Optimal Hub Motor design? Break down of BionX D-Series

That weekend i will lace it up and try it out with a Max-E and 22s.
More pics:

side by side with MXUS stator (340/11,5 vs 198/45mm):

ESyguK.jpg


sidecover with noise dampening foil (it's name is dynamat and i believe this is normally used in the car hifi area):

7jJUtI.jpg


2CF9S9.jpg


brake rotor flange:

11OG8A.jpg


Dual bearing on both sides. those are on the freewheel side and it looks like they suffering corrosion and they also do not spin that easy.
I will definitely replace them later:

t2GPvN.jpg
 
liveforphysics said:
Miles has some incredibly impressive motor magnetic designs. Miles hubmotor magnetic simulation is about perfectly what a hubmotor needs to be. It just comes with the structural packaging challenges to be solved in a way that doesn't substantially increase stator thermal impedance.
Yeah i know about this magnetic designs. you mean this topic i guess: https://endless-sphere.com/forums/viewtopic.php?f=30&t=57371&start=250#p863515 (where i was quite present as well).
the last post there was in 2014 and no further development happened for the structual parts.
Miles motor design had very low iron losses, but if i compare the copper losses at given torque, the D-motor is not that much worse.
It's peak ETA of above 94% is fantastic! it's a pity that this motor never was fabricated or developed to the end.

That motor likely goes up 4-5x in continuous power handling through a widely vented cover, but it adds aero drag and decreases weather and debris protection.

Yes you could be right with that. That stator has a large surface area (especially the end turns) and there will be plenty of airflow just from centrifugal forces if it is spinning. I will first try it on the bike, and later i will think about how to improve the thermal resistance.
 
madin88 said:
Justin, i wonder if you get the same 92,5% peak eta as the spreadsheet says, and what about the eddy losses at higher RPM?
are they considered in the simulation?

That's exactly what this is about:
justin_le said:
In any case it gets us in the right ballpark and looks like a hysteresis drag of ~0.44 Nm (very similar to the Bionx PL350) with a slope of 0.008 Nm/(rad/sec) (a fair bit worse than the 0.003 with the PL350, but not suprising given the 4x higher eRPMs of 44 poles vs 11 poles)

The data would be better if you had more data points on the no load power draw over a wider range of RPMs, but from the 3 data points you gave it looks to be about 0.008 Nm of eddie current drag for each rad/sec of wheel speed, and yes that is very much part of the simulation. At say 300 rpm, the drag from eddy losses is 300 rpm * 2 Pi / 60 (rad/sec)/rpm * 0.008 Nm/(rad/sec) = 0.25 Nm. At 1200 rpm, you'd have 1.0 Nm of eddie current drag, and so forth.

The online simulator shows the net efficiency from battery output to mechanical motor power, so it includes the controller losses too. If you want to see the predicted motor-only efficiency, then you need to do a custom controller and set the controller resistance to 0 ohms. I'm seeing more like 88-89% than 92% with the values put in but at this level of detail we'd really want to dyno test to be sure. So the peak efficiency isn't that different from other more standard DD hubs with 0.35mm laminations, but the efficiency under higher loads is quite good given the low winding resistance for that KV.
 
madin88 said:
@ Justin, i'm not familiar with the new functions of your simulator, but could it be that you forgot to set human input to zero on system B?

Indeed I noticed that too after uploading the screen capture, although that just changes where the cursor location is placed by default after simulating. In this case I was showing the entire graphs and not any cursor data so it doesn't affect anything.
 
madin88 said:
liveforphysics said:
Miles has some incredibly impressive motor magnetic designs. Miles hubmotor magnetic simulation is about perfectly what a hubmotor needs to be. It just comes with the structural packaging challenges to be solved in a way that doesn't substantially increase stator thermal impedance.
Yeah i know about this magnetic designs. you mean this topic i guess: https://endless-sphere.com/forums/viewtopic.php?f=30&t=57371&start=250#p863515 (where i was quite present as well).
the last post there was in 2014 and no further development happened for the structual parts.
Miles motor design had very low iron losses, but if i compare the copper losses at given torque, the D-motor is not that much worse.
It's peak ETA of above 94% is fantastic! it's a pity that this motor never was fabricated or developed to the end.

That motor likely goes up 4-5x in continuous power handling through a widely vented cover, but it adds aero drag and decreases weather and debris protection.

Yes you could be right with that. That stator has a large surface area (especially the end turns) and there will be plenty of airflow just from centrifugal forces if it is spinning. I will first try it on the bike, and later i will think about how to improve the thermal resistance.



That was my fault for dropping the ball on stator MFG for Miles motor. Fortunately, I have a friend who recently got a very bad ass laser cutter with a big enough table to do 4' * 8' sheets of material with high precision. :)
 
some news:

bofore i'm sharing more pics and info about that motor conversion, i want to report how the test run with Adaptto controller was.

in short: DISAPPOINTING, but now from the beginning on..

