Your take is correct. Freewheeling is a critical component of retro-direct transmissions, and so regen and reverse will never be in the picture.

We spared you the 5 minutes of struggling before they finally did the trick!

😂😂😂

😂😂😂🎉

@@GrinTechnologies KNIPEX Tools - 6 Piece Circlip Pliers are amazing

Yeah if you have a diode you won't be able to do regen, but you really don't need a diode with parallel connected batteries. Just charge them independently and plug them together for discharging and they're is very little risk. The temp sensor upgrade depends on the controller, but on those that can take a multiplexed speed and temp signal is pretty straightforward.

Speaking from seeing 100's of thousands of km of logged ebike ride data, I can tell you that the only situation with an ebike where a freehweeling motor will get you more range than a regen capable motor is when your riding style involves a majority (>> 50%) of the distance pedaling without using the motor and don't have any significant stop and go or long downhill stretches. In that case the extra drag from the wheel has more of a negative effect than the gains from regen. Even in the case of coasting fast down a hill and building up momentum, vs going slower and collecting regen energy, you will generally go much farther powering on the flats from the saved regen energy than you would from the extra inertia of the higher downhill velocity achieved through coasting. Storing energy via inertia is 100% efficient, but the quadratic rise of air drag with speed very quickly negates any benefits.

@@GrinTechnologies Thank you for the detailed reply, very much appreciate it. The work you guys do on bringing electric vehicles forward as practical transport is admirable and your enthusiasm for the technology behind them is genuinely infectious. Wish you and Grin every success.

Correct. The GMAC uses a larger axle diameter on and a thicker tight fitting key stock and hasn't shown any issues in this regard. For the G60/G62 motors, the verdict is still out on how hard you can push it with just the stock key. You can definitely feel the little bit of play and 'clunk' sound when the core switches modes, and it slowly gets worse over time. But so far none of the units on the field have gotten so bad that it either trips up the controller or causes mechanical failure. Everything feels a lot nicer with zero play though so we'd recommend using the 648 retaining compound regardless. For the RH212 of course there is no need as the stator is quite literally welded to the axle.

GMAC regen is huge !

That's neat to hear, what is the brand of the bike? It's pretty rare to see this on a factory ebike.

@@GrinTechnologies It’s one of those old Currie folders with small diameter wheels and 24V of lead-acid batteries. They are suitable only for low-speed operation, and the regen makes speed modulation easy enough that brakes aren’t needed most of the time. Btw, I’m an OG from endless-sphere forum. You guys at GRIN are the GOATS!

We've seen the welds fail when done poorly, so the glue would have to be pretty tough. Plus it's quite hard to clean out all the grease to get a glueable surface. No reason you can't try though if you think you have a viable approach

Hey Knoxie and thanks for sharing your field experience. FWIW you almost always 'gain back' more than you loose having regen vs freewheeling, see kzfaq.info/get/bejne/nb9znZeS2NadoHk.html bewteen the 10 to 15 minute mark. The torque arm / nuts loosening situation is for sure the most annoying downside with modding motors for regen. It's a consequence of all generic hubs using lame axle flats rather than a spline for the TA interface, but this here does the trick to solve this with a single arm: ebikes.ca/v7-regen-torque-arm-12mm.html We all think you should have another crack at this in spite of your experiences from years ago :-)

The easiest way to understand why this is not an issue or concern at all is to realize that the forces present on the frame from regen braking are identical to those on the frame or fork from disk brakes brakes. Only disk brakes can put upwards off 200-300 Nm of torque, while regen is rarely able to hit more than~80 Nm. If bikes aren't falling apart from fatigue with disk brakes, they definitely won't from regen either

@@GrinTechnologies I've heard about some vintage bikes that shouldn't be upgraded to disc brakes and figured it would be the same for hub motors, but I didn't realize the motors are actually gentler than the brakes. Thanks for clearing that up!

A 4mm bolt will have nowhere near the strength to resist the rotating forces on a axle. 80 Nm of torque at the surface of an 12mm axle is about 3000 pounds of shear force.

Ok,that is more then I expected, thank you for the Info.😊

Thank you! 😃

The motor heat is not high at all by grease standards (like 100-110 oC max) and electrically conductive graphite dust floating around the motor can very much mess up hall signaling, the same way water in the motor can cause leakage currents. There is almost no wear on the insides of the motors, and with the nylon gears we honestly are not convinced grease is even needed at all.

