Tech Tuesday: Belt Drive Full Suspension Singlespeed MTB

I've been single-speeding for nearly 30 years (not including BMX bikes as a kid), with a few years of derailleurs thrown in for good measure. I don't dislike multi-speed bikes on a philosophical level, nor has a derailleur ever truly stranded me on a ride, insulted me or killed a family member. I just don't like the noise(s) they make. And so, the few periods where derailleured bikes were in the stable, I quickly converted those bikes to single-speed (or sold them altogether) simply because of the extra chatter that a derailleur and its pulleys make.

However, my aural fixations have not been without irritation in my current era of bikepacking and ultra-distance bicycle racing. The long distances, inclement weather and water crossings that frequent many of the routes in the southeast has made it a challenge to keep a chain clean, lubed and quiet for the duration of a ride. I often packed a small bottle of Squirt to re-lube the chain every hundred miles or so, but occasionally fell victim to my own poor planning (laziness) if I forgot to refill or even bring the lube, leaving the remainder of the ride an unbearable cocophany of gnashing metal.

A few years ago, I acknowledged the appeal of the Gates CarbonDrive belt-drive system and its supposed quiet operation and reduced need for maintenance. I first chose to convert my (t)rusty Ritchey p29er. If old MacDonald had a bike, it might be this (guinea) pig of a workhorse with a braze braze here, and a weld weld there, cut and weld and braze and weld, and everywhere a rattle can spray bomb...you get the picture. 

Oink oink.

A 50x20 for the hills of Florida.

For all of the warnings about frame alignment and lateral rigidity, perhaps that plagued the earlier iterations of the Gates belt systems, the converted Ritchey has been relatively trouble-free. It's not a particularly stiff frame, and the horizontal drop-outs I welded in certainly do not meet any reasonable spec for alignment. The only issues I've had are an occasional slip of the hub due to not torquing the hub's bolts sufficiently. Otherwise, the belt-drive has been dead reliable, and quiet!!!

My trio of bicycles also includes a (used) 2018 Orbea Occam...full-suspension and single-speed. I rode this for two years with a conventional chain and Rohloff tensioner. It mainly sees miles on routes with lots of technical singletrack. For my Mountain 420 ride in Tennessee in 2023, I didn't have any Squirt to top off my mini-bottle prior to the ride. I did have some Maxima chain lube which was supposedly paraffin-based. And since I soak my chains in paraffin wax, I didn't think much of this fact. Unfortunately, whatever solvent Maxima uses to suspend the paraffin in the lube managed to solvate the paraffin that was on/in the rollers of the chain, completely stripping the chain of its paraffin about halfway through the ride. And, where Squirt would typically give me about 100 miles of drivetrain quietness, the Maxima lube was cooked in about 30. I quickly ran through the remainder of the lube, and spent the last 200 miles pedaling a Lowes shopping cart. 

Fast-forward about 9 months, I began contemplating a tensioner that would allow me to run a Gates belt-drive on the Orbea. No aftermarket tensioner exists for this purpose, and the few full-suspension bikes that utilize a Gates belt incorporate a tensioner mounted at the chainring/gearbox (i.e. Zerode). Lacking an elevated chainstay on the drive-side, utilizing the Zerode tensioner at the chainring is not an option. So, I dug into my box of derailleurs I never use, and began hacking them up to prototype a tensioner. Would this even be possible, especially with the high belt-tension you typically need to prevent the belt from skipping? 

Here, I've taken a SRAM Force derailleur, removed its pulleys, and made a custom carrier for a Gates 22t cog. Version 1.0 provides tension to the toothed-side of the belt, and the belt tracks quietly. Unfortunately, it does not provide enough tension to prevent the belt from skipping on the cog under moderate pedaling force. No bueno.

 

Here, I've added an idler pulley to add some tension against the backside of the belt and increase the amount of belt "wrap" around the cog. Unfortunately, Version 1.1 was also insufficient to prevent belt skipping under moderate loads.

