Current Avinox MG Prototype Bikes Don't Support Battery Regen

[cream]

So most of us aggressively take sides to what? To something that didn't hit the market yet? Where's the rush?

As @Fangs2k said e-cvt can coexist with standard systems. And it won't be cassettes/derailleurs available only for low end spec bikes. There's yet plenty to be made from both types of transmissions.

Even now, you can buy an M3 CS with a manual transmission, and it's not cheap


All in all, I'm interested in how it feels...if it parts from the normal bike feel, then it's not for me. At least nor for my eMTB. I might use it on a touring/city/commuting where I would be more interested of the system efficiency at the expense of how I feel it. Until I have tried it myself I am a bit reserved. It's how I'm wired don't judge me, just hear me out, ok ?

Regen in an electric MOUNTAIN BIKE also not appealing to me. Why would I change something that would impact how the bike will feel, complicate the predictions my brain could make on the trail.

But the e-cvt is a nice electro/mechanial device. Still haven't found an answer on how the system will function with no battery or a fully depleted one. Anybody ?

 

[Powerslider]

So most of us aggressively take sides to what? To something that didn't hit the market yet? Where's the rush?

As @Fangs2k said e-cvt can coexist with standard systems. And it won't be cassettes/derailleurs available only for low end spec bikes. There's yet plenty to be made from both types of transmissions.

Even now, you can buy an M3 CS with a manual transmission, and it's not cheap


All in all, I'm interested in how it feels...if it parts from the normal bike feel, then it's not for me. At least nor for my eMTB. I might use it on a touring/city/commuting where I would be more interested of the system efficiency at the expense of how I feel it. Until I have tried it myself I am a bit reserved. It's how I'm wired don't judge me, just hear me out, ok ?

Regen in an electric MOUNTAIN BIKE also not appealing to me. Why would I change something that would impact how the bike will feel, complicate the predictions my brain could make on the trail.

But the e-cvt is a nice electro/mechanial device. Still haven't found an answer on how the system will function with no battery or a fully depleted one. Anybody ?

Specialized's current Turbo Levo system walls off a small part of the battery so the derailleur stays powered even if the battery is depleted. This would work the same way. Nothing new really.

 

[sethimus]

There doesn’t need to be an alternative does there? The derailleur and cassette etc has served cycling well for decades and is simple, effective and cheap. People don’t have to buy SRAM XO groupsets at the end of the day.

I think like everything in cycling, there’s room for all kinds of difference, why can’t MGUs exist alongside the current default solutions?

I know the answer, and it’s got little to do with solving problems but everything to do with keeping factories busy and retailers tills ringing.

I’m not a luddite, I’m interested in MGUs but not because I struggle to be in the right gear at the bottom of a descent or start of a climb or even because the derailleur is ‘ugly’. Mind you, some of those belt and tensioner setups don’t look great either!

I’ve had a bit of time on one of my brothers e-commuter bikes, a Leonardo with a Gen3 Zehus hub motor with regen (yes, he lives in London and has that kind of moustache and lifestyle!) and it’s a cool thing for the cycleways of that city.

He reckons he can add about 10% of battery range just by normal braking regen, without any back pedalling. Doesn’t sound a lot to translate into much extra range for an EMTB even if the technical challenges to enable it can be overcome with a full suspension bike.

I’m often reminded of the legendary engineer and designer Colin Chapman whose mantra was always ‘simplify, then add lightness’. It feels like MGUs are doing neither, replacing one set of complexities with another.

I’ll be watching with interest though, to see how each of these early concepts develop over time.

if you don’t see the advantages you need to ride harder trails. simple as that. in my last finale holiday i needed to re-adjust my axs derailleur 4x due to hitting stuff on the trails…

 

[Fangs2k]

if you don’t see the advantages you need to ride harder trails. simple as that. in my last finale holiday i needed to re-adjust my axs derailleur 4x due to hitting stuff on the trails…

Yet I did a week in the mountains around Kaprun and didn’t need to touch anything?

I’m not saying I don’t see advantages, but I also see downsides, why is that so controversial?

