'lightest.bike' 1.7kg 1000w mid drive

[Pilot Engineer]

I recently decommissioned my recumbent bike and decided to give fitting the drive on a regular bike another try.

Half a day was spent getting the best chainline, tension, etc to no avail.. the best i can do, is with the medium mount, cover 1/4 of a 46T sprocket. 46T is the absolute limit when accounting for the chain not being able to touch the top of the chainring.
View attachment 361765

Even in this scenario, with the power off, pedaling causes the chain tensioner to move forward significantly, the chain skips on it, then the chain skips on the chainring.. regardless of gear. It will skip with maybe 5% of my leg power applied.

I decided fine.. let's go back to the dual chainwheel configuration.
But.. despite my alignment being +/- 1mm.. the wobble in the lightest bike cranks causes the drive chain to constantly fall off and i have no idea why.

View attachment 361766

Next day after cooling off from the long running frustration, i try a poor man's inframe mount, trying to maximize chainring contact..

View attachment 361775

..now i finally have ~40% chainwrap on a 48T.. maybe enough?

In the tallest rear gear, the chain is not slipping along the gears anymore, but as i shift down to lower gears, the chain starts skipping over the chainring again.. on the lowest gear, the bike is unpedalable and just as bad as the above setup.

When i do have the bike in the largest rear gear, and it's not slipping, i find the feeling of sludgy notchiness coming from the idler quite irritating.

Looking at lightest's page about the different mounts, i notice that even the bike with the inframe mount, which should give an ideal amount of chainwrap, has ~30% chainwrap, which isn't enough to prevent this problem.. so i'm betting this is a problem with all installations.

View attachment 361777


Luckily one of the permutations of dual chainring setup actually works for me w/o the motor chain dropping itself.. it's identical to the the setup that worked on my recumbent.

View attachment 361776

Quite a bummer that there's no way for me to avoid this dual chainring setup.. i actually don't see how the single chainring setup could work well..

@neptronix if the idler arm has the plastic sprocket in a too advanced position, then you can move that in the other holes, so that the movement when you pedal will be less. If your chain is skipping, you can install the other chainguides.

 

[neptronix]

Okay, i'll try that.
I have much better crankarms on the way too.

 

[neptronix]

Update from the indiegogo:

42 motors have been shipped! Many of you will soon receive your Lightest ebike kit, and we can’t wait to hear your feedback. If you’ve already received yours, let us know how it’s performing!

New gears are here – and they’re even quieter! We’ve just received the latest gear samples, and after testing them, we’re thrilled with the results: they reduce noise by 12 dB! This marks a significant improvement, making your riding experience even smoother and more enjoyable.

Even more importantly, this reduction happens in the frequency range where the human ear is most sensitive—between 1000 and 2000 Hz—which means a much quieter and more comfortable riding experience.

? What’s next?
We still need to complete full-scale testing over the next two weeks before proceeding with production orders. If everything goes as planned, the motors will start shipping from our facility with the new gears within six weeks.

We’re continuously working to enhance the performance and reliability of the Lightest kit. Thank you for your patience and support—it means the world to us! ??

Stay tuned for more updates, and happy riding! ?‍?

 

[neptronix]

170mm Shun freewheeling crank installed on mine.
Keeping on the freewheel that comes on the Shun crank, the crank wobble is reduced to what i'd expect is normal for a bike crank ( ~1.5mm of total lateral wobble )

I would feel confident about trying the original single chain configuration with these cranks.

Outward offset on the cranks is +1-2mm versus the original 165mm Bikee cranks. Good, that's not hurting the Q factor much.



I have the legspan of a 6" 2' person so the 165mm pedals never felt right, like i couldn't get enough power into the cranks. The 170mm is just allright in that aspect. I really want 175mm.

The Shanmashu big dog pedals i recently bought help counteract the added Q factor. They have the pedal closest to the axle. So if your foot wants to go inward, it can while still having something to leverage against. I notice i ride the inner edge of this pedal.

