Timing Belt: Update #6 (Finale)

As mentioned in Update #5 last week, after using the sync belt and 33t sprockets for about 500 miles we switched back to the sync chain and timing rings for about a 35 mile ride with some friends on Saturday just to get re-calibrated.  After doing so, both Debbie and I came to the conclusion the chain had a noticeable reduction in drag compared to the belt.  It seemed as though our cranks were spinning with almost no resistance, a noticeable difference from previous rides with the belt. The latter was somewhat disconcerting as when we started the evaluation and switched over from the chain to the belt, neither Debbie nor I detected any difference in pedaling effort / drag from the belt.

Bear in mind, our daVinci cranks use 34t timing rings vs. the more common 42t rings found on most tandems.


As for why we didn’t detect any difference in drag at the beginning of the evaluation, I’m left to conclude either the belt’s very quiet and vibration-free performance created a placebo effect that caused us to not recognize the change in drag, OR the increase in drag developed during the testing.  In regard to the latter, as I swapped-out the bottom brackets while converting back to chain I noted the Phil Wood stainless steel rear bottom bracket’s timing side bearing had a gritty feel that it didn’t have when I first installed it to support the belt .   The front BB — a Phil Wood Ti-Mag model — also had a slightly gritty feel on the timing side that I didn’t detect when we started the testing as well.

Therefore, at present I suspect the 33t sprockets were just too small to be practical for use on a sync drive, which would also be true of a sync chain supported by 24t timing rings.  In other words, it wasn’t a belt thing, it was a small sprocket/chain ring thing remembering that, the smaller the sprocket or chain ring, the higher the force on the belt or chain… and when sprockets or chain rings get really small, the force they generate gets really big.

So, in summary, I think our experience with this particular experimental 33t sprocket w/12mm wide Gates 8MGT-1792-12 belt system was less successful than I’d hoped for a couple of reasons:

  1. Small sprockets and chain rings just aren’t a good choice; they simply generate a lot of added force on the sprockets and belt for no real benefit and often times to the detriment to other equipment and sync drive efficiency.
  2. Our Calfee — even with its couplings — may not be as robust as it needs to be to support a high-force generating, smaller timing ring/belt sprocket sync drive.

Even when the belt tension was set at the higher preload values that precluded ratcheting, e.g., 15lb or there abouts, if I locked the rear brake and applied a moderate load to my captain’s crank the bottom run of the belt would go slack… for whatever reason, there was just that much elasticity in our Calfee’s frame and bottom bracket bearings and axles.  As noted in early updates, the additional force generated by the small sprockets clearly caused the eccentric to rotate, which was more evidence that the small sprockets were stressing the sync drive and associated components under peak load.  As for the wear and tear on the bearings, I suspect a chain with a pair of 24t chain rings would have been equally hard on the Phil Wood BB’, etc.   Belt pre-load alone just didn’t strike me as a major contributing factor, even on experimental system with the small sprockets and higher pre-load.  After all, the drive-side bearing on the same bottom bracket has to deal with some truly massive loads and I’d guess that our Calfee’s Phil Wood Ti-Mag rear bottom bracket has somewhere around 6k miles of pretty stout wear and tear at this point.

As for trying a larger set of sprockets and longer belt, our friend Bob came up with an alternative configuration based on a 59t sprocket and longer belt that would have worked.  However, it may have required even wider bottom bracket axles vice the ones we were using with the 33t.  After mulling it over, and given that we were both at our limit for Q-factor on the Calfee with the 33t sprockets, we decided the 59t sprocket configuration was pretty much a non-starter for us. Bear in mind, I’m most comfortable using bottom brackets that are no wider than 108mm in combination with our daVinci cranks where most tandems come with 118mm sprockets and a proportionately wide tread or Q-Factor, if you prefer. Debbie is most at home with 108-113mm and was very happy when I switched the 119mm bottom bracket out for her 111mm one at the same time we put the chain back on.   I think I mentioned in the Paketa V2r thread that Calfee has apparently adopted a different stay design that’s deeply dimpled to work with the Gates 69t sprockets without using a super-wide set of bottom brackets, so they too recognized that sync belts do best with some special frame accommodations.

So, we are now at an end to our sync belt experiment and the belt and sprockets are on their way back to Bob.  While I am somewhat disappointed that our experience was truncated, I think it was worthwhile to explore the alternative system and smaller sprockets, if only to see if they might be viable. We could have taken a second run at it with more conventionally sized sprockets but, even though there were many positives for the belt, unless we found ourselves riding in adverse weather and doing a lot of sync chain maintenance, the tried-and-true sync chain drive simply works just fine for us.


About TG

I've been around a bit and done a few things, have a couple kids and a few grandkids. I tend to be curmudgeonly, matter-of-fact and not predisposed to self-serving chit-chat. Thankfully, my wife's as nice as can be otherwise we'd have no friends. My interests are somewhat eclectic, but whose aren't?
This entry was posted in Analysis, Technology & Equip.. Bookmark the permalink.

