Neumotor 8057/75 middrive project, belt drive, adding halls, 10kW+

Punx0r said:
Check how much power your motor actually outputs at 500rpm - it will be a lot less than 14kW!

As a sense check I would take motor stall torque, multiply by driver pulley radius and check how much belt-pull force that will give. You may find the belt is actually within the belt spec :)

Thats absolutely right.
The phase amps will anyway limit torque (and input power) at take off and the second thing is saturation of the core.
during acceleration it will be like this:
1) controller operates at phase amp limit
2) controller operates at battery amp limit (-> maximum power)
3) no more limitation after acceleration -> hopefully! :mrgreen:
 
ridethelightning said:
large driven pully not included in list?
let me guesse, it takes balls of steel just to think about the cost?... :mrgreen:
:) I could have bought a 3k hubmotor for what i payed for the pulley alone. you can expect minimum 200-250USD for the fabrication of such part.
The good thing is that i have CAD and could do this part by myself. It is nice to design parts on the PC, send it via internet to one manufacturer, and two weeks later it arrives by post.
altogether awesome solution in evolution from hub drive. i have no experience here, what do you think the difference in performance dynamics will be compared to the hub, more in terms of power, with the unsprung mass issue aside?

im thinking better accelleration and hill climbing at low speeds, perhaps less torque at higher speed and maybe less top speed?

Regarding torque, or losses per given torque output it can be calculated. Lets do a theoretical comparion between my belt drive, MXUS 3k and QSV3.
I would like to have at least 200Nm on the rear wheel, so lets compare the copper losses at this level of torque.

QSV3 50h 4T:
11.4kV and 45mOhm phase to phase resistance (measured)

kT = 0.84Nm/A -> so it neeeds 238A for 200Nm of torque at the wheel
238² x 0.045Ohm = 2500W of heat

MXUS V2 45h 3T:
12kV and 55mOhm (measured)
kT = 0.8 Nm/A -> 250A for 200Nm
250² x 0.055 = 3400W of heat

Neumotor 8057 / 75 with 1:6 reduction:
75kV / 6 = 12.5kV and 23mOhm
kT = 0.127 x 6 = 0,76Nm/A -> 263A for 200Nm
263² x 0.023 = 1600W of heat

This is only theoretical without taking saturation (or kT curve) into account.
At 200Nm the MXUS is definitely well in saturation, and the QS and Neumotor (at 33Nm) are probabaly also close to the limit.

How my belt drive will perform at high speeds i'll see when it runs, but the low no-load losses of the Neu looking promising so far.
With its kV of 12.5 it should be a bit faster as the MXUS which did around 80kmh without any field weakening on the bike.

The drive has lowered the weight by more than 5kg compared to the MXUS, or >8kG compared to the QS 8)
As the motor still is located on the swingarm (and located not that close to the pivot in my case) it still adds partly to the unsprung mass, but it should lead to a much more comfortable ride especially at not so smooth roads.
 
DanGT86 said:
How did you generate the tooth profile for the cad work?
I have asked Gates and they where so kind sending me the profile for 108T.
But for CNC fabrication it is essential that you adapt the shape (concave radius) so they can use at least a >2mm diameter mill for it.
Normally belt pulleys are gear hobbed and not milled. Thats the reason why guide flanges are separate parts which will be attached later.
 
I have some good news for you guys!

On Thursday this week i installed the swingarm back to the bike, hooked everything up and did a test run.
Test run probably is the wrong describing as it was going without any worth mentioning issues and it was so fun that i almost would have need to shove the bike home because i overseen battery soc. It was a 60km ride in a hilly area.

First i must mention that i wasn't able to set more than 200A phase because the thin 12AWG motor wires (i used a grin tech motor wire only temporarly for the test) got so damn hot it almost melted when i was going uphill, or after a few hard accelerations.
The battery amps were set to 120A and i was seeing above 9kW peak.