This can allow you to step -10mm inwards if that's what your body likes doing.


Between these two changes, i should be effectively within a few millimeters of the bike's stock Q factor so that's pretty awesome.

Love the swappability of parts on this thing.

 

[amberwolf]

New gears are here – and they’re even quieter! We’ve just received the latest gear samples, and after testing them, we’re thrilled with the results: they reduce noise by 12 dB! This marks a significant improvement, making your riding experience even smoother and more enjoyable.

Even more importantly, this reduction happens in the frequency range where the human ear is most sensitive—between 1000 and 2000 Hz—which means a much quieter and more comfortable riding experience.

I wonder: is this because they are made more precisely and better mesh? Or is it because they are made of something that deforms under load more in some way to dampen the vibrations generated?

 

[neptronix]

I'm told that the machining is improved.

 

[neptronix]

Bought a 29er and now that i have cranks that are straight, i'm going to attempt a single reduction configuration so that i can get the 'director's cut' experience

 

[Atos62]

Found something nice:

 

[Atos62]

https://www.ratiozero.com/VID_20230623_082049-2.mp4

 

[neptronix]

Bad news on my end.

I installed the rear tensioner and the chain slippage went away but now as a result, i have a loud chain, and i don't like single chainring mode.
I removed the torque sensor/idler and the racket was reduced, but still sounds like a broken bike.

Well whatever, i wanted to see the stock experience so i trudged on..

The power connector to the controller has been chronically loose since the beginning, and i've had to bend the pins to get a good contact again before.

This time i bent them, part of the positive one snapped.

Frustrated with this motor after all the hassle.

So okay, i'm out of warranty so i decided to take a peek inside.

The mosfet cooling solution is pretty smart. We have the controller heat on a big chunk of metal on the the opposite end of the motor, nice.

The motor appears to be sensorless. I cannot find anything on the board that looks like a hall sensor. And no additional wires out the motor.



^-- btw this capacitor is 80v rated

The motor winding has a lot of loose copper doing nothing on top, That's disappointing because it means the motor is less efficient than it could be.



Due to the position of the gears ( right above the motor's phase wires ), there is no easy way to route the phase wires out, other than to drill into this big heatsink portion shown here. to reach where the phase wires connect internally.



I am generally pretty impressed by the build quality and would like to continue using it anyway, i think an external controller can unlock some more power

I think that since this motor is very high RPM, a VESC would probably be the best driver of it instead of a phaserunner.

Let me lick my wounds, find an appropriate urethane sealant, and i'll take a stab at running this externally.

 

[PaulD]

Thanks for the teardown pics. I'm also really surprised by the loose copper floating around in there. Looks like it would rattle around. Maybe it increase the inductance a bit?
Do I count 16 magnets? And do you mind counting the gear teeth?

 

[neptronix]

Sorry but I would feel bad giving a competitor detailed info. I actually hope the company pulls through and improves their design, i don't want to mess that up by spilling a second helping of beans.

 

[amberwolf]

I do wonder why they didn't tie down the wires; that was done in even the crappiest hubmotor I've had. :?

 

[neptronix]

The controller board actually compresses it flat.

I don't know enough about motor engineering to remove some of the extra wire or not tbh, maybe it's only worth -1% efficiency IDK. maybe it is there to equalize winding length or something.

I'm tempted to reterminate the windings to as short as i can get them but i really want to test a totally stock motor so that i have a baseline before and after. I think it's reasonable to have 12 gauge wire coming out the end so that the external controller is adding only a speck of additional resistance in the winding.

 

[uhmwpolyethylene]

Bad news on my end.

I installed the rear tensioner and the chain slippage went away but now as a result, i have a loud chain, and i don't like single chainring mode.
I removed the torque sensor/idler and the racket was reduced, but still sounds like a broken bike.

Well whatever, i wanted to see the stock experience so i trudged on..

The power connector to the controller has been chronically loose since the beginning, and i've had to bend the pins to get a good contact again before.