I swapped the phase and hall wires as long as the controller did show a "hall offset of 0 degree". I did so because it was mentioned that it could have an effect on performance if the offset is 60°, or 300° or something like that..
Ok, than i did manual tuning and everything seemed to be fine, like the no-load speed were exactly how it should be concidered the battery voltage, motor kV and about minus 7% because of sine wave.
the no-load speed was 75kmh at 84 volt, and the display showed a consunption of 120-160watt which should be ok.

Already during the setup i noticed the terrible noise the motor makes. When i did the tests with the bare motor, the sound was quite normal, but now as it is laced up, it is a differnt thing.
It seems like the wheel with all it's parts (rim, motor housing) amplifies the switching frequency like a loudspeaker so much that it becomes definitely annoying. It doesn't matter if WITH or WITHOUT throttle (when the motor is spinning down), the sound is always the same and very loud.

With the bike on the road, the noise wasn't that annoying anymore, but something definitely must be wrong as the performance is very poor.
I have set 60/100A in boost mode, so above 5kW, but it felt like maybe a 500-1000W powered bike, and the temperature limit of 100°C was reached very fast.

It looks like Adaptto doesn't work with that motor, and i personally don't know on what settings i could turn further to make it work properly.
It is a pity!

Whats planned now:

As soon as i received the 100V 12Fet controller from PV, i will hook it up on that bike together with a CA.
Adaptto is known to have issues with a few motors (like Crystalyte TC series), so there might be a good chance that it will work.

here are the rest of the pics i took during the conversion:

the axle is a piece of art. it has 3 wire channels and nowhere sharp edges. Bravo Bionx!

4SzFuG.jpg


the wire channels also have the function for torque transmission.

0YntpB.jpg


wrVn6a.jpg


i was able to fit 3 x 1,5mm² wires into one single channel, so the total cross section is 3mm² per phase.
the third channel was for the sensor cable

ni7jsh.jpg


the location of the two thermistors (one is NTC 10k and the other KTY type)

9IFj7v.jpg


lacing straight pull spokes is really easy compared to normal J-bend spokes

96dnwY.jpg


TBLDSi.jpg


that special torque arm design doesn't shift the axle center (like usually high power hubs do with M14 or M16 flat part), so no problems with the installation of the disc brake caliper

8n9giZ.jpg


wT8i4C.jpg


and finally the D-motor on the bike

arKm2w.jpg


rS128T.jpg


Guys, i don't think that large diameter motor looks that ugly or strange :)
whats your thoughts?
 
I think it looks fantastic! Very well done indeed.

Motor design is excellent IMO and numbers look great. Adaptto is definitely not a great fit for this motor, seems to be a very narrow range of motors that work flawlessly with current adaptto firmware. Very interested to see results with another controller.

Do you think it would be possible to create a double wide motor using two rotor rings, double stack of laminations and a re-wind?
 
Ohbse said:
Do you think it would be possible to create a double wide motor using two rotor rings, double stack of laminations and a re-wind?

Basically you would need a custom axle (M14 or better M16) with somekind of a larg "disc" in the middle where you can bolt on the two stators.
Than you would need to make the bearing flanges larger on a lath so you can fit larger bearings for the new axle. There should be enough room for that..
Making the rotor double as wide with two existing ones and stock sidecovers would be quite easy to make. the 12mm gap between the sidecovers could be filled with epoxy and carbon rovings, or a custom aluminium ring with cooling fins (-> $)

however, following things would keep me from doing so (at least at the moment without any successfull test runs):

- the loud noise (which wasn't controller related)
- the low capeable RPM because of the laminations (which are only suitable for maximum of 400Hz regarding Miles comment in another thread).
- i am not sure if the area where the sidecovers are attached to the spoke flanges would be strong enough for twice the torque and probabaly 10 times the power they where initially made for :lol:
 
I have to say this is the coolest hub motor ive seen in a long time! Cannot wait to see the performance results/vid

Thanks for the pics/info
 
You can look up Farfle's goldenmotor (magic pie?) dual-stator build thread for some of the issues you might run into on doing such a rebuild.


Regarding the low-rpm, if all you were after was low-speed torque it wouldn't really matter, either for use in a wheel on a cargobike or trike, or for use in a middrive with transmission and/or gearing up to whatever speed you were after.
 
It's a nice motor and you did excellent work with it. I'm excited to see how it runs when you find a controller pair that's a good fit.
 
liveforphysics said:
It's a nice motor and you did excellent work with it. I'm excited to see how it runs when you find a controller pair that's a good fit.

Yeah it would be a pity to not get it running.
The design and fabrication quality of the axle, stator spider and flanges is awesome - even the part where the brake rotor is bolted to has improved form stability (for beeing lightweight) :)
Only thing i don't like so much are the plastic sidecovers which are likely the main reason for the loud noise. It is really stupid to go for a lightweight material and than add heavy noise damping foils.

btw: weight of the case is 1780g incl sprocket
and the stator 2270g with all my wiring
edit
entire wheel: 6,3kg (MXUS 3000 e.g. is twice as heavy also with bicycle wheel)

I wonder if the bipolar hall sensors could be the reason why it doesn't work with Adaptto?
They sense both, north and south pole, while usual hub motor sensors like the SS41 only react to one pole.
As there is a gap between the magnets, i guess the singals will look different, but i don't have the equipment to measure it...
 
madin88 said:
liveforphysics said:
It's a nice motor and you did excellent work with it. I'm excited to see how it runs when you find a controller pair that's a good fit.

