, the belt should flex no more the 1/2 inch. The spec worries me because it has a maximum amount of deflection allowed but no minimum. How tight can the belt tensioned without causing damage to the belt and/or bottom bracket bearings? I press down with approx. 10lbs of force and watch the belt deflection. No measuring tools used (yet).

3) I use an open-faced bottom bracket tool to tighten the belt. It is very easy to adjust the tension with this tool.



4) N/A

5) N/A

Carbonfiberboy 

1) Yes. It was a retrofit installed by my LBS, but see below.
2) Yes.
a) Yes.
b) Finger pressure
c) Used a kitchen scale to calibrate finger
3) Allen wrench. The CoMo eccentric tightens by rotating a crankarm against the Allen wrench, as you know. So really very little pressure on the crank.
4) Yes.

So - the nitty gritty. Neither Stoker nor I are big wattage stompers. The LBS adjusted it much too loose. I tightened it to the CoMo spec: 10 lbs. = 1/2" deflection. Part of the install was new BBs and cranks. Riding the bike, it seemed to me that there were greater frictional losses in the system than there were before. I removed the belt and drive chain and spun the cranks. No drag there. I added the chain to the stoker's crank. Some drag there, but little or no difference from my single. I tested the drag on stoker's crank with a postal scale. I reinstalled the belt to the CoMo spec and tested drag again. I was disconcerted to find that the drag had more than doubled. I did notice a variation in belt tension with crank rotation, but it was small. I backed off the belt tension to more like the 10 lbs. = 3/4" that Brad is saying. I found the drag much reduced, so that's what we've been using. I have checked belt tension, but have not found any slackening. Also have not seen any wear on the belt. Also have never slipped a cog. About 1500 miles on the belt.

Source of the drag is a bit of a mystery to me. I can't imagine that small belt tension causing measurable drag on the BB bearings. So it much be internal drag in the belt. But if so, why does lower tension decrease the drag?

Brad Bedell

CFboy: this is a bit of thread drift, but it sounds like you're experiencing something similar to what I experienced. I think I'll set up the dial indicator when I do the install for the new tandem. I *suspect* the pulleys and or cranks are out of round. Cranks are more likely to be the issue as the tolerances would be more difficult with crank machining than machining out the pulleys. Plus, it seems familiar to me that I've seen chain rings bounce a bit on cranksets. It might take finding the high/low spots and rotating them around a bit to help things out.

FSA carbon cranks front and rear. (tandem cranks rear, SLK-Light fronts, but the problem was there with the bontrager Octalink cranks also)

Other notes, having the chain too tight netted me a similar drag as the belt.

Chris_W

1. Yes. Installed and re-installed belt multiple times (to change drive-side chainrings, to service BB, to disassemble the coupled tandem for flying, to switch between IP and OOP, etc.). Installation is complicated because our rings are the wrong size. We got an early set of the Gates system (ordered November 2008), which came with 71 tooth rings. It is JUST possible to get the belt on with the eccentric BB in the slackest position, but to do so still requires using the ill-advised method of rolling the belt onto the rings. This is on a stock-size 2008 Co-motion Speedster, so distance between BBs should be what the system was designed for. I contacted Gates about this and they said they're aware of the problem and so later versions were shipped with 69 tooth rings, which they said I could buy for $175 EACH! I see it as a design flaw, and so should be fixed for free, but they seem to think it's not a real problem. So far using the ill-advised roll-on method has not caused any notable problems with the belt for us, but I'm as careful as I can be when doing so.

2. I checked initial tension as per guidelines from Co-Motion / Gates using an object of measured weight and assessing belt deflection. Initial tension seemed to be slightly above the recommended amount, but again there is no way to remove any tension. Since then I've occasionally checked with finger pressure to see if there has been any change, and I've noted how difficult it is to put the belt on each time. I was hoping that their will be a small amount of stretch over time to make the tension more spot on and make installation easier. After about 5000 km (3000 miles) there has not been a notable change (and the belt is still a b1tch to get on).

3. See above - no tension adjustment possible because rings are too large / belt too short.

4 & 5. N/A.