Using my head a little bit here...some research turned up a "snubber wheel" solution for Rohloff internally-geared hubs. Apparently, the high-belt tension typically required for the Gates belt was destroying the bearings within Rohloff hubs. Their solution was to reduce the belt tension, and run a small wheel on the backside of the belt to prevent the belt from riding up the cog. I created a "snubber wheel" using a skateboard bearing (and other aluminum scraps I had in the garage). Version 1.0a worked...I was unable to make the belt skip no matter how much load I put through the pedals.

Not an AI image. Here, I tried to get fancy and remove the "parallelogram" from the derailleur. In doing so, I discovered the links of the derailleur are not a parallelogram, as its movement (when shifting) is not along a simple plane. Version 2.0 here required an inordinate amount of hand-filing to create, and in the end, I was unable to fine-tune the alignment to my liking (i.e. too much belt noise). So, I reverted back to Version 1.0a.

 

The maiden voyage of Version 1.0a was the Pisgah 99. Failure here would guarantee me at least a half-day's walk to get back to the cabin. It didn't fail.

Lots of unlawful stuff to disturb our National forests are in the works, I'm afraid.

The prototype was a success. While the tensioner was not sufficient on its own to provide enough tension to keep the belt from skipping, the addition of a "snubber wheel" on the backside of the cog prevented the belt from riding up the cog under heavy pedaling loads. While I didn't drop the belt during the ride, there were some disconcerting "thuds" that suggested suspension movement under harsh terrain was causing the tensioner to "loft" as the belt was quickly loaded. The SRAM Force derailleur clearly wasn't designed for this.

I sourced a SB One tensioner from Germany as a potential solution. This is a single-speed tensioner designed for downhill mountain bikes, with adjustable tension and an adjustable clutch mechanism. Its construction is impressive but it, too, isn't ready to go (for a belt) right out of the box.

The SB One tensioner is designed for small cogs. Downhill MTBs go...downhill. My rear cog is significantly larger than what the SB One was designed for. Here, I've roughed out the basic shape of the adapter on the mill. This locates the SB One far enough forward to not be detrimental to belt-wrap around the cog, and incorporated a mount for the snubber wheel behind the cog...

...like so.

The snubber wheel can also be pushed 'outboard' to the cog so that the wheel can be easily removed.

A few pics of the initial roughing, and the final massaging.

Measure thrice, cut twice.

I also replaced the 22t tensioner cog with a 20t cog. 

Quite a bit of work for 20 grams. I'm not well equipped to turn or mill stainless on my machines. I'd like to find an aluminum version of this cog. Or an .stl file...maybe try a 3D printed cog in the future???

Verson 3.1 is not lightweight.

But it looks presentable.

And it works.

But something seems off...

Don't need that any longer...now the overall weight is comparable to a high-end derailleur.

Much better.

 42x28 and a custom spiderweb-clearing stick.

The tensioner worked flawlessly across the Virgina Rockstar and the TNGA during 2024...hundreds of miles of quiet operation, nearly maintenance-free*, keeping my (in)sanity in check. Only things I plan to change are 1) anodize the adapter mount black and 2) lighter tensioner pulley (at some point).

*Gates belts neither like mud nor ice, as these clog up the center-track of the belt. Fortunately, water is about all the maintenance they need during a 300+ mile bike ride. Always carry a spare belt...because of Murphy.

Race Report: HuRaCaN 15

The highest density of bad golfers in the U.S. live right there in the middle.

I tried not to think about HuRaCaN for the week following CFiTT, instead focusing on recovering and getting fatter on Christmas treats. The first weekend in January was planned for The Spanish, a 525 mile route from St. Augustine to Fort DeSoto, which I was going to ride on a conventional single-speed...coasting from coast-to-coast. I was still committed to ride the HuRaCaN fixed, but promised myself it would be the last time. After all, I'm just here to grab low-hanging fruit. Nobody's done it, set the bar low, then sit back and revel in the suffering of others who come afterwards (that's my motivation).

The HuRaCaN is Singletrack Samurai's (Karlos) marquee event. Held on the first weekend of February, multiple route options (Lite 335mi, Original 380mi, WTF 400mi) and nearly 200 attendees, including some of the fastest ultra-endurance riders in the southeast and lots of adventurers and party-pacers. It feels like a reunion of sorts, lots of comfort in seeing familiar faces.