Setting up the Adaptto controller for optimal efficiency is definitely a nightmare.
Autodedect "detected" a hall angle of -2,5° to be optimal, but it wasn't (i adjusted the halls mechanically to be exactly 0° with the oscilloscope).
During the ride it turned out that an angle between 0 and -0,8° is optimal (unfortunately a value between cannot be chosen as the numbers jump).
Same with PWR timing. 0,35 - 0,53 - 0,71.
At 0,71 together with half open throttle the power was pumping while i was maintaining crusing speed. 0,35 was a bit too low as i noticed a bit more heat in the motor, so i set it to 0,53°and got good results.

Well, how is the performance, noise etc?
It is on the same, rather better overall performance level as the MXUS 3k which was previously installed on the bike.
The 8057 does heat up a little bit quicker due to much lower copper and iron mass, but it was staying below a dangerous temperature most of the time due to good cooling.
When i was riding around 75kmh continuously on the road in a hilly area, the temperature was alternating between 65 and 90°C.
Definitely not bad for a 2,7kg motor with a reduction of 1:6 8)

But the best thing, or the thing which was noticeable the most was the MUCH better feeling due to the lower unsprung mass.
I can now lean into corners without need to be afraid to loose traction if the road isn't totally smooth. When riding over potholes with the MXUS it felt like someone is giving me a looong and hard kick in the ass. Now the kick feels short and very well dampened by the suspensions.
It is no comparison!! It rides soooo much more comfortable and better now 8)

Now i need to swap the phase wires from 12AWG to something like 8AWG so i can crank up the phase amps to 300A+.
I also noticed that there is some strange noise coming from the bearing on the shaft side. It was defect out of the box as it has had some play from the beginning on.
Aside from the noise a defect bearing makes it also makes the rotor vibrating too. I need to swap it out for i high quality SKF.
The belt stays on the pulleys like it should. There is absolutely no risk that it could go off.
Just on the small drive pulley it sometimes is moving by 1-2mm away from the guide flange, but i believe this can be fixed by mounting the motor with a little twist.

8me10q.jpg


5AFE4v.jpg


QbDPyK.jpg


20sbI8.jpg
 
that is so awesome.
your pics are so sharp. how do you get them like that?
i thought the forum would not allow hd files like hat to be uploaded, atleast i have failed to so far.

so even though the motor is still unsprung strictly speaking, it is only 2+kg and closer to the pivot point aswell as being more centralised in relation to the rest of the bike mass, which basically makes it disappear as far a handling is concerned. well done!

how is the noise??? it would have to be better than chain drive, but does the motor sound loud due to high rpm?

i will be keen to hear how it performs with more power :twisted:
keeping the motor temps down will be the limit factor but a little cooling should be easy to set up id say...wet rag+cableties :D
 
Sweet. I guess with a little more tweaking and new wires you might even get faster acceleration and it should climb better then the previous mxus setup.

Could you do a video of the bike running? Very interested in what kind of sound that motor is making, and how easy to live that sound is. This your first mid drive? If so I guess you as me are well set in the way of silent dd hubs, noisy motors might take some getting used to?

Guess that bike is transformed in a big way now that you got all the "fat" out of the rear wheel.
 
ridethelightning said:
your pics are so sharp. how do you get them like that?
i thought the forum would not allow hd files like hat to be uploaded, atleast i have failed to so far.
i am uploading the pics to imageshack.com and share them by adding the link to the text.
When i created the account some years ago the service was free, but now they charge about 15USD per year. If i don't pay them, all images i ever linked will disappear.
how is the noise??? it would have to be better than chain drive, but does the motor sound loud due to high rpm?
It is not as quiet as a hub drive with sine wave. Probabaly similar to hub drive with trapez controller.
Above 50kmh the sound is barely hearable as the wind is getting louder.
Below that speed i can hear it and at around 15-20kmh it makes the most noise which probably has to do with the vibrations (oscillations) of the rotor and the defect bearing.
Close to 100kmh the noise is becoming louder again. Motor is spinning than close to 5000rpm :twisted:
It is hard to say at the moment which part of the noise is coming from the motor and which from the belt drive.
But it is definitely nothing which annoys me, more the opposite. I like the sound, and now people hear me coming and looking to the side if they walk over the street :)

macribs said:
Could you do a video of the bike running? Very interested in what kind of sound that motor is making, and how easy to live that sound is. This your first mid drive? If so I guess you as me are well set in the way of silent dd hubs, noisy motors might take some getting used to?
I will!
Yes, the new thing was that more pedestrians were looking at me. That wasn't the case with the hub drive, but that isn't a bad thing entirely as this bike anyway has a licence plate and insurance, so i don't need to be that stealthy anymore. I am not afraid of the cops.
 
