niboreriam

@hamncheez: There absolutely is a generator phase, and the article itself points it out. To provide a higher gear ratio (overdrive) when you are above the legal assist speed limit, the system cannot pull power from the battery to spin the second motor, because that would be illegal motor assistance. Instead, it must use the torque motor as a generator, driven by your pedaling, to power the speed motor. This mechanical-to-electrical-to-mechanical conversion loop is exactly what destroys the overall efficiency.

@Mtbdialed: You are 100% right, and your argument actually proves my overall point. A simple software micro-cut in the motor's torque, exactly like a motorcycle quickshifter, solves the shifting-under-load issue on a standard MGU. So why adopt an incredibly complex, heavy, and inefficient dual-motor eCVT system when a few lines of code on a classic mechanical gearbox fix the main issue perfectly?

@phops: Splitting the load across two motors is not more efficient. You are now running two stators, generating two sources of heat, dealing with double the copper wire losses, plus the constant mechanical drag of the planetary gearset. Thermodynamically, it is worse. Regarding the dead battery: yes, pedaling a 50 lb ebike without assistance is hard. But with a standard gearbox or derailleur, you can still shift into the granny gear and spin your way up a hill. If an eCVT dies, it has no power to maintain the virtual gear ratio. You will be stuck pushing a 50 lb bike in a random, likely very hard gear. That is a massive difference in real-world trail riding.

This article gets excited about trade show talk, but it misses the real constraints. The eCVT on a performance mountain bike poses big problems. First, the efficiency. A power-split eCVT (like the one in the Prius, which inspires these systems) has an intrinsically lower efficiency than a classic mechanical transmission, it is historically recognized: as soon as part of the power goes through the electrical path (motor -> generator -> motor), we accumulate losses that do not exist on a direct mechanical path. And we must not forget to add the chain or the belt at the output to compare apples with apples. As a result, afterwards you have to oversize to compensate, even if it means putting out 1500-watt motors that rip your hands off, as the article itself reports. Losing a significant part of the rider's and battery's energy in heat and friction on a mountain bike is starting to be a lot, not to mention that two motors running continuously (including one potentially as a permanent generator above the assistance limit) also heat up much more than a classic motor that only works in bursts. On the energy recovery mentioned in the article, the orders of magnitude are misunderstood, it is marketing. In fact, there will never be enough juice recovered to recharge anything beyond the bike lights eventually, not even a smartphone. It is negligible, the mass and speeds involved on a bike have nothing to do with a car. Next, the article thinks that the eCVT is the solution because current MGU gearboxes sometimes struggle to shift gears under load, and it is right. The truth is that designing a compact and reliable mechanical gearbox is extremely difficult, but not impossible. We just need to stop demanding gear shifts under a disproportionate load and accept to ease up on the pedaling for a fraction of a second, like for a derailleur at the price of a rent, which does not prevent it from happening under load on all gears, I have lived and felt it myself. The eCVT bypasses the problem instead of solving it. It replaces a mechanical difficulty with a heavy block, which needs a complex software and overpowered motors just to simulate the gears, with the bonus of power electronics that are much more expensive to produce and almost impossible to repair simply. Finally, the article misses a crucial detail about running out of battery. On a classic gearbox system (MGU), the 0% energy reserve is just used to trigger the gear shift for a fraction of a second, we are talking about a small motor and small actuators. So we can continue to shift gears to get back home. On an eCVT, the planetary gearset needs to consume electricity continuously just to maintain the engaged gear. The small reserve will empty much faster and the bike will freeze on a gear ratio that is impossible to pedal (unless the problem is solved...).

You can really spot the automotive engineer syndrome here: used to having a massive engine under the hood to mask power losses, and mechanics safely tucked away inside a sealed housing. But on a mountain bike, you can't get away with that. Out on the trail, it's a triple technical disaster: Replacing a solid mechanical engagement (teeth on rollers with ~98% efficiency) with lateral friction is a joke when the human engine only cranks out 250W. The power loss while pedaling will be massive. With a constantly changing front chainring diameter, you have no more control over anti-squat than you do with a standard cassette. That is precisely the perk of existing MTB gearboxes: the exact same anti-squat curve regardless of the gear, for better or worse. Optimizing a main or virtual pivot point in these conditions becomes a massive headache... Smooth flanges sliding in the open air, inches from the ground, in the dust and mud of our trails? Slippage is guaranteed at the first mud puddle, along with extremely fast wear for that fancy proprietary belt. Sure, you could put a cover on it, but then you're building a tank... Bottom line: We are just relocating an existing problem to the bottom bracket, ending up with fewer benefits than current gearboxes, and arguably worse execution. At least on paper anyway; maybe a more refined version will drop someday...

An impressive technological feat on paper, but the conclusion of this article is alarmingly naive. Reducing the risk of trail closures to a simple "be courteous and don't be a jerk" is reductive at best, and intellectually dishonest at worst. The real danger doesn't come from respectful riders, but from the general public's perception. A local politician, a hunter, or a hiker isn't reading Avinox spec sheets to tell the difference between nominal and peak power. When faced with a machine climbing at 25 km/h with no apparent effort and a turbine noise, they will see a light motorcycle, end of the game. Political inertia means it is infinitely easier to sign a local decree banning access to a forest than it is to lift that restriction years later. By releasing 1500 W and 150 Nm monsters (which, incidentally, are a mechanical heresy that will shred our traditional drivetrains), the industry is handing them the stick to beat us with. If regulations hit, they will punish everyone blindly, including acoustic mountain bikes and lightweight e-bikes.

I have been a massive fan since the Coastal Crew days. I loved their approach to kinematics and eventually bought a brand new Warden, which was my absolute dream bike at the time. However, starting my career as an engineer in the MTB industry made me realize why Knolly is unfortunately a textbook example of what not to do. They stubbornly stuck to geometries that appeal to almost nobody. Worse, they completely missed the e-bike train. Even if an e-bike was among the new models in the pipeline for this year, it came way too late. Instead of developing one when it was crucial, they chose to recently "modernize" their entire acoustic lineup. It feels like they keep releasing the same bikes without any clear and radical evolution. It's not that there is no added value (I was a customer!), but the brand seems locked in an outdated dogma and a vision that just doesn't sell anymore. The post-COVID market correction has undeniably been brutal for everyone, but I refuse to believe it is the sole reason for this situation. I've known companies struggling in this sector during and after COVID, and it honestly breaks my heart to see Knolly in trouble, truly. But I share a more "Porsche" approach to things: sell some Cayenne and Macan to have fun with 911, and save the company...

@cxfahrer2: Assembly line isn't manufacturing line, Hepha frames are manufactured in Asia, and that's not a bad thing imo, regarding the pricepoint they are offering ;)

The motor is a rebranded Gobao P120, same as Hepha, who was in partnership with Gobao in a same way, back in the days.

Homeplace 3 And that's not "Tartes des mirtilles" but "Tarte aux mirtilles" haha