It is indeed possible. MAC motors have an extra complexity that the motor stator is also a bit loose on the axle, and not just the planet carrier. So you may need to loctite 648 on the actual stator too as that will wear down from back and forth torques in a normal MAC. We had to have the stator core material in the GMAC's modified with a steel sleeve insert as the stock stator core that was all aluminum would eventually get quite loose.

We've discussed this topic in various presentations and videos about regen. In general, it's not an issue at all. You'll always be drawing much more than putting back in. In some special situations (like you live at the top of a mountain and start every trip with a really long downhill), then you are aware of the situation and simply charge your battery to like 85 or 90% so that it has a reserve for regen out of the gate. If you have a programmable controller, then you set the max regen voltage to not exceed the full charge battery voltage and the regen will automatically roll back as required. Single port BMS circuits will cut off regen, dual port BMS circuits won't. Neither will get fried from regen unless they are terribly designed BMS boards.

With access to CNC You can definitely custom machine your own planet carrier and make it much lighter since it doesn't have a clutch mechanism. A bit of axial play in the planet gears is of no consequence. The noise you hear is probably the motor simply winding down from it's own inertia after you let go of the throttle.

@@GrinTechnologies do you happen to know what kind of steel they are using for the planet carrier?

Any short between the phase wires will cause very high drag to turn the wheel. This is common if the cable gets severed or torn as the exposed phase wires can come in contact

Yes indeed, those two things go hand in hand. In fact, the only reason that ezee had a stock carrier without a clutch is because they had a high volume customer that required reverse mode. Regen wasn't part of the picture.

It's a good question actually. We were worried that the heat would damage the loctite but it's usually possible especialy with steel to keep the heat quite localized just to the weld area. Worth an experiment for sure.

@@GrinTechnologies I couldn't get it off the shaft anyways and didn't want to try a puller, so I'll just wait a bit and see if it loosens up on its own and then loctite it.

This isn't just a hypothetically good idea on paper. We have firsthand data from over 20 years of actively using regen on ebikes in all kinds of conditions to back up the merits! The actual numbers show that in the majority of riding situations you get more energy back from regen than you expend from lack of freewheeling. Usually it's a factor or 2-4 times or higher. City riding in Vancouver on my normal bike I average between 1 to 1.5 Wh/km of energy repacture from regen. The energy lost to motor drag when pedaling without the motor running is no more than 0.3-0.5 Wh/km. But even that's mostly beside the point because it's 100% worth it just to have the benefits of electronic braking with no mechanical wear of the brake pads. The primary scenario where there is a net benefit to freewheeling over regen is people who are doing long distance riding in flat terrain without any stop and go and where the motor is used sparingly, like bike touring in the countryside and only saving the motor for occasional headwinds etc. For more details, watch: kzfaq.info/get/bejne/nb9znZeS2NadoHk.html or read: endless-sphere.com/sphere/threads/analysis-of-regen-on-an-ebike.7891/ A controlled rather than overrunning clutch of course gets you both (freewheeling and regen), but it comes with a lot of complexity if you work on all the mechanical and electrical details of implementing one. That is likely the primary reason why no actively clutched hub motors have every made it to market yet, though there is nothing stopping someone from building one!

It's best to start your ride with a warm battery by keeping it indoors till you head out on the ride. You get better performance on discharge (lower internal resistance and much less voltage sag) and no problem handling high regen currents either.

You could try, but be mindful of the torque levels at play and the shear strength of your screw.

Direct drive is far easier and blue reliable. But if lightweight is most paramount for you, then geared is the way to go. Shengyi SX1 is a really good option for that

We went with a wider / stronger keyway on the GMAC planet carrier than the regular MAC motors, so it wouldn't be quite as straightforward.

The prelocked ones are from eZee bicycle. For Shengyi they don't have that as an option, in spite of us doing many requests. Welding yours shut isn't a bad way to turn a nonworking motor into a working motor at least.

@@GrinTechnologies as an alternative do you know if u can swap other clutches? Is it just a matter of moving over the 30t planetary gears if the OD and ID of the clutch match? Or is there more to it than that?

There's nothing off the shelf that will have the same planet carrier pins at the same location. It's not a generic part, it's something specifically engineered by Shengyi.

It does, but luckily it also has a pair of deep groove radial ball bearings in each planet gear that can cope with the thrust no problem. This is what makes it quite one of a kind. No other heroism gear hub motor we've seen on the market would be regen compatible this way

No and no, as a motor it works and behaves exactly the same. There is no change to the gear ratio as its the same set of gears.