Joe (left) and Matt (middle); idiot (right). 

I camp out with the usual suspects. Joe Wharton and I have pedaled hundreds of miles together across Florida. Some days we're competing, other days one of us is having a bad day. Joe would finish the Original route in first this year! Matt Mustin is another good friend who's shared hundreds of miles riding about Tennessee, North Carolina and Virginia. He's recently taken to single-speed riding and this is his first HuRaCaN on one gear. We rode together for about half of the route. Matt took a wrong turn early on and rode the Diggins loop backwards (Santos) and didn't catch up to me until the Ghost Trail. We rode together through ONF, Seminole and eventually crashed a Waffle House in Apopka; he then sought out a sheltered location to take a nap and I rode on towards Clermont. We would ride separately for the next 22 hours or so, probably never more than about 10 miles apart. Matt would catch me again in the Croom singletrack, and we rode the remaining 70 miles together to the finish.

Maggie Jones Rd (photo: Matt Mustin)

Tricycle, homestretch in Santos (photo: Matt Mustin)

Entering Shangri La for the finish (photo: Joe Wharton)

Matt and I finished in 48h:39m (45h:39m including the 3-hour bonus for the Ghost Trail), unofficially tied for 4th place, and unofficial fixed-gear FKT. The singletrack in Santos, Paisley Trails and Croom (all but a few trails) were low-tech and could be ridden fast(!) on a fixed-gear. But, like with CFiTT, riding fixed over 380+ miles was an extra level of hurt from not being able to coast, not being able to rest your legs, and any attempt to get pressure off your butt requiring extra leg-fatigue. It's done and I'm done with it. 

It was great to see lots of friends and familiar faces. Big congrats for Joe to pull off the win (unofficial) after a frustrating and challenging previous year. Peter Kraft set a tremendous FKT on the WTF route. Ben Carpenter picked up the win on the Lite route, and Mike Root was laying down a blistering time on the Original route before a nail killed his tire. I leap-frogged with Tim Fahy and Kristin Apotsos throughout the race, which kept the competition interesting...Matt and I got the edge by taking a short nap and getting out of Lake Lindsey while Tim and Kristin were bivied up. Gary Cors pre-empted his weekend with an ITT on the Original route, finishing on Friday; then, started the Lite route on Saturday...700 miles of riding! He set a high bar with his 'Cat 2 HuRaCaN' and is hoping somebody comes out to try a 'Cat 3'. That's the kind of stupidity that piques my interest! Chritine Stone knocked out her 2nd ride towards completing her Samurai Cross within a calendar year...one long ride to go! And, always great to see Rasch, Andy Roberts, Kim Ranallo, Chris Hudson, Eric Henderson, and Jerry Metz! Great riding, all of you!

Obligatory selfies at Shockley Heights, Suncreek Brewery and Ridge Manor.

Selfies at Lake Lindsey Mall and Shangri La Campground finish.

I'll rochambeau you for it.

  

Race Report: CFiTT 2024 (CFiXED)

 

The Cross Florida Individual Time Trial (CFiTT) rolls out on the 2nd weekend in December, begins at dawn (usually) in New Smyrna Beach with a rear-wheel dip in the Atlantic Ocean, and ends with a front-wheel dip in the Gulf of Mexico (MEXICO) near Inglis/Clear River. In recent years, the ferry at St. Johns River has been out of service, and the route has looped around Palatka before dropping into the Ocala National Forest, totaling about 300 miles. This year, 'Ranger' Tim Fahy offered to rent a pontoon boat and ferry riders across the St. Johns, shortening the route to 245 miles. This seemed like a reasonable trade-off for my first off-road fixed-gear ride...if it all went to heck, at least I'm not in it for many miles.

The weather in New Smyrna was cool and dry; forecast for Saturday would be similar, but a chance of light showers in the late morning.

The starting and ending wheel-dips, Atlantic to Gulf.

