RE: asterisked part (tried to use BB code for holding, but seems they cripple the forum software here…). I don’t think there is any hardware improvements of note.
Looking at the pair of replacement torque sensors (this took a while to find, came up empty on ali and elsewhere for the M600 version..), they look pretty similar, although without real specs (how many times/second a signal is sent etc., or if it’s constant and in the controller to sample appropriately…don’t know), tough to guess. The M600 uses the same Tq sensor as the M500, while the Ultra’s is seemingly it’s own part, but same wire count, same plug, seemingly near identical construction - where manufacturing is involved, it is HIGHLY beneficial to source the same component parts as much as possible for inventory, manufacturing and general cost savings, so I’d wager the internals are probably identical.
bafang ultra torque sensor / Bafang Ultra M620 torque sensor with axle for replacement.
www.greenbikekit.com
bafang m600 m500 m510 m420 G332 mid motor torque sensor for replacement
www.greenbikekit.com
I’m not sure why you’re focusing on mechanical advantage here, at least in the motor. The controllers used by Bafang are pretty simple, but can do ‘ok’ when able to be adjusted by users. The Archon/Innotrace, Phaserunner, Ludi v2/VESC and the like are sine wave, FOC controllers, possibly with more compute power (although I’d expect them all to be relatively low in compute - a handful of signals inbound being processed N times/second doesn’t need all that much, but things like FOC code and not-exposed motor signals/sensors adds to this), and are in general - are better all around. They still rely on the same underlying base motors and sensors (although some augment, e.g. GPS, gyro, etc.). There are levels of signal processing being done inside the controller firmware which may be a level of ‘secret sauce’ (such as field weakening algorithms, or the exact matrix of input values vs output actions) and those things not currently exposed via the configuration (you’re not really ‘programming’ the controller, just setting parameters/values) options exposed to the user.
The ‘mechanical’ advantage I believe is solely down to the mass of the 620 in dealing with thermals and it’s ability to take more power - if that’s important to someone. Meanwhile, the non-mechanical advantage is down to the controller and ability to move beyond what comes out of the box from Bafang. It’s possible, for example, the tq sensor provides a constant signal (I don’t know for sure - it’s difficult to locate actual technical docs on Bafang stuff, and haven’t broken out an o’scope and meters - maybe some year when I have spare motors to play with - although I’m SURE someone has..), but the Bafang controller is under-specced compute-wise, so it samples/considers only a few times per second, or acts on it within a few times//second window at best. Someone at a 90-rpm cadence is rotating the crank 420 degrees every second, so for example, if the bafang stock controller is only reading or acting on torque sensor input 4x a second, that would mean minimum crank rotation of 105*, or a bit less than 1/3rd crank rotation, would be required before able to add assistance. Conversely, the Innotrace/Archon/Ludiv2 may be reading the exact same signal many more times per second, giving the capability to react ‘nearly instantly,’ subject to whatever algorithms/behavior or configuration is applied within the controller.
The above is just an example - would love to know the real specs on internal components, but you can hopefully see how an aftermarket controller COULD make quite a bit of difference…and as a bonus, there are some efficiency gains for both the Innotrace and Ludi controllers over the OE.