@@GrinTechnologies Oh yes, now I get it. The ratio would change if the gears would not spin. It only affects the freewheeling clutch.

You totally could, but then as you say it means future servicing of the motor gears is impossible, and you would also distort the axle with weld bead right near where the ball bearing of the shell needs to fit. It would have to be a tight job or one that is ground down after the fact. Brazing would likley be better.

Yup. Several times and it never came close to doing the trick, even when we tried to completely fill all the gap space in the keystock. Epoxy does not bond to metal very well, nor does it have much strength as a general gap filler to take up play.

Would loctite 684 maybe bond strongly enough to the metal for this?

@@3737HHOh I see, you are talking about using a glue/adhesives inside the clutch assembly to prevent it from freewheeling, rather than using it as a bonding agent between the planet carrier and the axle to prevent play. No, we have not tried this. The clutch is full of grease and rivetted together, so you can't really get at it to clean and flush out the inside of all grease so that a glue would stick. That said, we HAVE seen some clutches on old motors with water damage where the rust itself caused the clutch to become seized. So perhaps a strategic soaking in salt water could do for those without access to welding supplies. We have also tried drilling holes and pinning the clutch mechanically, but without being able to see the inner workings it's hard to do this with proper effect.

Potentially. You can tell if the freewheel is working or not by unplugging the motor controller and then spin the wheel by hand forwards and backwards. If the drag feels the same both ways then you don't have a freewheel, while if there is more drag in reverse then it means your freewheel is working.

Hey Bruce, it's true that there is more drag when pedaling without a battery (similar to the rolling resistance of your tires if you want to quantify it). But you are dead wrong about regen gains being less the coasting benefits. See kzfaq.info/get/bejne/nb9znZeS2NadoHk.html. We've been selling non-freewheeling geared motors with the GMAC system for nearly 5 years no, with no instances of people breaking gears from being airborne or doing small jumps. I could see the sudden change in RPM causing a shock load on the gears due to the motor's inertia beyond what you'd normally get from motor torque, but I wouldn't universally say this will break gears as a blanket statement.

Yes, pretty effectively.

@@GrinTechnologies could you please test this out and show how effective it is please, thx for when we are locking the bike to protect from thieves

Uh, you definitely need a motor controller that can do regenerative braking via some input or another, you need a torque arm that can handle bidirectional torque, and you need to ensure that your planter carrier is affixed to the shaft such that the back and forth rocking won't shear the keystock.

@@GrinTechnologies alright thanks

@@GrinTechnologies can i send u somewhere a DM?

The controller almost for sure will be fine with the clutch disabled.

There's been more than a few that have come and gone over the years, but these aren't usually available as separate parts, more typically it's part of a proprietary system.

That is exactly the same as the G062 shown in this video. So yes. But you will most likely need to change the controller on your ebike to take advantage of the regen

@@GrinTechnologies If the controller is left unchanged then you still get the braking, but not power back into the battery, correct?

@@kitesailor No, no regen, no braking.

In principle you can do this with most geared hub motors, but it's risky and complicated and not something we would recommend for 99% of people.

Not in any meaningful way, no. The pedal drive chain is unaffected, but there is now a bit of motor drag (0.5-1Nm) to overcome while pedalling without assistance.

Thank you for replying

A geared hub motor of the same power class generally will, yes, since they have lower winding resistance for a given KV, so less of the regen energy is absorbed by the copper winding and more is available to go back into the battery pack.

Yes, almost by definition a motor that can run in reverse can also regenerate, as it doesn't have a clutch and can apply backwards torque to the wheel.

@@GrinTechnologies cool just need to figure how to make it charge the battery

In theory yes but in practice not very easily and not with normal bicycle drivechain hardware.

No the pedaling system is totally unaffected by this. However, locking the clutch will result in a small amount of resistance to turn the motor. Depending on the model of hub it can be anywhere from 0.5 to 1.5 Newton-meters. It's roughly comparable to the rolling drag of the tires on the ground.

Zero indications or signs of that so far, and there isn't any reason why the nylon gears would care which direction the torque is in. So far all the 'issues' stem from metal components that slip or wiggle a bit when alternating torque direction, and that includes the planet carrier/key stock and axle itself when secured with just a normal torque arm. Both of those are addressed in this video.