This drive system goes LIKE HELL! :D

I swapped the phase wires for 8AWG and wired up everything clean. The strange thing was that on the first hill with WOT the motor did overheat extremely quick.
Guess what was wrong!
It was PWR timing. Yes, obviously if you upgrade the phase wires for larger ones (1:1 nothing swapped), this settings needs to be adjusted. I believe it is because of the lower resistance and inductance the larger wires bring with it.
While with the 12AWG wires a timing of 0,5x was ideal, i had to readjust it to 0,8x now for optimal ETA.
It is unbelievable that such a small change of this setting can have such a huge effect.
As everything seems to work well now, i upped the amps to 130/260A and ridden it hard at WOT.

10kW was no problem :D
The motor defintely can take this power with very high efficiency. It is insane what this small 2,7kg motor can do. Neumotor did a really good job. It works fantastic!
On the flat the speed (without field weakening) was above 95kmh. Acceleration is awesome. I would say it can be compared to a QSV3 at same input power and wheel size, and it is definitely better as MXUS 3k. I don't know if it is better as the MXUS (which was previously installed on this bike ) due the better cooling, or if it is due to the lower resistance of the windings.
The neumotor heats up quick from lets say 50°C to 80° when riding uphill, but the temp stays stable which tells me that efficiency must be quite high and it cools down really quick. I like it!

Due to the high kV of 12.5, the 1:6 reduction and my conservative phase amp setting of "only" 260A it doesn't pull up the front, but from about 5-10kmh it accelerates bad ass. The lack of "kick ass torque" is anyway a "problem" with adaptto, isn't it?

The 1:6 reduction is optimal for my demands and i would say for a street bike like this it is a good balance.

Regen braking:
It works like a hub drive. Very smooth with similar strength as one 3k hubdrive, and without any play or "click" sounds. It is fantastic!
The drag:
It is definitely higher as hub drive and noticeable when rolling down hills wihtout any braking. The bike is a bit slower now, but not much.
But i don't care about that as i don't need to pedal this bike (it has no pedals anymore), it just takes getting used to it.
 
Excellent thread Madin88
Especially the direct comparison hub to reduction and all your before and after detailed specs calcs and performance results.
Very use-full and a lot of similarities to my current build especially the belt drive single reduction.

I was worried about motor size [ mine will be a 55mm wide magnet 5kg inrunner ] but after seeing your motor performance comparable to a mxus 3t and v3 QS im happy.
The only thing i can add is my drive hub has the same m5 44mm bcd mount and ive seen online calcs saying its good for 200nm.
So real world i hope 150 to 170 , but you will see if it holds before me :) .
Keep the data flowing. iF you can highlight torque comparisons id appreciate it.
 
Emoto said:
I was worried about motor size [ mine will be a 55mm wide magnet 5kg inrunner ] but after seeing your motor performance comparable to a mxus 3t and v3 QS im happy.
Is your motor a so called big block with 80mm diameter arigap?
Yeah the Neumotor with 1:6 reduction definitely plays in the league of QSV3 and MXUS 3k.
If you like to compare your motor in terms of torque, you just would need to know the kV and phase to phase resistance.
The only thing i can add is my drive hub has the same m5 44mm bcd mount and ive seen online calcs saying its good for 200nm.
So real world i hope 150 to 170 , but you will see if it holds before me :) .
Keep the data flowing. iF you can highlight torque comparisons id appreciate it.
The LMX hub features M6 bolts on the sprocket side :)
Also the rest of the hub looks strong so i have no worries that it could break on my usage.