This time i bent them, part of the positive one snapped.

Frustrated with this motor after all the hassle.

So okay, i'm out of warranty so i decided to take a peek inside.

The mosfet cooling solution is pretty smart. We have the controller heat on a big chunk of metal on the the opposite end of the motor, nice.

The motor appears to be sensorless. I cannot find anything on the board that looks like a hall sensor. And no additional wires out the motor.

View attachment 367033

^-- btw this capacitor is 80v rated

The motor winding has a lot of loose copper doing nothing on top, That's disappointing because it means the motor is less efficient than it could be.

View attachment 367032

Due to the position of the gears ( right above the motor's phase wires ), there is no easy way to route the phase wires out, other than to drill into this big heatsink portion shown here. to reach where the phase wires connect internally.

View attachment 367031

I am generally pretty impressed by the build quality and would like to continue using it anyway, i think an external controller can unlock some more power

I think that since this motor is very high RPM, a VESC would probably be the best driver of it instead of a phaserunner.

Let me lick my wounds, find an appropriate urethane sealant, and i'll take a stab at running this externally.

Disappointing for five years of wait

Are there any other mid drives that use the same type of design as this, I've heard a later iteration of the cyclone does? If so, what serial number version etc?

 

[neptronix]

There is nothing like this on the market, i've been looking for about a decade.
Closest thing is PaulD's mid drive which is 650w rated and 36v only.

Cyclone has kits that are somewhat like this, but they weigh a ton and are not known for reliability, are very outdated, are getting spendy, have no specs listed, and have broken webpages all over the site. how confidence inspiring is that

It's unfortunate that the release of this drive hasn't gone well. I really hope they pull through and improve the drive because the experience on the customer end could be 3x better with some minor changes. The core idea is extremely good - making a drive that can adapt to pretty much any gearing/bottom bracket situation.

 

[neptronix]

I pulled at the wad of extra wire. There appears to be ~4 inches of wire for each phase, whose main purpose is to link from one part of the stator to the other. I'd estimate there is a foot of extra wire in here and that is a significant amount.



On most stators i've seen, the wire between each stator tooth will take a straight shot from one point to the other and not loop around.

Example: MAC motor



After some looking about.. I think the best way to get the phases power out is to drill into the case straight at this point where the 2 wires from each phase meet & get soldered to the connector.

I was hoping that this black plastic-ish holder of the phase terminals could be pulled out a bit, but apparently it's glued on somehow and i don't want to screw with it, so it stays.



Okay, now to build up the gumption to drill into this..

 

[zizajaun]

I haven't read all 41 pages of this thread but the tone seems pretty critical about the Lightest. Can someone please give me a brief summary of how good and reliable this motor is? On paper it seems great (light, powerful, fits lots of frames), but I can't find any independent reviews. It seems to be on sale, which makes it cheaper than the CYC Photon.

 

[glennb]

I haven't read all 41 pages of this thread but the tone seems pretty critical about the Lightest. Can someone please give me a brief summary of how good and reliable this motor is? On paper it seems great (light, powerful, fits lots of frames), but I can't find any independent reviews. It seems to be on sale, which makes it cheaper than the CYC Photon.

Well put. I'm in the same boat, not willing to sink with the guinea pigs, but keen to jump on board.

 

[neptronix]

I haven't read all 41 pages of this thread but the tone seems pretty critical about the Lightest. Can someone please give me a brief summary of how good and reliable this motor is? On paper it seems great (light, powerful, fits lots of frames), but I can't find any independent reviews. It seems to be on sale, which makes it cheaper than the CYC Photon.

I haven't heard of any significant mileage being put on these, i have maybe 300 miles of mine.