We’ve been waiting for a replacement for derailleurs for years, this might be the time a viable alternative for some bikes and riders emerges, but it won’t be universal and won’t be for everybody.

 

[TorAtle]

I asked AI how much regen force a Gobao X1P could possibly generate. 0.4G peak, apparently. There's a lot of variables and the math could be off. But if true, that's pretty good.. So paired with a hydraulic front brake, we'd probably only need a one pot 140mm rear brake. The pads would last for ages.

 

[Astro66]

PS. Not sure what the heck you are doing to your derailleurs! I used to break them often when I was a novice but as time went on I learned to not smash them into things.

I'm not breaking them. I'm bending the cages.

I'll give you an example. I rode across a narrow timber plank that had been placed on the trail to cross a wet patch. The person in front slowed, so I needed to drop off the edge unexpectedly, just before the end of the plank. The rear wheel dropped off to the left, and the derailleur caught the edge of the plank, pushing it upwards and bending it like a banana.

Immediately I noticed gear shifts were poor, and the chain was jumping up and down on the sprockets, ruining the day, and requiring a new cage.

Another example. Crossing a trail where you had to ride over the top of several rounded rocks. One rock was quite mossy, so the rear wheel slid off to left, impacting the derailleur cage as it slid off. Another bent cage.

All these incidents, I didn't fall or have an accident. It's just the rear wheel slid off an obstacle to the left, and with the derailleur cage so low to the ground on that side, it gets impacted.

Then of course there are the impacts when you are weaving between rocks in a narrow gully. I have learnt to lean the bike away from close rocks on the right. But if you are going hard. You don't always have that option. And I ride to push limits. And the derailleur is forcing me to slow down, so I can always avoid impacting it. The derailleur is reducing my ability to push limits.

 

[Astro66]

But the e-cvt is a nice electro/mechanial device. Still haven't found an answer on how the system will function with no battery or a fully depleted one. Anybody ?

Who rides their EMTB with no battery ? Who rides their EMTB battery to 0% ?

I have had EMTBs for 8 years. I have never run my battery to empty. As I run electronic shifting, powered by the motor battery, on my last 4 EMTBs. On the 2 occasions I have slightly misjudged the ride. I turn the pedal assist OFF at 2%. And just rode the last little bit of the trail on leg power, so I maintain power for the electronics.

You are going to run out of battery anyway. So why run the battery to zero ? Turn the pedal assist OFF before you get to zero. It's not rocket science.

BTW. On the Avinox Motors. The Smoothshift is turned off automatically, when you select no pedal assist. So the motor does not spool on gear changes when no pedal assist is selected.

 

[sethimus]

BTW. On the Avinox Motors. The Smoothshift is turned off automatically, when you select no pedal assist. So the motor does not spool on gear changes when no pedal assist is selected.

does it still shift?

 

[cream]

You are going to run out of battery anyway. So why run the battery to zero ? Turn the pedal assist OFF before you get to zero. It's not rocket science.

I saw something about Avinox getting to Mars so maybe it's rocket science...

I don't care what kind of science it is as long as you can explain to me how this Avinox e-CVT (which is a prototype that you haven't tried I'm sure) works in different modes of operation (inputs/outputs) and how it will shift or be locked in a specific gear ratio without using electricity (more than a servo on a AXS/DI2). If I turn off assistance at let's say 5%, whats' the range (let's call it shifting range)I would have before being unable to shift or pedal? I'm not saying it's not possible but you seem to have a better grasp of things, please share.
Also, you're being defensive no particular reason. I'm not a Bosch advocate, don't hurt me, please

BTW. On the Avinox Motors. The Smoothshift is turned off automatically, when you select no pedal assist. So the motor does not spool on gear changes when no pedal assist is selected.

Ok, so what? What's the relevance to this thread? You just wanted to freely educate me on another topic? In that case, thank you.

 

[slickrock]

Yes mesh gears mgu design has some defects but they can be resolved. It probably needs the parnership of a leading bike brand and an automotive engineering company to do that.