I don't have any other photos of the ride because it sucked. Good grief, I've taken the privilege of coasting for granted. Those 245 miles riding fixed-gear felt like 400. A lot of that feeling was due to not being able to coast, but some of it was due to a conservative tire choice that back-fired. I opted for a set of 2.6" Bontrager SE2/XR2 tires instead of my usual 2.1/2.2" Schwalbe Thunder Burts. I figured the added tire volume would give me some leeway to bulldoze my way around Chuck Lennon and Santos, as I wouldn't be able to coast and finesse my way around roots and rocks. The downside is they roll slower than the Thunder Burts, and running higher pressures to reduce the rolling resistance caused an unbearably harsh ride. My concerns about finesse through Santos were overblown...riding fixed on a rigid MTB on those trails is not an issue. But the harsh ride took its toll and I struggled to maintain interest to even finish the ride during the last 30 miles. My butt hurt that much. Even my seat wasn't happy, as my inability to stand (and coast) to unload the saddle over harsh terrain was impaired, which led to bending the seat rails. Nonetheless, I ground it out and wrapped it up, and set a fixed-gear FKT in the process (27h:21m). And, I have a taste of what's to come in February.

Tech Tuesday: Fixed-Gear MTB

The end of the Fall Semester drew near which usually means a new season of Florida ultra-racing, starting with the Cross Florida individual Time Trial (CFiTT) early in December. The last few years of strong finishes inched me closer to my goal of a sub-24 hour ride on the 300 mile route. But not this year...I hadn't ridden much since TNGA in August, so I'm fat and out of shape. Let's be honest...I was fat and out of shape for TNGA, too. My waning motivation was further defeated by the news of a ferry ride across St. Johns River, shortening the route to ~245 miles. Granted, this was an optional bypass, but the race start was pushed up to 4am to help as many riders get to the boat ride as possible, which meant even a sub-24h attempt on the long course wouldn't be racing the sun-rise as usual. Again, I'm fat and out of shape, and now debating whether to even ride CFiTT. Or HuRaCaN.

Suffering is usually enough for motivation, but I just wasn't feeling it. However, a stupid idea popped into mind. I remembered that Corey Hilliard came down from New York to ride the 2024 HuRaCaN on a fixed gear. Crazy! I cut my chops in competitive cycling riding the track back in the late 90s, and continued to ride fixed-gears exclusively for the next decade...until I started mountain biking (albeit, single-speed). I heard of others riding fixed off-road, but hadn't personally considered it, and I was excited that Corey might be the first to accomplish it. He seemed prepared -physically and mentally- for the challenge ahead but, alas, did not complete the route. I was disappointed, perhaps most for thinking I could've done it. So, why not try it? And, why not try it first on the shorter, flatter CFiTT route?

The conversion of a single-speed MTB to fixed-gear isn't a difficult task. The easiest method is to flip the rear wheel around, remove the brake rotor from the 6-bolt ISO mount, and attach a bolt-on cog in its place. Fortunately, Gates makes a 22-tooth combo ISO/HG cog that allows it to either bolt it to the ISO rotor mount, or use it on the HG driver (freewheel) when I return to normal, sane, single-speeding.

 Gates ISO/HG cog, bolted to the hub.

The next issue is what to do about brakes. Granted, it's a fixed-gear...the pedals are always moving when the wheel is moving. Slowing your pedaling, or even skip-stops, could suffice. But, I'd be riding a loaded bicycle and, after a couple-hundred miles, I might not be keen on tasking my knees with that duty. I rummaged through my bin of parts and found an XTR cassette I no longer use. I drilled out the rivets on the carrier and removed its cogs, then turned down the arms of the carrier on the lathe, leaving enough material to tap for M5 hardware. I then drilled a new bolt-pattern on the brake rotor to match the pattern on the carrier.

The rotor/adapter mount to the cassette body, which is now on the non-drive side of the frame. While moving forward, braking forces are applied to the cassette body in the same direction as pedaling forces would otherwise be applied, so in that direction it works flawlessly. However, if rolling backwards, applying the rear brake just causes the hub to freewheel. Consider you have one large gear and you stall out on a very steep climb...I'm thinking about sections of Nayles North in Santos while zombie-riding with cooked legs. 

There, I fixed it. I pedal, wheel spins...wheel spins, I pedal.