'lightest.bike' 1.7kg 1000w mid drive

[neptronix]

This sounds about right compared to non AI sources i've read in the past.
I notably demagnetized a mac a long time ago at probably >140C

Neodymium Temperature Ratings



Don't expect the manufacturer to tell you what grade of magnet is in the motor so you have to assume the worst.

95C has always been my safeword; i like lower temperatures because the copper becomes less efficient as the heat goes up, so you can get an upward spiral of heat happening when you run a hot tune like i do, if you don't stop early

 

[raylo32]

Interesting... what characteristics or additives determine magnet grade?

 

[amberwolf]

what characteristics or additives determine magnet grade?

There's a bunch of different types of magnetic alloys, and different amounts of each part of hte alloy change it's magnetic characteristics, including the curie temperature, etc.

I don't know offhand what any of the specific alloys are, but you can go to various magnet-company sites and there are good articles about this sort of thing if you really want to know.

 

[Pilot Engineer]

This sounds about right compared to non AI sources i've read in the past.
I notably demagnetized a mac a long time ago at probably >140C

Neodymium Temperature Ratings

View attachment 374510

Don't expect the manufacturer to tell you what grade of magnet is in the motor so you have to assume the worst.

95C has always been my safeword; i like lower temperatures because the copper becomes less efficient as the heat goes up, so you can get an upward spiral of heat happening when you run a hot tune like i do, if you don't stop early

View attachment 374509

We use H grade magnet!
Also, we've been using a motor on a bike with 40A phase current for 4 years now. No problem, it just starts to heat after 10-15 minutes of very harsh riding.

 

[neptronix]

After buying every tooth count from 36t to 42t, i found out that with a 37T, if you take advantage of the 2mm of play in the medium mount's bottom bracket connection, you can get a perfect amount of tension. At this amount of tension, the motor's gear can be removed with the chain on it,

This should provide slightly too high of cadence on full charge, but i actually want that combined with a RPM limit in the VESC, so that even at low rpm, the candence is still good.



Nice that we can remove the tiny bit of mechanical latency, added noise, power required to run this idler. It also provides a conveinent inlet/outlet into the motor if i chose to run copper tubes through it.

Pendulum pedals should be here in a few days. These should negate the seat height gain from the 170mm pedals ( i usually like 175-180mm ) and allow more foot to road contact when seated and stopped.



I think short periods of 37mph is possible. I'm interested in finding out how long it can sustain 35mph.

More later.

 

[limiky]

Oooooh that pedals! Tell us how they feel/perform!

 

[neptronix]

Will do.
BTW have some $ to buy your unit next month.
I need a sacrificial lamb for the cooling experiments and maybe some of your badass parts.

1800w peak or bust!

 

[zizajaun]

On a 500W motor: Pedalling up a short but loose and fairly steep hill there was a metallic snapping sound, a bit like a chain or spoke breaking (but not exactly the same). I checked the bike and the outside of the motor but I couldn't see anything missing or broken, and I couldn't find anything metal on the ground (though it was a rocky area so it might have got lost). But the motor is no longer supplying power to the "motor shaft" as the manual calls it (the one with the external sprocket on that drives the chain). Everything else is the same: when pedalling (or manually pressing the torque sensor), the motor spins and sounds like it normally would; it just isn't driving the chain. There are no error messages on the screen. Turning it off and on makes no difference. Any ideas?

I took the side off and was expecting to see something obviously broken, but not much is visible. The big cog meshes with the small one, but the small one is very stiff so I can't turn it. I'm not quite sure what the function of the big cog is, as it seems to just spin around on its bearings when the small cog turns; it doesn't seem to be driving anything, and it doesn't turn when I manually turn the motor shaft. When I remove the big cog and turn the motor shaft manually, the cogs inside also turn, and I can't see anything broken. However, I can't see much because the black layer is stuck down; it seems to contain electronics, so I'm a bit wary about removing it.

Please advise.

 

[neptronix]

Do you have any ideas on how many miles you have on it?
What options did you select when you bought it?

I'm thinking that you broke the clutch inside the big plastic gear. @limiky what do you think..

 

[zizajaun]

Do you have any ideas on how many miles you have on it?
What options did you select when you bought it?

I'm thinking that you broke the clutch inside the big plastic gear. @limiky what do you think..

Are you referring to the big plastic gear in my picture? I could only see needle bearings inside (see attached pic) but I didn't look closely. Does that contain the clutch? I can remove it again for a closer look tomorrow if that would help.

I got it off ebay from someone who bought it new but never used it. It has less than 500 km on it. It is the 500 W 90 Nm version with no extras (e.g. no Super Hero app so I can't see what the settings are). I was using it with a 52v battery at the time, which is within spec. No throttle installed, only torque sensing. No brake or gear sensors. Medium mount.