@@GrinTechnologies ohhh that's good to know, I'm opening a repair shop of my own after working as a repair tech for this mom and pop shop for the last 3 years, and recently there's been a spade of ig hubs coming back with planetary gear stripping, i assume it's cuz of the lack of grease cuz wen i service them is bone dry inside

@@edgefx1 We've not seen any correlation between greasing and gear damage, nylon is largely self lubricating after all. We have seen batch-to-batch variations when clearly one batch of gears had poor process control in the plastic injection molding or whatever and was more prone to failure when previous lots of the same hub held up just fine. And of course issues with people putting more torque in the motor than it can sustain. If the motor is run quite hot, the plastic gets correspondingly weaker, and most gear stripping happens during a long climb or other heavy usage situations where the motor is toasty.

@@GrinTechnologies hopefully that's the issue, a batch of bad gears causing a bunch of costumers to come in for such repairs🫣, if it's just a bad batch then id love to be able to offer this service without the customers coming back saying their gears stripped

Either your gears are stripped or your freewheel clutch is not engaging and is slipping instead. Open it up to have a look and see which it is!

@@GrinTechnologies I opened but i cant find solution, everything looks good, freewheel rolls and strings..

You just do all the steps shown here in reverse.

There's one on my desk for just this purpose! Of course, locking the clutch on the BBSHD turns it into a pedal generator not into a regen brake that can slow down the bicycle. But you can eliminate the drivechain completely and just pedal to charge the battery if you want.

What if we lock the freewheel on the rear cassette? Well I guess this will only work on a fixie bike, without a rear derailleur... That S curve on the chain would snap... Has anyone used a BBSHD on a fixie?

That is true, but the grease inside the freewheel no longer serves any function because it's no longer a rotating component that can benefit from lube. So damaged or not damaged doesn't really matter.

@@GrinTechnologies might reduce warping also not sure though, I've used that technique to cool quite a bit, its effective

Nope. It's totally a DIY project, not something that we or anyone else we are aware of offers as a paid service. So if it's outside of your comfort zone to do this kind of work yourself you would be better off just purchasing a ready made factory ebike that already had regenerative braking well implemented. An used but good condition BionX system is a good bet in that regards.

It depends on the speed. Most DD hubs have 0.5-0.8 Nm of drag loss, so you can multiply that by your RPM ( but in rad/sec) to get a watts value at any given wheel speed.

@@GrinTechnologies that's not too bad. I was expecting it to be like double that. 🙂

@@nebulous962 Well some hubs are about double that, like the big honking Crysatlyte motors were like 1.5 Nm. One of the lowest we measured was the original BionX PL350's, at about 0.4 Nm.

Agreed, if we had a mig we'd have used it for sure.

Not dumb, but so far any implementation idea that gets hashed out seems to add enough complexity to be "not worth it". Especially when the drag losses are pretty insignificant -much less than the energy you get back into the system from regen-. The more sensible approach to reducing the drag losses to almost zero would be switching from permananent magnet motors a technology without PM's, like induction or switched reluctance, or go ironless like maxon motors. Without magnets moving over iron, the unpowered motor can spin freely without resistance, yet still stay always engaged for regen.

@@GrinTechnologies Oh I didnt mention the dumb stuff I wanted to do, I was talking more pushing over 120v through it again now that I have amassed quite a few 60v batteries. Although, you make a great point, I thought it would add too much complexity, but then I started thinking about the Rholoff Hubs and the way they work and thought eh why not, one day we might be able to solve that riddle more simply. If you ever see any hubs running induction motors I would be very happy to see them, my Escooters would be much less of a pain to walk home when I get flats then. Thanks for the response :)

@@GrinTechnologies Btw, I very much do need a replacement for my BPM clutch assembly and have been having one hell of a time trying to find one.

Lol I just finished writing my comment and I had the same exact question!

​@@GrinTechnologies Interesting, I had the same question! That answers it. So it's hard to imagine a simple mechanical solution to offer this actionnable mechanism. The benefit would be greater for heavier direct drive motors but their stators are solidly fixed to the axle, so it sounds unplausible to find an elegant solution to disengage heavy PM motors mechanically. I've personally experienced running out of battery it on my commute, and would have benefited from this solution in that case :)

kzfaq.info/get/bejne/nb9znZeS2NadoHk.html At the 12 minute mark

@GrinTechnologies Hi Grin, I really love the work you do, I wish I could learn from you; I don't know much about technical stuff but I would at your place for free just to learn.

kzfaq.info/get/bejne/nb9znZeS2NadoHk.html Obviously it depends, but here in Vancouver we typically see between 5% to 15% depending on the user, terrain, and riding style.