Although I'm not happy with the sizing and installation problems, I've been happy with the other aspects of the belt drive (minimal maintenance once installed, clean non-drive side of the bike, minimal wear, huge bling factor).

One thing that I've learned is that the belt is directional, the writing on ours needs to be readable when standing on the drive side of the bike. When the belt is installed in the reverse direction then it makes a lot of unusual noises when power is put into the captain's pedals. When installed in the other direction, there is no noise. I'm assuming that the noise-less direction is the correct direction. I didn't notice anything from Gates saying that direction makes a difference and I don't know if this is common or not.

If anyone wants to buy a set of 71 tooth rings for use on a tandem with a slightly shorter BB distance than a CoMotion then let me know!

tredlodz

1. It's been on and off 4 times (Coupled bike - travel)

2. Yes and no. I marked the eccentric position from my LBS' initial install and aligned to that position for subsequent installs. A finger test shows moderate deflection with moderate force, probably a little tighter than I've run a chain before. But no sophisticated meter was used.

3. Tension was adjusted with a pin spanner, but it wasn't really necessary. It would be possible to push it into position by hand.

4/5 NA

Additional comments:
I've see no sign of stretch in around 2000+ miles use, still as tight as ever.

I do notice variable tension as the cranks spin. Our 69T rings don't look that beautifully made, so it could be ring concentricity, or it could be concentricity issues in the crank attachment. There is a little drag if you just spin the cranks, but I'm not sure how this translates to the real world when the Captain tensions the belt - If anything tight timing chains always seemed more efficient than loose ones to me.

I've considered reducing tension a bit, to reduce drag and maybe help the bearings, but so far the belt hasn't slipped while standing and I'm not keen for this to happen.

BTW, our belt seems to run nicely in-line about 1mm from the flanges, but our friends have a belt that walks so far from both flanges that it overhangs the pulley on the inside (both pulleys, front and back so eccentric adjustment doen't help). Has anyone found a solution to this? (they have a Co-Mo Supremo and the 71T pulleys which are super tight as has been noted).

TandemGeek

Can you elaborate on this?

I'm still working to fully grasp the various realities of how these belts are actually 'working' on bicycles, but in general the Gate's GT belt tooth profile works most efficiently only when it can fully seat in the sprocket grooves, which is very different from how a chain works. Chain rollers pull against only the driving side of the chain ring teeth, which is why you see asymmetrical chain ring pitch elongation and can eventually flip around a timing ring to get a second life from it.

Quoting from Gate's White Paper on synchronous drive belts, "The belt’s modified curvilinear teeth enter and exit the sprocket grooves cleanly, eliminating speed variation and vibration as the pitch line of the belt moves around the sprocket relative to the pitch line of the sprocket." In plain speak, what I believe this means is that when the belt is properly tensioned -- not too tight nor too lose, but just right & per spec -- the shape of the belt's teeth and the sprockets grooves (aka, the patented Gates GT tooth profile) allow the belt teeth to engage the sprocket's grooves without making contact with the upper edges of the grooves. So, in theory, the belt only fully engages the sprocket when the rounded profile of the belt teeth seat in the matching profile of the sprocket grooves (See attached photo from Gates Design Manual). Therefore, what I've come to believe is that if the belt isn't tight enough or is too tight, you don't get that smooth engagement between the teeth and sprocket grooves, hence... you get increased friction and vibration as the belt engages the sprockets more like a chain instead of a precision gear mesh.

Where I'm beginning to struggle a bit with the belt concept as a tandem sync drive is the realities of how a tandem's bottom brackets and frame react under pedal loads. It was Dave Walker's V2r, right-side-drive design and the ensuing dialog that created a whole bunch of questions in my mind, as well as his. While the belt drive -- in theory -- should be just fine as a tandem sync drive and as a primary drive on a bike, I'm now wondering if current bicycle / tandem frames and their bottom brackets / axles really are stiff enough to translate the theoretical low-drag performance that allows the belt to match a chain drive's very high efficiency into reality. Let's face it, while the individual component and frame deflection values of boom tubes, bottom bracket bearings, axles, and cranks are all very low, there is a stack-up effect that causes the distance between the bottom brackets to shorten under loads, albeit a very small amount. So, if we've gone to all the trouble of making sure our belts are at the precise tension required for maximum efficiency, what happens when that tension drops once we begin to apply power to the pedals?