Both SRAM and Shimano have posted MGU patents last year, and these are of the gearbox variety not too dissimilar to Pinion. Let's see if these patents are just posturing or do they really intend to bring them to market. I really doubt it. And then there's Intradrive desperately trying to woo takers to their full gearbox MGU and they've been at it for how may years now? And then there's Pinion itself squeezing as much as it can with version 1.0 MGU. I've been waiting for a seriously improved V2.0, but nothing so far. For reasons I fully elaborated on the ECVT thread, ECVTs will will crush the fledgling gearbox-based MGU market. That Avinox realized this and went with ECVT is good enough reason alone.

You also have to question why the leading brands did not include mgus into their ranges of bikes. The same reasons may apply to adoption of ecvt mgus.

They haven't because gearbox MGUs are just so wanting: heavy, noisy, expensive, lossy, bulky, ungainly, and too mechanically complex. Most motor companies like to work with motors rather than complex gear carriages that are in gearboxes. They would rather stick with motors. ECVT is motor-laden (there are two of them in each housing). Same goes for pedal-by-wire: The Niche motor has two independent motor/generators, much like the TM-B. There are lot more of motor companies that pull off a simple planetary transmission than complex, costly, or more difficult to produce gearboxes.

 

[Rob Rides EMTB]

PS. Not sure what the heck you are doing to your derailleurs! I used to break them often when I was a novice but as time went on I learned to not smash them into things.

Mech damage is real. Novice right through to pro.

The issue with the current SRAM cages is that they are so prone to bending or moving out of alignment slightly with the lightest of scrapes on a rock, and this effects shifting at some point across the cassette range. Maybe only 1-2 gears, but the issue is real.

On bigger impacts the cage is too deformed to work properly.

In the past couple of years personally I've damaged around 8-10. I've also witnessed Pro riders blow up their cages multiple times on press camps; its not just a novice issue. At the Orbea press camp on the new Wild, at least 3-4 riders bent their cages, one totally wrapped around the cassette the wrong way (one of the Pyrenees guides, a pro!) and one of the Orbea designers also broke his.

Riding unfamiliar terrain where you just do not know the correct line and squeezing through a rock garden, or taking a hard right corner and scraping it on a tree stump. There's real issue that happen, that can and do happen to any skill level.

The current SRAM Transmission works very very well when new and perfectly setup. But its ridiculously fragile, and not fit for purpose when sending a big enduro rig down a mountain.

 

[Fangs2k]

The current SRAM Transmission works very very well when new and perfectly setup. But its ridiculously fragile, and not fit for purpose when sending a big enduro rig down a mountain.

Which is completely the opposite to what all the reviews said when Transmission came to market, where it was praised for being strong and people were standing on it.

Wonder why some of us are sceptical when the new big thing turns up.

Won’t some of the tensioner placements and designs be prone to damage and bending as well?

 

[DylanJM]

Won’t some of the tensioner placements and designs be prone to damage and bending as well?

Yes, but that is the case with anything on a bike. There's always a chance of something breaking.

Most of the gearbox solutions use a bottom bracket mounted solution which more out of the way than a derailleur. Even a dropout mounted tensioner would be far smaller and out of the way compared to a standard derailleur also so it would be still an improvement.

 

[Rob Rides EMTB]

Which is completely the opposite to what all the reviews said when Transmission came to market, where it was praised for being strong and people were standing on it.

The actual body of it is very strong - it can take a good whack and still work perfectly fine. The issue is the cages bend too easily.

Won’t some of the tensioner placements and designs be prone to damage and bending as well?

Possibly yes. But in my experience they are a lot stronger than the cages. The Gates one on my Nicolai was built like a tank.

This one on the new Commencal is also pretty small and way less prone than the mech cage sitting close to the ground.

 

[DylanJM]

That Commencal one is defo one of the nicer tensioner solutions I've seen. It's interesting that most of the Avinox protos were spec'd with chain over a belt too. I think this would be mt preference also. You would still save some maintenance compared to a normal drivetrain. Belts sounds great but some of the owner experience I have read indicate they are not the be all and end all some might have you believe.