The last ride i thought lets see how many phase amps i can take, but at 340A the Max-E gave up. Three FETs blew, and the other three of the same phase had a short, so 6 dead FETs in total.
the conducting paths near the blown parts turned into plama as well, but thats not hard to repair. The best thing is the driver boards seem to be OK (at least no visable damage). I already have new FETs on order and i hope i can repair it the next days.
Tomorrow i will share some pics of the disaster.
 
340 amps through that motor is pretty amazing......so was it a monster before the MaxE gave out? Most things run best right before they die! Think 2 stroke!!!

Got any video yet?

Tom
 
Yeah 2 stroke the effect is well known, for electric motors I am not so sure. Right before the magic smoke we likely see increase of temperature, increased resistance, which would decrease power rather then adding power. So for electric my guess would be the other way around. Right before melt down and plasma out power would likely be less.
 
madin88 said:
Is your motor a so called big block with 80mm diameter arigap?

Yes but bigger than a big block :D with a magnet/stator width of 55mm, Air gap would be around 80.
If their rpm is correct a kv of 62.5, but i dont have phase to phase resistance.
If i can get it in the mxus 3t league for torque that should be a good starting point , its a custom stand up scooter , build thread link is in my sig.

The LMX hub features M6 bolts on the sprocket side :)
Also the rest of the hub looks strong so i have no worries that it could break on my usage.

Ow thats good ...Thats funny as my plan was to redrill / tap my hub to m6. for peace of mind.
Keeping an :shock: on your evolution.
 
litespeed said:
340 amps through that motor is pretty amazing......so was it a monster before the MaxE gave out? Most things run best right before they die! Think 2 stroke!!!

Got any video yet?

Yes it had much more punch at take off, but the controller quit already at the first try after 20m.
When the bike is running again the first thing i will do is to make a vid

Emoto said:
Yes but bigger than a big block :D with a magnet/stator width of 55mm, Air gap would be around 80.
If their rpm is correct a kv of 62.5, but i dont have phase to phase resistance.

nice! Looking forward to your reports about the performance.

If you like to measure the resistance without buying special equipment you could do the following:
Charging a battery through the motor and measure the voltage drop with a normal DMM.
Connect charger minus directly to the battery, and charger plus to one phase wire. Than close the circuit from any other of the two remaining phase wires to battery plus. Turn the charger on, note the current and measure voltage drop between those two phase wires.
 
Some pics of the properly wiring:

BnFZRM.jpg


dual 4mm² PFTE phase wires

k5aJol.jpg


j6guus.jpg


oLuoLB.jpg


sensor wires run through another tube

e6ZwVT.jpg
 
It is crazy how much soot there was inside the controller. Near the blown FETs the board got toasted and traces melted down.
I think that i have time later today to solder new FETs.

Peak power was 32kW when it blew. who can top it? :lol:

OyyCoI.jpg


rDvJAg.jpg


i already cleaned the board and broke off the dead FETs

VPGh0j.jpg


eunmfY.jpg
 
Thanks for posting the detailed images. Inspirational for an amateur like me. I hope you can successfully repair the controller. Did you choose the PTFE to reduce size or just for general durability?
 
32kw! no wonder it blew to crap!!
ill be following the max-e repair wth interest. i have a mini-e that i need to repair that is in similar condition :|
 
32 KW is just crazy. At your voltage that is just a crazy amount of amps. With me only hitting 14 kw my MaxE isn’t being stressed even a little bit.