My simply "test" of this theory remains a locked brake application of pedal loads to the pilot's cranks to see how tension changes in chains and, now, belts. With chains, there's an awful lot of slack inherent in a chain that contributes to the loss of tension in the non-driving side under load, but with the Carbon Drive GT belts... there is nearly zero stretch under load, nada. However, When I apply my locked rear brake test and load up the front crank by putting downward pedal pressure on it -- pretty much what happens when you start a tandem from a dead stop, sprinting, or climbing a steep grade -- there's a dramatic shift in belt tension between the upper and lower runs. In theory, and again assuming that the belt is running on a pair of axles that have a fixed distance, the net change in tension between the higher and lower load bearing runs of the belt should net each other out, i.e., if you had a value of 100 in the top and 100 in the bottom for a total load value of 200, when the belt goes under a driving load on the top run that moves it to 150, the lower run should still have a value of 50... and the preload value should keep the non-driving run of the belt from falling below a minimum value of some sort, i.e., it shouldn't go slack.

This is what Dave and I have been going back and forth over for the past few days. While I can now see how a single-side-drive will mitigate the bending effect on the rear axle between the bottom brackets of the sync drive, you'll still have the pilot's front bottom bracket deflection as well as the net effect of any right side bowing that extends from the rear axle to the front bottom bracket as both the drive chain and sync belt go under load.

Yeah, it makes my head hurt too. But, the reason I'm poking extra hard on this is because my experimental system with it's very small 33t sprockets exacerbates the static preload impact on the bottom bracket bearings, as the Gates spec. is not 5-10 lbs of downward pressure for .5" deflection, it's 17-20 lbs of downward pressure. Remember, the smaller the sprocket diameter, the higher the belt tension goes under load. Frankly, I haven't been able to bring myself to bring my belt up to the full spec for tension and have been using the 15 lb value, presumably to no ill effect. But, then again, I don't think I'm 'calibrated' well enough to know if we're experiencing any loss in sync drive efficiency from the belt vs. the chain, or a potential decrease in belt life just yet. I just know that I'm now spending a lot more time thinking about a part of our tandem that I never gave a second thought.

So, this is the source of my unusual and sudden renewed interest in belt experiences and tension setting.

TandemGeek

If they haven't tried it yet, flip the belt around so the white printing is facing the other way. As noted above, these belts do seem to exhibit an inherent desire to run better in one direction vs. the other. Weird, but I've seen it first hand.

Bent In El Paso

I have also noticed the increase in drag if the belt is too tight. If the belt is too loose, it seems to make noise and wants to wander in and out on the pulleys. There does seem to be a "sweet" spot where everything works as advertised. I have not noticed the directional issue with the belt. Whenever I take the belt off, I always put it back on in the opposite orientation thinking this may "even" the wear on the belt and hopefully extend its life. I have not seen anything either on the belt itself or the Gate's specifications which mentions a specific direction.

I tend to be pretty accurate using judgement in making mechanical adjustments. Just to be sure, I think I may try using my fish scale to check the belt tension. I think by hooking it over the top of the belt I can place 10 lbs of force on the belt and measure deflection using a ruler. It's a crude method but should get me in the ball park.

The first couple of times I adjusted the tension, I used the crank against the allen tool method. With the location of the allen tool against the FSA spider, I was not able to make this method work too well. Maybe I was not doing it correctly. I also did not want to mar up carbon finish. That is why I went to the bottom bracket tool. In my experience, it is much easier to use this tool to get the correct tension on the belt.

Carbonfiberboy 

I put the platform of a small spring-type postal scale under a horizonatal pedal and pushed on the pedal until it moved in the backwards direction. IIRC, I got 3.5 oz. on the stoker's 170 crank with drive chain on and belt removed. I was not able to get a measurement on the captain's pedal, as the BB rotated too easily. With the belt added and at the CoMo spec tension, I got 7.5 oz. on the captain's pedal, 172.5 crank. With the belt detensioned to the 3/4" deflection, I got 5.5 oz. on the captain's pedal. I thought that was OK. I'd be interested in hearing from others who make these measurements.