 

[Fangs2k]

Yes, but that is the case with anything on a bike. There's always a chance of something breaking.

Absolutely, which is kind of my point, EMTBs will never be maintenance or mishap free.

I’m looking forward to reading the reviews of how they ride in the wild when they’re actually released.

 

[Astro66]

does it still shift?

Yes, the derailleur shifts. But you have to manually pedal the chain up or down the cassette.

 

[Ndanger]

, some of us are still rather attached to the mechanical simplicity of the traditional 'bike'.

I like the tangible idiosyncrasies of my mechanical drive train. It has a clicky clacky feedback and is soooo simple. Not sure that I will feel comfortable more than a days hike from the nearest road as temps drop at the end of the day with complex computerized systems. A smooth quiet system that eliminates my direct input would put me to sleep.

 

[patdam]

That Commencal one is defo one of the nicer tensioner solutions I've seen. It's interesting that most of the Avinox protos were spec'd with chain over a belt too. I think this would be mt preference also. You would still save some maintenance compared to a normal drivetrain. Belts sounds great but some of the owner experience I have read indicate they are not the be all and end all some might have you believe.

I have a belt on one motorcycle. That woorking very weel, until i ride on graveled road (ofen on moutain). Gravel gets into the pulley and punctures the belt; it needs to be replaced before it snaps. very painfull job (necessity to dismantle the frame). never a belt for me in TT.

 

[Astro66]

Ok, so what? What's the relevance to this thread? You just wanted to freely educate me on another topic? In that case, thank you.

The relevance was to your question. If you turn assist off. You won't be using the drive motor. Only the motor in the derailleur, which is a servo motor and only draws tiny amounts of energy.

I don't care what kind of science it is as long as you can explain to me how this Avinox e-CVT (which is a prototype that you haven't tried I'm sure) works in different modes of operation (inputs/outputs) and how it will shift or be locked in a specific gear ratio without using electricity (more than a servo on a AXS/DI2). If I turn off assistance at let's say 5%, whats' the range (let's call it shifting range)I would have before being unable to shift or pedal? I'm not saying it's not possible but you seem to have a better grasp of things, please share.
Also, you're being defensive no particular reason. I'm not a Bosch advocate, don't hurt me, please

No one is releasing exactly how their E-CVT works. But in a traditional e-cvt, one of the motors can act as a generator to recharge the battery. So I'm guessing that if you turn assist off, part of the energy from your legs will be to maintain the battery at it's current state of charge, to power the electronics, and maintain the ability to vary the gear ratio. Then the rest will go into driving the rear wheel.

As the E-CVT is like 95-97% efficient. The point is that if the battery is not being used to drive the main drive motor, and providing no assist. The feeling will be very similar to riding an unassisted derailleured EMTB, whilst still maintaining the ability to change gear ratio.

 

[Ndanger]

But in a traditional (non free wheel ) e-cvt, one of the motors can act as a generator to recharge the battery.

I took the liberty of adding a conditional item.

 

[slickrock]

I took the liberty of adding a conditional item.

Actually, in the particular use case Astro is talking about, you don't need a fixed wheel. There's no problem with human pedal power charging the battery with a free-wheel out back per inherent ECVT design. ECVT can transfer pedal power into either the drivetrain or the motor (acting as generator) or both at the same time, as in the case of running out of battery power. However, you do need a fixed wheel, plus tensioner-less belt/chain suspension (or hardtail) design if you want forward momentum to charge the battery.

 

[Fangs2k]

Actually, in the particular use case Astro is talking about, you don't need a fixed wheel. There's no problem with human pedal power charging the battery with a free-wheel out back per inherent ECVT design. ECVT can transfer pedal power into either the drivetrain or the motor (acting as generator) or both at the same time, as in the case of running out of battery power. However, you do need a fixed wheel, plus tensioner-less belt/chain suspension (or hardtail) design if you want forward momentum to charge the battery.

Welcome to 2026 folks, where chat is how humans can charge the battery to power the motor that moves the bike, rather than pedalling the bike directly.

I love progress!