7 Responses to Timing Belt: Update #6 (Finale)

  1. wayne sulak says:

    i. find it interesting that you find the calfee too flexible for the small sprockets. i would estimate that a 34 tooth chainring would have 75% of the torque of a 24 tooth ring. have you considered going to a steel bottom bracket which should be twice as stiff as a high quality ti bottom bracket axle?

    • TG says:

      The point was, a 24t timing ring or 33t sprocket is too small for just about any frame and bottom brackets; again, there’s no juice for the squeeze. Moreover, as mentioned in my first belt report and the summary, I used a 119 +5R Phil Wood bottom bracket with a stainless steel axle for this project. The front BB used Phil Wood’s proprietary titanium axle with a magnesium/titanium/aluminum shell which, while not as stiff as the stainless steel axle, was more than adequate for the kind of loads that I can generate as a 160lb sport/recreational rider. I suspect an ISIS bottom bracket spindle or one of the MegaExo axles would have been more robust, but the bearings may or may not have been any more resistant to the increased torque from the very small diameter sprockets. Lord knows, folks with ISIS and MegaExo bottom brackets and chain drives & 42t timing rings have more than their fair share of reliability issues… more so than many of us who are still using square tapers, so it’s no guarantee that they would have improved the bottom end performance over the Phil Wood BBs.

    • Leland says:

      According to the engineer at Santana both the Calfee and Co-Motion lat tubeless bikes don’t pass our standards for lateral stiffness so I wasn’t surprised to hear this. I also immediately thought about how flexy Ti square taper BB axles are. Combine those two and this wasn’t the best test bed for such a setup.

      • TG says:

        You mean to tell me that a test developed by a company to provide evidence that its own designs did a better job of meeting its own specifications than their competitor’s designs actually supported it’s objective? Shocking!!! Next thing someone is going to tell me is that Gates has conducted studies where it could be shown that belts were more efficient than chains! Seriously, it would take a pretty objective series of tests by some unbiased tandem enthusiasts (and that in and of itself may be an oxymoron) who didn’t know whose tandems they were riding to figure out how stock tandems from the major builders stacked-up.

        The only other thing I’ll offer here about belts is that once the initial pricing for sprockets and belts was scrapped for something more reasonable, they became a nice option for anyone who was looking for an alternative to chain drive. They are definitely quiet, offer exceptional “connectivity” between the captain and stoker, are nearly maintenance free and lighter weight than the average OEM tandem cross-over / timing ring & sync chain systems.

        As for the selection of our Calfee as the test bed, I had to use what we owned at the time which, if you think about it, is a pretty good way to figure out how something is going to work on the tandem you own!

  2. Dave Walker says:

    Interesting observations; thanks for the in-depth analysis. I don’t mean to wax pedantic (although that’s often the bane of being an engineer), but just to clarify the technical explanation, as you change the diameter of the sprockets the TORQUE (=force*radius) stays the same but the FORCE (tension) on the belt or chain varies inversely with diameter. The end result is as you describe; smaller sprockets lead to much higher forces on the bottom bracket axles and bearings because of the higher timing belt/chain tension under load.

    I had to read back through the earlier posts regarding your custom belt-drive setup to understand the entire sequence of events. Most of what you’ve described makes sense to me, but there’s one key missing element. In Update #3 you say, “…Gates has designed the 8 mm x 12 mm carbon GT belt we are using to work on a two axis wrap with 30t pulleys and easily accommodate 1.5 hp… far more than I could generate, noting 1 hp = 600 watts and being mindful the sync drive only transmits the captains power to the rear crank axle.” First of all, 1 hp = 746 Watts, approximately, but that’s a minor point. The Gates spec’s are really written from the perspective of motor-driven assemblies such as motor vehicles and electrical drive systems, where the maximum power is usually developed at high rotational rates; often thousands of rpms, but almost always in the hundreds or higher. Bicycles are an entirely different application, where the rpms are low but the TORQUE can be very high. In my previous posts regarding the Paketa V2r belt drive, I used the example of the captain putting out 1000 W at 90 rpm, and showed that the force on a 69T timing belt or, equivalently, a 44T chain ring is 272 lb-force (ignoring any belt pre-tension issues). To drive the point home about torque vs. power, consider this: at zero rpm, with the captain pushing just as hard on the cranks, the tension on the belt or chain is the same as before yet the power delivered is now…zero. Another easy way to think of it is that Power=Force*Distance/time, in this case D is how much the timing chain or belt moves per second (or whatever period of time you choose) on the bike; i.e., NOT the bike’s velocity, but the timing chain or belt’s velocity relative to the bike.
    As you noted, with the small-diameter sprockets you used with your custom belt setup, the belt moves very slowly; barely 1/2 the linear rate of a standard tandem belt setup with 69 or 71T sprockets, meaning the tension is nearly twice as high for your setup as it’d be with a conventional tandem timing belt.
    Humans are closer to electric motors than internal-combustion engines in one respect: they produce maximum torque at zero rpm. In my youth, I once leg-pressed over 1000 lbs. in the gym—and I’ve never been a weight-lifter type. While no one can manage quite that on a bike (mainly because you’d need something to push against, and one leg is pushing down while the other is pulling up), my point is that a well-conditioned athlete can produce incredible forces on the pedals; perhaps 600 lbs. if you assume 450 lbs. pushing on one pedal while drawing up 150 lbs. on the opposite leg at zero rpm. I think I could still do that today, at age 52, in a static setting. That would produce 1200 lbs. of tension in the belt or chain—again at zero rpm, so the power output is still zero. You only transfer power when the cranks are turning and the belt or chain is moving.