On principle yes but it is both nearly impossible to source replacement axles of a given motor type, and the actual swap is a royal pita since everything (wiring, soldering, interference press fitting etc) needs to be redone

The motor drives the sun in this topology. The pedals aren't connected to the gearing at all, they drive the hub shell directly.

No, he´s not. Just a bloke pressing his agenda.

3-5% is still more energy you got back than you ever expended overcoming the small amount of motor drag when pedaling without the motor. Not sure what you mean by throttle control, you can coast just like any normal bike by not using the throttle, and if the ~0.5-1Nm of torque is an issue then you can use the "virtual electronic freewheeling" feature of the Phaserunner/baserunner controller to automatically cancel that with a small amount of current injection from the battery pack. As we've been showing since 2008 or so, that is more than covered by what you get from the "small" amount of regen. Remind us what is the downside that makes it "not worth it" ? Definitely with higher weight to drag ratio vehicles you get higher and higher regen percentages. Cargo ebikes are routinely in the 15% + regen realm even in just stop and go riding without major hills.

@@GrinTechnologies I dunno what ebikes with regen are like, but with my Tesla once you stop accelerating you let back on the "gas" peddle just enough to coast... but if you let back too much - like all the way - the regen kicks in and it starts to slow you down. This is how I assume regen bikes work. If not, then maybe you coast as with normal vehicles when not accelerating but since regen recovers energy by slowing you down the only other way to do it I can see is by activating the regen system (thus slowing you down) through the brakes or some other notification system. All in all I'd rather be able to let off the gas and coast and then begin regen once I touch the brakes and actually engage the brakes once I push far/hard enough. Now, if this is the way regen ebike brakes work and they're not heavy and not expensive and have no draw-backs then sign me up. Also, can regen braking be set-up on my cheap-o 1000w or 3000w Chinese ebike conversions? If so, how are they implemented, how much do they weigh and how much do they cost? You mentioned a controller --- is that all I need? Thanks

That too. Sometimes the clutches even fail in a way that they are locked, and then you get the mod for free!

>> you cannot add any meaningfull charge to a battery on something as low mass as an ebike I guess that depends on your definition of meaningful. The % recapture rate on a bike is not as high as a on heavy dense vehicle like a car, but it is still north of 10% or more in most situations with an ebike in either stop and go riding or riding with lots of hills. It's better of course with loaded cargo bikes.

You don't actually need to generate a higher voltage than the battery if you are using a well designed bldc controller. Also, not having as much mass is kind of a moot point because the energy used for accelerating and decelerating are proportional to mass

@@noobulon4334If you charge your battery with your mains powered battery charger for 4 minutes ...how much have you charged it ?.... 4 minutes of regen cannot give you anywhere near a constant 2 amps like your battery charger ... Mass inertia is the only stored energy you have . so it only makes sense on locomotives .. On a low mass vehicle Why not just call them magnetic breaks. Calling them "regen" is just a marketing slogan , As I said before ,who is going "fast" downhill with their breaks on ?? you are just wasting efficient momentum and turning into inefficiently milli amps ...that is actually wasting energy ! I will connect my solar charge monitor to my battery output and go down the longest hill I know with my "regen breaks "on and give you the current ,speed,,duration data ...when it stops raining.

@@marxman00 you're barking up the wrong tree here and for two reasons. First is that we *have* batteries that can charge in excess of 20C, (so 0-80% in less than 3min) the particular chemistry is Lithium Titanate Oxide (LTO), and as long as you have a big enough battery it can take in whatever current you need it to Secondly the magnitude of energy in the system is irrelevant, what matters is how efficiently you can take in energy and release it. Lots of bikers don't like to stop or slow down even when they really should for their own safety because it is a pain to have to build speed again. If you can efficiently put that energy into the battery and use it to get back up to speed then you basically eliminate the physical penalty of stopping or slowing without killing your range

@@noobulon4334 Dream on, What slice of the total energy pie are you using ?accelerating and decelerating is not a journey ..if what you said worked every electric vehicle would already never need to charge ..because the start inertia would magically last forever .we dont live in a frictionless vacuum , perhaps you are suggesting dream power,..without all the nasty physics getting in the way.