I think we’re all a bit premature, we’ll know how the various manufacturers attempts at the tech works soon enough, most of us are all guessing at this point.

There’ll be tests from the usual suspects soon enough, then we’ll know, and we’ll see how quickly they scale from the announcements of ‘2027’ and see how they do longer term.

 

[Karve]

Mech damage is real. Novice right through to pro.

The issue with the current SRAM cages is that they are so prone to bending or moving out of alignment slightly with the lightest of scrapes on a rock, and this effects shifting at some point across the cassette range. Maybe only 1-2 gears, but the issue is real.

On bigger impacts the cage is too deformed to work properly.

In the past couple of years personally I've damaged around 8-10. I've also witnessed Pro riders blow up their cages multiple times on press camps; its not just a novice issue. At the Orbea press camp on the new Wild, at least 3-4 riders bent their cages, one totally wrapped around the cassette the wrong way (one of the Pyrenees guides, a pro!) and one of the Orbea designers also broke his.

Riding unfamiliar terrain where you just do not know the correct line and squeezing through a rock garden, or taking a hard right corner and scraping it on a tree stump. There's real issue that happen, that can and do happen to any skill level.

The current SRAM Transmission works very very well when new and perfectly setup. But its ridiculously fragile, and not fit for purpose when sending a big enduro rig down a mountain.

100% agree with this. The current mech setups are the ultimate definition of polishing a turd. I’m breaking them all the time, or riding with friends who are breaking theirs. They need to go. That said, I do think the mech will hang around — it is the right tool for XC and other endurance disciplines. Not sure where this leaves the big meat powered enduro rigs which need the protection a gearbox/cvt provides but dont need the weight or drag. As EMTB is driving innovation they may end up having to take whatever is built for XC

 

[slickrock]

Welcome to 2026 folks, where chat is how humans can charge the battery to power the motor that moves the bike, rather than pedalling the bike directly.

I love progress!

I think we’re all a bit premature, we’ll know how the various manufacturers attempts at the tech works soon enough, most of us are all guessing at this point.

There’ll be tests from the usual suspects soon enough, then we’ll know, and we’ll see how quickly they scale from the announcements of ‘2027’ and see how they do longer term.

Well that's just the beauty of Hybrid Drive ECT - there's really nothing premature abou it. How do you think hybrid vehicles (HEV) charge their batteries? The gas engine powers both wheels and charges battery. When you are idle at a stop sign, the gas engine stays on and charges the battery.

So in the case of a drained battery degrade mode, you can still pedal the bike to move the bike forward as well as drive the generator to keep enough battery juice on tap to keep the gearing and system running. Silly as it sounds, you could theoretically turn your ECVT into an indoor trainer and charge up the battery while you idly pedal the bike.

Battery, motor, pedaling, wheels are all interlocked with ECVT, so all of the following are possible and pretty much can happen all at the same time:

• Battery powering wheels
• Pedaling power wheels
• Pedaling charging battery
• Wheels charging battery

 

[Ndanger]

Actually, in the particular use case Astro is talking about, you don't need a fixed wheel. There's no problem with human pedal power charging the battery with a free-wheel out back per inherent ECVT design. ECVT can transfer pedal power into either the drivetrain or the motor (acting as generator) or both at the same time, as in the case of running out of battery power. However, you do need a fixed wheel, plus tensioner-less belt/chain suspension (or hardtail) design if you want forward momentum to charge the battery.

Thank you for follow up, and yes anytime the pedals rotate it provides the opportunity to create current. My belief is that it is the recovery of kinetic energy wasted on brakes to create electricity instead of heat is the practical application of regen.

Reality check A typical e-bike battery holds about 500 to 750 Watt-hours (Wh). To charge a dead 500Wh battery at a sustainable human output of 150W (yielding about 60W of stored power), you would need to pedal continuously for over 8 hours.

 

[Fangs2k]

Well that's just the beauty of Hybrid Drive ECT - there's really nothing premature abou it. How do you think hybrid vehicles (HEV) charge their batteries? The gas engine powers both wheels and charges battery. When you are idle at a stop sign, the gas engine stays on and charges the battery.