The climb was short (about 1 minute) but very loose, causing the wheel to intermittently spin then get traction, which I guess is hard on some parts of the transmission. (I don't really know how ebike motors work, but I have recently replaced the clutch and gearbox on my diesel van.) I was in level 5 at the time, so around 500 W, but in my lowest gear (32t/50t). I'm about 100kg kitted up and the bike will have been about 26kg.

Earlier, I had noticed the temp was around 65 C after that climb. A few weeks ago the motor shut down completely during a ride, seemingly due to overheating, but that was after about 10 minutes at the full 500 W on a very hot day. I doubt that heat was a major factor here.

 

[neptronix]

Sounds like you lugged the heck out of the motor. I was unable to intentionally overheat my 1000w unit by riding it in one gear smaller than what's ideal ( by keeping high pedaling cadence )

Congratulations on overheating a stock model. Possibly these early 250-500w units had smaller stators and aren't as tough.

I believe the clutch is inside the gear and the needle bearings move it. @limiky knows a lot about these

 

[zizajaun]

Sounds like you lugged the heck out of the motor. I was unable to intentionally overheat my 1000w unit by riding it in one gear smaller than what's ideal ( by keeping high pedaling cadence )

So using a higher cadence/lower gear is harder on the motor? I assumed it was the opposite: I was intentionally using a lower gear so as not to strain the motor and burn through the battery.

Congratulations on overheating a stock model. Possibly these early 250-500w units had smaller stators and aren't as tough.

When I asked about upgrading to 1000w, Victor said it was just a firmware upgrade (for about 200 euros), so presumably the hardware is the same. I think the hardware is different on the 250w but I'm not sure; I know you can't use 36v batteries with the 500w+ whereas you can with the 250.

I believe the clutch is inside the gear and the needle bearings move it. @limiky knows a lot about these

I look forward to hearing from him/her/them.

 

[limiky]

Definetly needle roller bearing went kaput. It is a difficult failure to visually inspect, because the tolerances of the rollers on the hardened steel shaft are so little that if the roller change position because of cage break, you cannot clearly see them.
One thing also could be that the 2026 needle roller one way bearing before mentioned detached from the big plastic helicoidal gear (i believe it is glued to it and press fitted too).
If the motor behaves like normal, and pushing by hand the torque sensor makes it spin normally but without driving the chain, it is a slippage of that 2026 needle roller bearing, like in my unit happened 5 times, 99percent probability.
In some days i'm going to open my unit and post some photos of the bikee bike solution to it that uses a much stronger proper freewheel bearing capable of far superior loads.
I didn't use the motor after another breakage with the new freewheel installed, this time it was the inner 20mm shaft supporter bearing that broke, and by discourage at the sixth breakage of the unit i completely changed plans on the bike.
I'll give some feedback on the new freewheel assembly.
Hope that helped.
Oh, and bikee bike told me failures like mine are not common, they tolde only me and one other italian guy had them, but on hard downhill use and with the most powerful units.
Contact bikee bike, amd inform them that it is not your case: you don't have the high torque unit in 1000w config.
They repaired free my unit every time it broke for free (only paying shipment to them).

 

[zizajaun]

Definetly needle roller bearing went kaput. It is a difficult failure to visually inspect, because the tolerances of the rollers on the hardened steel shaft are so little that if the roller change position because of cage break, you cannot clearly see them.
One thing also could be that the 2026 needle roller one way bearing before mentioned detached from the big plastic helicoidal gear (i believe it is glued to it and press fitted too).
If the motor behaves like normal, and pushing by hand the torque sensor makes it spin normally but without driving the chain, it is a slippage of that 2026 needle roller bearing, like in my unit happened 5 times, 99percent probability.
In some days i'm going to open my unit and post some photos of the bikee bike solution to it that uses a much stronger proper freewheel bearing capable of far superior loads.
I didn't use the motor after another breakage with the new freewheel installed, this time it was the inner 20mm shaft supporter bearing that broke, and by discourage at the sixth breakage of the unit i completely changed plans on the bike.
I'll give some feedback on the new freewheel assembly.
Hope that helped.
Oh, and bikee bike told me failures like mine are not common, they tolde only me and one other italian guy had them, but on hard downhill use and with the most powerful units.
Contact bikee bike, amd inform them that it is not your case: you don't have the high torque unit in 1000w config.
They repaired free my unit every time it broke for free (only paying shipment to them).