Bullet points version of that:

The good:

1. Weight is as advertised; exceptional power density.
2. Excellent thermals, you can't overheat it - with modification there's room to increase the power ( not sure if the drivetrain is the limit though )
3. The Q factor is exceptionally good, if you have issues with most mid drives that make you pedal bow legged/lopsided, this drive will work for you.
4. Adaptability to different bicycles is excellent thanks to the outer drive mechanisms just using bike parts. This will fit bikes the BBSxx etc won't fit.
5. Build quality of the drive unit is good, with only 1 possible wart known ( what looks like excessive non-useful copper in the stator )

The bad:

1. I had to pay ~$250 extra to get an upgrade to unlock the speed limiter and add a throttle. I didn't know this up front. I bought an offroad unit.
2. The stock power connector isn't keyed and is physically weak.
3. My 1000w unit is hard limited at 900w, and the marketing materials say this unit peaks at 1200w.. all attempts to tune the power don't do anything, and i failed to get help with getting the promised power 3 times.
4. In single chainring mode, expect a loud chainline and much stuggle to get to work; this is the most difficult route and the dual chainring mode is 10x easier to install. But keep in mind that you lose torque sensing in dual chainring mode.
5. Stock cranks are not straight and in single chainring mode, you will hear the chain making tortured sounds or possibly skipping due to it - with these bad quality cranks, i've only been able to get dual chainring mode working. I ended up replacing the stock ones with Shun freewheeling cranks. You may also not like the 165mm size of the stock cranks.
6. Some have complained about how the torque sensor/tensioner works as stock - it produces a rhythmic motor droning and pressure and this is annoying. Not a problem if using throttle like me.
7. The app has been incredibly buggy for me on a new android phone... it randomly crashes all the time.. the drive unit can also shut itself off after a few minutes of sitting idle, regardless of settings. I don't know if these problems are fixed or not yet.

Known unknowns:

1. There's been a lot of complaints about loudness of the motor itself. The company recently made new gears and supposedly fix this, but we don't have any reports on how that fared yet. Personally, i think the noise level is average.
2. Improvements to the drive seem to be happening at a crawl. The company doesn't express interest in fixing a number of the existing problems, so i don't know if they will be addressed.

I would only recommend buying this if you have a high pain tolerance at this time.

 

[macil]

I ordered my kit on January 11, 2025, and it arrived within ten days. They just didn't send me the complete order, some of the components were sent to me additionally.
I ordered directly from the website Lightest ebike kit shop. Your mid drive motor with torque sensor.

 

[neptronix]

Thanks for that, that's really good news, updated my post.
Let us know how it goes with your kit.

 

[glennb]

Thanks for the summary.

I want something for my road bike purely for climbing assistance.

Criteria are:
- small
- light
- neat
- quiet
- efficient
- preferably configurable power and throttle response
- preferably open source software
- prefrably not have to convert to 1X

36v 500w peak would be plenty.

"Lightest" serms closest fit, particularly if VESC can be hacked on to their controller.

 

[neptronix]

Quiet is unknown now that they have new gears, efficient? i'd say so; above average.
Stock controller has 5 different power levels, that's the only way to tune it that i know works.
Dual chainring will be quiet, single chainring will be a racket during powered and unpowered use due to chain noise and from a roadbiker perspective, you'll probably hate that.
Very small amount of drag on the dual chainwheel version, you might lose a mph when not powered.
No open source software. The controller looks fully custom.
The narrowest square taper bottom bracket is 123mm wide and fits into either a 73mm or 68mm BB. Not sure about the other BBs.

The only other alternative is the paulD mid drive for your needs. It also has low Q factor.

Compatible with VESC? on my end i have a temporary motor and now knowledge that i can get another unit in a reasonable amount of time if i screw it up, so my gumption level is increasing and i'm ready to FAFO!

 

[maiz]

Thanks for the summary.

I want something for my road bike purely for climbing assistance.

Criteria are:
- small
- light
- neat
- quiet
- efficient
- preferably configurable power and throttle response
- preferably open source software
- prefrably not have to convert to 1X

36v 500w peak would be plenty.

"Lightest" serms closest fit, particularly if VESC can be hacked on to their controller.

I would add the bikeon to your list. Keep your bb, keep your rear wheel. The only sacrifice is your two largest rear cogs.