    Turning to the Gates belt-drive spec wherein they say the belt can handle 1.5 hp, the obvious question is, “at what rpm?” Let’s assume something like 1500 rpm, a low cruising rpm for an automobile engine. You see what I’m getting at: the torque required is only 1/16.7 that required at 90 rpm as in our bicycle example to deliver the same power! A sprinter putting out 1000 W at 90 rpm places about 15X as much tension on the belt or chain as the 1.5 hp motor does at 1500 rpm. In other words, I’m not surprised you had belt slippage if you were using anything close to the tension spec for a motor-driven application using the same belt.

    The breaking strength of a Gates Carbon Drive (or Poly GT) belt is clearly pretty high; more than adequate for cycling as well as motor-driven applications. Avoiding belt slippage is a high priority, not only because your cranks end up being out-of-phase on a tandem, but also because it leads to premature belt wear.

    Final note: I’ve found, just like you, that when using a timing belt the eccentric needs to be much tighter than with a chain in order to ensure the eccentric doesn’t slip. We’ve had it slip, with disastrous consequences in one case very early on in our trials: the belt popped off, we were far from home, and I didn’t have the right tools with me to deal with it. We waited for quite a while for a friend to come pick us up with their vehicle. I’ve learned to tighten the eccentric bolts religiously!

    • TG says:

      You are, of course, right about torque vs force; old f**t brain f**t. I’ve corrected my entry.

      As for the eccentric, I’m pretty sure a chain running on 24t timing rings (similar in diameter to the 33t sprockets I was using) would have caused the eccentric to rotate under out-of-the-saddle climbing loads as readily as the belt did. I didn’t see it as a belt thing and bringing the Bushnell’s fixing bolt up to the proper torque spec. solved that problem. In fact, I was somewhat surprised to find I’d been under-torquing the eccentric to begin with: I’ve apparently lost my ‘internal torque wench feel’.

      I will say that, the more I’ve thought about same side 2×10 drive the more I’ve warmed up to the configuration for teams who don’t need the big climbing gears. It will be interesting to see how many teams show up at next year’s Co-Motion Classic with 2×10 and/or same-side drive (word gets around, don’t ya know).’

      We’ll need to play around with a 11x36t cassette before I can convince myself that the gear spacing wouldn’t be bothersome. Of course, to do that I’ll need to fit a Shimano XT or XTR derailleur with a Jtek Shiftmate #3 to our Calfee + the 11x36t cassette. Not sure if and when I might do that, so I may have to keep my eyes open for reports by other folks who adopt the 2×10 wide-range systems.

  3. Dave Walker says:

    Hi, Mark,

    Not to worry, as I suffer from CRS, too.

    As for your comment about 24T (chain) rings perhaps also causing similar issues with the eccentric coming loose, I can only say that, given we’ve only used the 69 or 71T versions of the Gates belt drive, I’m not prepared to say it’s just the smaller diameter that’s to blame: the eccentric definitely is more prone to movement with a belt drive than with a chain. Whereas we’ve not had issues with a timing chain, with the belt the eccentric has creeped on us more than once: it definitely needs to be tightened more than with a timing chain to stay put. I attribute this to the static pretension that’s constantly pulling on the eccentric.

    I’m still perplexed by Gates’ recommendation of 70-80 lbs. tension on the (tandem) belt. This is so far off of the 30-40 lbs. constituting the “green range” on the Gates “Krikit” tension tester that I’m left scratching my head what’s the optimal pretension?

    Considering the potential loads that the belt might experience, as I elucidated above (caution: serious brain damage may result), I’m of the opinion that the maximum load that a strong cyclist might impose is so much higher than the average load under normal circumstances that the only way to view it is that one should tension the belt so that it doesn’t slip under YOUR OWN maximum effort, at which point you should call it good. I respect Gates’ expertise, but at the same time I don’t think anyone within the company really knows the answer to this question, given the unusual mechanical strains that cyclists put on the components. It’s quite amazing, actually, that the Gates Carbon Drive belt is able to withstand the extremely high peak loads of a cycling application and still survive. In time, we’ll have a better handle on what’s the best compromise between the belt tension required to avoid slippage for most users and the added bearing wear caused by the static pretension. In general, it’s better to have a lower pretension for maximum bearing lifetime, balanced with the need to ensure that the belt doesn’t skip under high loads.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s