Thanks for the reply. Yes, it spins freely like it's disengaged when I pedal or press the torque sensor. It sounds exactly as it should (normall spinning sound); it just doesn't give any power to the driveshaft. It's like revving a car with your foot on the clutch pedal, so it makes sense that it's a clutch issue.

I emailed Bikee about this on Wednesday but they didn't reply. I might just order a HFL2026 bearing and try to replace it myself. Is there a brand I should be looking for? Prices seem to vary widely so presumably quality differs too. I'm in the UK. Also, do you have any suggestions for how to remove the current one? Is it epoxied in or just press fit or what?

 

[limiky]

I don't know how it is fixed there. You should ask bikee bike for a hint. Shaeffler are good, they probably offer the HFL2026 too.
To remove it and reinstall it a simple manual headset press style instrument could work, bit i'm not sure they used some glueing agent there (a pressfit only seems not good there, as the bearing can grip the shaft on its hardened surface weel, but the plastic gear could slip).

 

[zizajaun]

I don't know how it is fixed there. You should ask bikee bike for a hint. Shaeffler are good, they probably offer the HFL2026 too.
To remove it and reinstall it a simple manual headset press style instrument could work, bit i'm not sure they used some glueing agent there (a pressfit only seems not good there, as the bearing can grip the shaft on its hardened surface weel, but the plastic gear could slip).

Are the long slots in the middle supposed to contain needle bearings? They feel empty. The narrow bearings at the sides turn freely in both directions, but the middle section doesn't turn in either direction; it is fixed to the plastic gear. So I guess (a) there were never any bearings there; (b) they have fallen out; or (c) they have recessed into the gear (maybe when it broke a space was created for the bearings?).

No bearings fell out when I opened the case, and nothing rattles or falls out when I shake it.

How hard/risky is it to remove the black layer so I can see inside the motor properly?

 

[neptronix]

How hard/risky is it to remove the black layer so I can see inside the motor properly?

It's not that bad. I got mine apart and back together and running. You will mess up some thermal potting between the controller's FETs and the case in the process, but the motor is thermally overbuilt, so it can tolerate losing some.

 

[zizajaun]

@limiky @neptronix
I just removed the old clutch by tapping it out with an 18mm socket. There was no adhesive - it was actually greased. However, it seems to be a special version of the HFL2026 with small protruding ridges for gripping the plastic. Do you know the name or part number for that so I can order one? I don't think a standard one will stay in place because the smooth part of this one slides in and out quite easily; it will probably just spin around under high torque. I guess I could try epoxy but it would be better to have the proper part.

 

[limiky]

Seems no knurled hfl2026 exists to be bought online..probably bikee bike does the knurling with a leathe on the roller bearing.
I don't like this, the part should be engineered to be user swappable when warranty period ends.

Anyway.
Are you sure it is the bearing rollers that failed? At this point seems that the contact area where torque transfer happens is only there because of the knurled surface..and it could be there that a decoupling happens, and surfaces slide..

In your photo, some rollers of the central row seem to be a bit recessed, but i've never had a new hfl2026 in my hands to compare the rollers position when new and surely functioning

 

[neptronix]

I bet epoxy would hold that thing really well with all that surface area.

 

[limiky]

I bet epoxy would hold that thing really well with all that surface area.

Even better than the plain interference fit it has originally

 

[limiky]

Writing from a phone is a pita, sorry for the double

 

[zizajaun]

Seems no knurled hfl2026 exists to be bought online..probably bikee bike does the knurling with a leathe on the roller bearing.
I don't like this, the part should be engineered to be user swappable when warranty period ends.

Anyway.
Are you sure it is the bearing rollers that failed? At this point seems that the contact area where torque transfer happens is only there because of the knurled surface..and it could be there that a decoupling happens, and surfaces slide..

In your photo, some rollers of the central row seem to be a bit recessed, but i've never had a new hfl2026 in my hands to compare the rollers position when new and surely functioning

Pretty sure the rollers failed and Bikee said so too. But I don't have a very good understanding of how the system works so I could be wrong.

Yes, I'm thinking the bearings recessed rather than falling out when the plastic broke. I have a new (non-knurled) HLF2026 arriving tomorrow so I'll try epoxy first, but Bikee said I need a whole new cog, so I've ordered one from them as well (55 euros shipped). It'll be good to have a spare in any case, given that this seems to be a common occurrence...though Bikee said it normally happens from using the walk assist or throttle, which I haven't even installed.

 

[limiky]

Sorry guys, life's busy theese days, as soon as i can i will open my unit and post photos of the current mod on it, so maybe someone can take inspiration from it. I'll install the motor on another bike (not the dh shredder) so a lighter use of it may be sufficent to not punish it till failure point.

 

[neptronix]

Would love to see it!