I wasn’t aware anyone thought e-bikes couldn’t have regenerative braking, but they’ve existed for years and just weren’t the norm because it wasn’t really worth the added cost, weight, and complexity compared to the small amount of range it added.
From what I understood, it mostly just saved you wear and tear on brakes during long descents (as long as your battery wasn’t fully charged of course). The article touches on this towards the end, but it should be no great mystery that regen is not standard on e-bikes.
They certainly can, it’s just harder to capture the energy so they tend to be less effective than in large vehicles. It’s enough to extend range notably (and cheaper than adding corresponding battery capacity) but it’s also more engineering effort so not always included
Can this be retrofit, or is it pretty much must-be-designed-in?
Also, I typically don’t use my brakes 90% of the time (preferring to coast to slow down, turtle-throttle up to stops). The exception is returning home as there’s a slight hill… so I’d definitely trade in my front brake for regen (front wheel hub unoccupied) just for non-mechanical braking. (I do wonder about scooter motors and braking force, then again even that is probably no chance)
Depends on the bike and the system I imagine but I have not looked into converting a regular bike or e-bike that lacked regen. The conversion would probably be expensive though.
For sure, I was thinking more specific to hub motors themselves. Like if some sort of mounting plate could be added, or if it’s more complex than that.
Setting aside the headline’s alarmism about “lies”, the method described for implementing regen capabilities is very clever. From an electronic perspective, it shouldn’t be hard for more ESCs to support returning energy to the battery pack, a feature that would be unused with freewheeling rear wheels.
That said, changing the conventional design of a hub motor to enable this does come with some engineering challenges. For instance, existing designs have only three protrusions through the motor case: the axle out the left side, the axle out the right side, and the inside of the axle out the right side that carries the power wires. Of these, the axle protrusions are rotating w.r.t. to the case, so these need some sort of dynamic seal, such as O-rings, to keep dirt and water out.
It is ideal for O-rings to be small, because for any constant angular speed, a smaller diameter means less linear velocity. Linear velocity is the speed that the case slides beneath the O-ring, and more speed means more friction. We want less friction if we can.
For the axle O-rings, conventional designs only need them to be a bit larger than the bearing that supports the axle, so maybe 2-3 cm wide. But for this design where the left side is attached to the planetary gear carrier, the necessary O-ring needs to be as large as that carrier’s bearing, which is typically as wide as the motor, so maybe 15 cm. That is a 5x increase in linear velocity. Can that be solved? Yes, but it does add to costs.
Separately, I’m wondering how sturdy this design is vis-a-vis brake performance. As the article says, the conventional design has the brake directly attached to the motor case, and the case directly attached to the spokes. So applying the rear brake creates torque through the motor case. With this regen-capable design, the torque now also always goes through the planetary gear carrier, and either through the inner teeth of the motor case or through the planetary gears to the motor.
I’m not too concerned about the planetary gears or the motor, because unless the motor is jammed shut, the ESC can vary the amount of regen, which limits the torque to these gears. I’m more concerned about the teeth to the motor case, because braking almost always causes more torque than accelerating. Will those teeth hold up, or do they need to be reinforced? If the latter, then that’s more cost and more weight to construct this design.
If hard braking could strip these teeth, that’s a big problem for something like a cargo ebike, which is a use-case where regen makes a lot of sense, because bike geometry will not limit the max braking from the rear wheel.
Overall, I do think it’s clever, but it probably won’t appear in standard-geometry ebikes, because it adds cost for something that isn’t too relevant to most riders. But for all other shapes of ebikes, it’s plausible that this becomes an option. I could even see e-assist bike trailers using these as their hub motors, because for any serious payloads on a bike trailer, brakes should be designed in.
I had considered getting the ENGWE Engine Pro because it had regen, but I went with the simpler EP2 Pro and don’t really regret it (though I will admit, the descent control regen offers does seem like it’d be nice sometimes).
As an engineer, I would be thrilled to have more features to play with. But also as an engineer, I’m trying to imagine the sort of downhill ride that regen would be used for. I consider three sorts of hill: 1) a fairly shallow hill where air and rolling resistance perfectly balances downhill force at a speed which is approximately the road speed limit, 2) a steeper hill where at least one brake or regen must be used, or else the bike will overspeed, and 3) an exceedingly steep hill that requires both brakes or max regen in order to maintain control of the bike.
Scenario 1 is easy: no brakes nor regen are needed. Scenario 2 is ideal for regen, but only if the battery has capacity to absorb the energy and that regen can be modulated to achieve equilibrium at a road-legal speed. Scenario 3 is where regen becomes dangerous, because too much will lock one wheel, although these sorts of hills are dangerous outright. Exceptional modulation of regen would be necessary here, and even then, I personally would still guard the opposite brake handle, because if regen acts up, there may be little time left to react.
Most of my riding is in scenario 1, with very few sweeping hills of scenario 2. And scenario 3 only really shows up for me on dirt tracks, where I already have to be on both brakes or else the bike will slide.
So I’m of the opinion that regen would only shine for non-standard bike geometries, or for paved road hills that can safely recoup the energy. Sadly, both do not apply to me at the moment.
Why is the bike geometry a factor?
I may have overloaded the term, but what I meant was different shapes of bicycles and bicycle accessories, but most notably by wheel count. So a trike would be a fairly good candidate for regen, since the (typically) rear axle will have two wheels and thus can recoup more energy from regen, where the risk of skidding the rear is less likely to cause the bicycle to topple over. Likewise, a bike trailer with its own assist motors can do a lot of deceleration, although setting up trailer braking is an art unto itself.
I think earlier I used “bike geometry” to refer to a bike that can flip over-the-handlebars, so that’s my bad for being confusing.
FYI - Trikes (even recumbents) can still definitely flip over the bars (or lack thereof), trikes are paradoxically less stable at speed than a bicycle, but I get what you’re saying about regen. I have a recumbent reverse trike I ride sometimes and I’ve flipped it cornering at maybe 12-15 mph (whereas on a bike I can easily corner at 20+).
trikes are paradoxically less stable at speed than a bicycle
100%! That they often have a high center of gravity and narrow wheelbase doesn’t help things at all. But I’m all-in on the idea that a healthy bicycle economy means having all sorts of bikes available for purchase.
Grin figured this out a bit ago! Glad to see other companies picking it up too.
My old copenhagen wheel was regenerative breaking, activated by backpedaling.
My cake osa has it too, but you have to choose between it and freewheeling (whenever you let go of the throttle, it begins braking, which you can turn off by changing the braking profile)
I want to put regen braking in my socks
Charge your phone with your socks!
So glad to see another individual with vision