Tom
 
ridethelightning said:
32kw! no wonder it blew to crap!!
ill be following the max-e repair wth interest. i have a mini-e that i need to repair that is in similar condition :|

That 32kw would be to good to be true. it was just the peak power at the moment when it blew. I have set batt amps to 150A so the real power was more like 12kW.
There is good luck that you can repair the Mini-E if the driver boards look ok :wink:

I thought about using IXYS IXFX420N10T FET's instead of the IRFP4468 as this should be a much better FET according to expert opinion, but i am not sure if the driver boards would be suitable and i would need to swap all 18 pcs which is much more work and cost about 130USD.

https://www.mouser.at/ProductDetail/IXYS/IXFX420N10T?qs=6UKyFUzHTlndB8hPmmnDsA==
http://52.25.253.50/forums/viewtopic.php?t=65297&start=475#p1327958

On a Mini-E i would definitely use them!

hmm i think i should start a new repair thread for Adaptto.
 
madin88 said:
hmm i think i should start a new repair thread for Adaptto.

that would be an awesome idea. also including cap upgrades etc.
so many people out there with adaptto controllers, sooner or later will probably break and repair service seems non existant at the moment.
it seemed to me that the actual power board is actually not that complex, just the traces are fine and use smd.
i even managed to source spare micromatch connectors for the ribbon cabling

the driver boards may be another matter i suppose. send to ch factory and get them copied? 8) :D
 
Rube said:
Did you choose the PTFE to reduce size or just for general durability?
Those PTFE wires are much thinner compared to silicon wires and they can also withstand quite high temperatures. On most hubmotors this kind of wires are used.
The thing is If you want to put maximum cross section inside a tube with given dia than you should go with six phase wires instead of three. For three wires the max size would have been 6mm², and so its 8mm².
 
Some news:

I repaired the melted power traces on the Max-E and soldered one FET to each side in order to see if it works or not.
The motor did spin up properly (i attached the wires 1:1, no reset or new adjustment of the settings were neccessary). BUT later there came the bad surprise!

After installing the remaining FETs and the caps (had to remove some for soldering), the first test run was successful, but later it turned out that it shows occasionally "protect", and when doing autodetect it never did not even finish the first step and showing "halls E1".
Strangly, two times out of three the motor did spin up, but otherwise it got stuck - no matter if i apply just a little bit or even full throttle.
So something else got distroyed when the FETs blew. I believe the driver boards.

Than, luckily a friend lend me his Midi-E and he allowed me to modify the power stage in order to make it more stable when using it on my Middrive.
The "problem" with the Neumotor is its low inductance of 35-55µH (it depends on rotor position).
I don't want to go too much in detail, but it means that current rise time (delta I / delta T) on this motor is about 5-10 x faster (less time) compared to a QS or MXUS hub motor which have inductance values in the range of 150-300µH.
The lower the inducantance, the quicker the controller must be able to control the phase current. Another thing is higher current ripple will lead to higher stress on FET's and caps.

To reduce the current and voltage ripple in the controller i have beefed up the copper traces on the power stage and added about 10 pcs ceramic caps in total as close as possible to the FET's. Wires swapped all for 8AWG. Electrolytic caps changed to Rubycon ZLJ.

The supply line with so much capacitance and it's low ESR should be now "stiff as hell" (less noise and better efficiency), but the best thing my bike is running again and i almost drove 400km since i have the Midi-E installed 8)

I am pushing now 250A phase into the motor.
note: the former Max-E died when i set above 300A, but this was one of the oldest controllers and Adaptto has revised the power stage later so it could be the case that it is stable now at max settings, but i will not try to find it out! No, i will left it at 250A and will be happy with the power.
The battery amps are maxed out and it does 8-10kw now.

I have done a side by side test ride with another bike with a QSV3 and i can tell you that this little motor can keep up with it 8) , but later more and also vids :wink:

some pics of what happened:

beefed up copper traces:

ElZDgI.jpg


Rubycon ZLJ caps with 6,8µF 100V 3025 size underneath:

D4wAza.jpg


8AWG wires:

xUxx8y.jpg


d3NgT6.jpg
 
Hi Madin, great news that you have it running again, and have figured out how to get the controller to run the motor well.

Those (and similar) motors have a lot of potential, but are obviously a bit trickier to drive well than the more "conventional" motors. It is really good that you have the experience and skills needed to understand and solve the problems you had.

Looks like you will get a lot of enjoyment from that bike!
 
Back
Top