Yeah it would be a pity to not get it running.
The design and fabrication quality of the axle, stator spider and flanges is awesome - even the part where the brake rotor is bolted to has improved form stability (for beeing lightweight) :)
Only thing i don't like so much are the plastic sidecovers which are likely the main reason for the loud noise. It is really stupid to go for a lightweight material and than add heavy noise damping foils.

btw: weight of the case is 1780g incl sprocket
and the stator 2270g with all my wiring
edit
entire wheel: 6,3kg (MXUS 3000 e.g. is twice as heavy also with bicycle wheel)

I wonder if the bipolar hall sensors could be the reason why it doesn't work with Adaptto?
They sense both, north and south pole, while usual hub motor sensors like the SS41 only react to one pole.
As there is a gap between the magnets, i guess the singals will look different, but i don't have the equipment to measure it...

Swap to SS41 Honeywell halls?
 
madin88 said:
I wonder if the bipolar hall sensors could be the reason why it doesn't work with Adaptto?
They sense both, north and south pole, while usual hub motor sensors like the SS41 only react to one pole.
As there is a gap between the magnets, i guess the singals will look different, but i don't have the equipment to measure it...

here is a error: SS41 are bipolar as well so they sense both north and south pole.
the confusion came because the sensors on a recently disassembled mxus motor's did only react to one pole, and same thing was the case when i was repairing a defect sensor on a different no-name hub.

liveforphysics said:
Swap to SS41 Honeywell halls?

thats the plan for a second try with adaptto. it will be SS413A type as it is basically a better SS41:

-improved thermal stability(drift)
-lower min supply voltage (3,8V vs 4,5V of the SS41)
-typically faster switching

the SS411A would have lower gauss sensitivity, but i don't know if that would have an effect on the fanction..
 
Hello madin99

Have also a defective D - engine I would like to open.
Your pictures and reports help me a lot.
Have some questions
Do you have the two engine halves glued together or otherwise connected?
With which controller the motor works satisfactorily
Thank you
Anton
 
The covers are designed that mainly one holds the magnet ring, and the other is bolted to it.
But there is also some glue between the covers.

I have used a heatgun and believe it was helpful as nothing broke.

About the controller i can't say anything because i have not tried it yet with the PV, but it definitely is a 3-phase BLDC with 120° hall sensors location.
 
madin99

Thanks the demontage is clear to me
Only the assembly of the two halves
Were they glued?

Anton
 
SeiTon said:
Only the assembly of the two halves
Were they glued?

The largest part of the magnet ring is attached to one cover only and onthing broke here, so there was no need to glue the two halfes together for functionality.
I will add glue later (for sealing it up) once i got it running well, but now for the test rides the bolts should be enough.
 
Hi. I am sorry for asking this here, am sure I could learn it alswhere, but please

How does torgue work on motor like this, only from one coil to one permanent magnet on the shell? Or do they work in pairs, how it changes when you accelerate and how it is at full speed

Edit: I think i figured it out, so rotor is always say south pole, and then you have 3 phases on stator, 1/3 of circuference, and that then moves part of rotor forward, as it moves away from north pole phase on stator, and so on? So its thirds of stator that move around and that pushes rotor forward?

Edit2: what if you had different sets of south poles on rotor, that you would turn with your pedals, could that help in anyway? I wonder how to best combine my pedaling with torque from motor
 
I rode a demo ebike at Interbike that used this BionX D-series unit, and it exceeded my expectations even in stock form.

It had a tiny handlebar mounted microswitch that worked as a throttle, and would give it full torque available whenever you pressed it.

I would say it was top 5 or so for speed and acceleration out if the many dozens of ebikes I tried. It also feels like stock it's at least a 1kW bike, but I didn't have anyway to measure its power in or out during the demo track laps.
 
Also, the D-series is super quiet. Years ago, it was the first hubmotor that really grabbed my attention with how quietly it runs. I investigated further, and that's when I first heard the term "sine wave" controller. Of course we now have many more options when it comes to choosing a sine wave controller, compared to just a few years ago...
 
I have a local BionX dealer in Laguna Beach, CA whose store is adjacent to a very steep hill. The 500D conversion that I tested flew up the hill with minimal pedaling. Also had impressive braking from regen on the descent.
 
I still dont understand it, when I turn the pedals, the rotor with permanent magnets turns around, and as that is hapenning the motor supplies current to the phases on stator which further moves the rotor around.

Couldnt torque sensing be better done with some sensor on permanent magnets of rotor. What I know about Bionx bike system is that it use strain gauge, wich i guess its to sense pressure on the pedals.

What am thinking possibly it can be better done with sensing the permanent magnets moving around the stator...
 
Back
Top