So in the case of a drained battery degrade mode, you can still pedal the bike to move the bike forward as well as drive the generator to keep enough battery juice on tap to keep the gearing and system running. Silly as it sounds, you could theoretically turn your ECVT into an indoor trainer and charge up the battery while you idly pedal the bike.

Battery, motor, pedaling, wheels are all interlocked with ECVT, so all of the following are possible and pretty much can happen all at the same time:

• Battery powering wheels
• Pedaling power wheels
• Pedaling charging battery
• Wheels charging battery

I know how series and parallel hybrids work, and I've owned an EV of some sort for over 10 years, so am embracing of ‘new’ technology.

The above are applications for motor vehicles though, I was relating it to cycling, and it was a light hearted comment at that.

Personally, it’s not a direction I’m particularly interested in for electrically assisted bicycles, but I do at least realise that the forum exists to discuss other types of E-bikes which is something I wish others would do.

I still find the tech interesting though which is why I’m keen to see the finished products.

 

[Paul F]

All the current announcements for the Avinox MG (and the Gobao XP1 motor as well), officially and through influencers, tout that their ECVT design can support battery regenation, which is actually the case due the inherent capabilities of ECVT Hybrid Drive design. However, I just realized with all these Avinox MG prototype bikes, in their current design form, battery regeneration would NOT be possible.

Not only do you require a fixed wheel hub (which is possible on these bikes) with the belt/chain always spinning, but the mere presence of chain tensioner disqualifies reverse tension on the belt/chain that is essential to battery regeneration. I forgot about this because I've had the Rivian/ALSO TM-B on my mind for a while now:

View attachment 187435

Instead you need to have a concentric wheel path rear suspension design (or just a plain hardtail like with a city bike). This design requires no tensioner (awesome in its own right) where the chain/belt can be at full tension always, and more importantly allow reverse tension with forces operating in the opposite direction from belt/chain rotation when the motor acting like a generator. The TM-B beautifully pulls this off because the drive axle is well above the pedal axle with the drive axle right behind the single pivot, which pulls off a HP design without an idler. True, it's pedal-by-wire instead of ECVT, but that doesn't matter. All that matters is the concentric wheel path (which exhibits zero chain growth) and motor that can act as a generator.

With the current crop of ECVT motor designs from Avinox and Gobao the pedal and drive shafts are concentric, so it will be quite difficult, if not impossible, to get a concentric pivot to surround the BB to get a concentric wheel path, like with the faux-bar suspension of the Lenz Milk Money (I own one, BTW):

View attachment 187438
Besides, having concentric BB pivot will kinda kill anti squat features. Instead you would need some kind of dual chain high-pivot design, similar (but not the same) to Pivot's downhill bike to get the the kind of anti-squat Enduro folks are after:

View attachment 187436

Keep in mind though the DW6 design used by pivot (seen above) still does not have a concentric wheel-path, so a tensioner would still be required a but with a true single-pivot arc. In this case that chainstay pivot would either need to be co-located with fixed idler or directly in front of it.

Technically, a ECVT motor could be designed to have non-concentric pedal and drive shafts, similar to the original Effigear gearbox design:

View attachment 187439
It would be awesome if some future generation ECVT motor would have a layout like the Effigear: you get a real HP rear suspension where you can finally ditch the lameness of belt/chain tensioner as well as fully open up battery regeneration capabilities.

As an aside, what I said about the above Avinox prototypes not supporting battery regeneration is a bit a of lie. You could always pedal the bike to charge the battery (mechanically allowable with ECVT), but I'm sure that's not what customers have in mind when they hear the term "battery regen".

BTW, Hub-based MGUs don't have this constraint because the regen forces happens inside the hub.

I expect that if battery regeneration is light, reliable, reliable and uncomplicated to maintain it might be the useful in some scenarios. But non of these are likely.

 

[TorAtle]

I expect that if battery regeneration is light, reliable, reliable and uncomplicated to maintain it might be the useful in some scenarios. But non of these are likely.

I guess we would need some new brake levers which relay the lever position to the control unit. Another component that needs to be powered