I realize that the rear hub drag is not a factor and that I should have measured the belt drag with drive chain removed, but it was too much bother.

And thanks once again to TG for exploring these issues that some of us just wonder about. How the teeth engage the sprockets does seem a reasonable explanation of the varying of drag with tension.

TandemGeek

I guess the question that comes to mind is, is there a direct correlation between sync chain or belt drag under what is essentially a static load vs. the dynamic loads experienced while actually riding? I would intuitive think that the typical sync chain would have very little drag 'at rest' compared to what it would experience during use. Conversely, I would expect the belt's drag to be higher "at rest" because of the preload; however, the question really is... how much change in drag occurs during use?

It's somewhat similar to rolling resistance vs. wind resistance. Quoting from the same Roues Artisanales article I linked to in my recent blog entry on tires:

Of course, unless someone has the time and $$ to instrument a test rig and collect dynamic performance data... we may never know.

Carbonfiberboy 

I wasn't only measuring static friction but also kinetic friction during some movement. As you point out though, that may not be the same as or proportional to the drag under load. I did think I could feel it, though that's always subject to question.

Here's a good thread on chain and pulley friction:
https://biketechreview.com/forum/1-ge...pinning-pulley

Oosbahnd&Weefay

1) Yes, several times (4 minimum)
2) Yes
a)1/2 inch deflection under 10 lbs
b)Finely calibrated index finger
c)Had it set to spec in a shop and then got used to that particular tension
3) Used the S+S coupler wrench to tighten it on the trip since I had nothing else available to me (at my own error)
a) of note: this method actually works quite well

edit:
Biggest takeaway from this puppy on our Pacific trip was that as I railed on the bike up the hills day in and day out, I managed to tighten the belt over time. This resulted in (very slightly) pulling the eccentric out of plane, which in turn created two problems. First would be more drag, but I only know that due to the physics; there wasn't an increase to the point where I knew it for certain, or where there was noticeably less coast in the pedals if rolled backwards. Secondly, there was an AWFUL squeak that came out of the BB. I'm sure it wasn't even 10% as loud as I recall, but hearing that squeak 4 hours a day for near a week had me near murderous at times.

tredlodz

Got my friends to try flipping the belt and it now runs next to the flanges. Good fix for anyone who has a problem with the belt walking away from the flanges at both ends.

akexpress 

We just returned from the Santana Barcelona to Lisbon tour with our coupled Calfee with belt drive. It is only the third time I have put the belt on after traveling and as others as stated when tight to specs its seems to have a lot of friction so I left it a little looser and sometime on the second day it skipped a tooth that neither of us noticed when it happened and in fact probably rode the next day with it out of phase. We think we know when it happened as we came around a blind corner and had a very steep on ramp and didn't have time to shift and just both stood up and made the hill but in a very high gear. We have no issues with it walking on the sprockets but have had the eccentric move if not tight enough. We love the clean no chain ring tattoo effect of the belt. The other tandems on the trip with belts had tensions all over the place and no one seemed to have any problems with them. There was no consensus among the teams as to what they should be and in fact most didn't even know gates had recommended tensions.

ianbal

Sorry for hijack but with my current number of posts I can't send a pm. I was wondering how that tour was. We were on the Paris-Amsterdam trip earlier this year.

Cheers
Ian

akexpress 

Ian
We had a great time but it was a very active trip with lots of time constraints due to the boat schedule. I think Bill and Jan were somewhat maxed out at times dealing with a new trip but they pulled it off with a good result. I am sure they will modify some of the rides next year to help with the timing. The use of GPS for routes worked well but had many trials and frustrations for some teams and Bill, again that will be easier for next year as much was learned. We were one of only about 12 teams that did all the rides and the longest ride each day, as there was much more climbing then most people expected. the boat and crew were awesome. I don't know why I can't send pms as I used to be able to now one needs 50 posts, seems rather arbitrary to me. you can send me an email at akexpress@mac.com for more info
Mark