mxus xf15 gear hub motor teardown

mechhead

10 W
Joined
Mar 10, 2017
Messages
90
Location
adelaide south australia
Here is a new geared hub I've been looking at as something better than the swx bafang for a variety of reasons that will become aparent.

having ordered two to test and hot up, they arrived swiftly direct from mxus factory

Photos to come

a cassett version to go in a wheel with steel gears (aliexpress) 47.5mm dia 36t
and a freewheel threaded version to convert to a crossbreak style mid drive

inside the motor

6 turns 10 strands .5mm copper
the usual nylon gears
a multi screw mounted sun gear ( good potential for regen mod )
20 pole 18 slot 120mmx18mm stator with .35mm laminates

it was a small battle to work out which version of this motor has the meatiest windings

when asked about motor speeds i eventually got the spec sheets for their two fastest motors
and they claimed that the 48v 386rpm version was fastest so i asked for the spec sheet for the fastest 36 volt one and ...

48v 386 rpm
36v 349rpm

ok convert to kv and actual motor rpm gear ratio 5:1

so motor a (48v 386rpm 10t) 5*386=1930 rpm of motor @48v

to derive KV : rpm/volts 1930/48=40.2833 rpm/v



so motor b (36v 349rpm 6t) 5*349=1745 rpm motor @36v

to derive KV :rpm/volts 1745/36=48.47222 rpm/v

so we have a winner the badly named "36v" motor is the badass with 6 turns to the others 10

so speed it up to 48v

motor b (36v349rpm) take motor KV and multiply by new voltage of 48



48.47222*48=2326.667rpm now add gearbox reduction to give hubshell rpm or that of thebuilt wheel at noload speed and you get 2326.667/5=465rpm of wheel



so to compare motors properly at same voltage noload speed and you get

motor A (48v386rpm) @48v = 386rpm

motor B (36v349rpm) @48v = 465rpm

so commutation frequency 2326rpm X 10 pole pairs / 60seconds = 387hz

correct me you lot if im wrong but isn't 387hz a way short of sensible frequency limit for .35mm lams ?

what if we delta it

any here are spec sheets of motor load curves if i can make the attachments work
 

Attachments

  • XF15 36V 349rpm.ods
    18.8 KB · Views: 433
I have been looking for 47.5mm steel gears 47.5mm for the mxus xf15 at aliexpress and can not find any,can you please post where you found them from?
I only found these https://www.aliexpress.com/item/32801000901.html?spm=a2g0o.productlist.0.0.42671799QfuqoS&algo_pvid=19a334e1-7a2f-4eb2-acb1-53dd6eed45a4&algo_expid=19a334e1-7a2f-4eb2-acb1-53dd6eed45a4-0&btsid=6cbd2067-f5a3-4496-bc11-8c2c77f082fd&ws_ab_test=searchweb0_0,searchweb201602_1,searchweb201603_55 witch say that they are 47.5mm 36T Thickness:12mm but the mxus xf15 that i have says that the nylon gears are 47.5mm 36T Thickness:13,5mm https://www.aliexpress.com/item/32830428964.html?spm=a2g0s.9042311.0.0.52ba4c4d3tr9gz
So will the 12mm thickness steel ones work?
 
Ordered this motor today, Putting it in a hardtail frame, want to run it at 36v at first and then maybe upgrade later to 48. I will post results also
 
vasam :

i have ordered a set of those 12mm wide gears, I rekon they will work well, a little extra clearance wont hurt, the last set of steel gears i got were for the smaller older bafang hubs, and they were .5mm wider than the originals and they rubbed on the face of the one way clutch housing. The clearance was close to perfect and only the deformations left from hobbing the gears were rubbing.

a lick on some coarse wet and dry sand paper removed the deformations and flattened the gear wheel and after polish no more rubbing.

So in light of that experience and steel being miles stronger than the nylon gears i doubt narrower by 1.5 mm will cause any problems

I will let you know how well they fit when they get here.

Enewb :

Which version of the motor did you get ?

There are at least three different winding specs available but the sellers dont know how to get an apples to apples comparison out of their own specs.

They give different speeds sold as 36 or 48v versions

The spec sheets give turn counts and the mxus factory will give the to you if you get talking to them on alibaba
above eas the spec sheets for the 6t and 10T.
 
I just found this complete gearset 47.5mm by 13.5mm 36t

https://www.aliexpress.com/item/4000370794990.html?spm=a2g0o.productlist.0.0.18045fde2PcbII&algo_pvid=b81053f7-5d10-4acc-a35e-0ba5f42eceb0&algo_expid=b81053f7-5d10-4acc-a35e-0ba5f42eceb0-21&btsid=cdad4d8d-f405-46b6-bf74-1b82201d095d&ws_ab_test=searchweb0_0,searchweb201602_7,searchweb201603_55
 
mechhead said:
Enewb :

Which version of the motor did you get ?

There are at least three different winding specs available but the sellers dont know how to get an apples to apples comparison out of their own specs.

They give different speeds sold as 36 or 48v versions

The spec sheets give turn counts and the mxus factory will give the to you if you get talking to them on alibaba
above eas the spec sheets for the 6t and 10T.

Unfortunately they couldn't give clear info on the turns or the difference between the 36/48 volt versions. All I could gather was that the motor supports both voltages, but the rpm on each was not mentioned, even if there was a difference.

Will just have to take a gamble
 
I orderd this kit https://www.aliexpress.com/item/4000258829938.html?spm=a2g0s.9042311.0.0.40ae4c4dHdGRlT and from what the seller told me the motor is this one https://www.aliexpress.com/item/32825136459.html?spm=2114.12010612.8148356.1.723776ddeHfVDK
witch it says it is 48V Cassette 425RPM .
So mechhead it looks like mine is the motor B (36v349rpm) @48v = 465rpm one.
(Customer reviews say that At 54 volts on the idling bike gives out 63 km/h(about 39 mph) on a 28 "wheel,and real top speed on 28" or 700c wheel is 44-46km/h(about 28mph).
 
It's usually a dead end to take the Chinese spec.s literaly and their graphs have been shown to contribute to early dementia.
There's a Ezee "350" very high speed range motor in the Motor list @ the Ebike CA (Grin) sim. if you want something to play around w/.
 
I have had this motor installed on my bike for 2 months already and have done almost 1000km on it ,with the 22 ah controller and a 48v battery at level 5 assist it outputs near 1200w ,for my first ebike build this much power is even too much for me and i am quite a heavy guy(250 pounds)!

But i cannot get the full potential of this motor because i have an issue since day 1 of the instalation of my kit,i cannot start with high level of assist at full throttle nor with pedal assist without the motor cuting off,and power cut offs acure even in other ridind conditions as like situations where i ask for high discharge!

After comunicating with the Chinese seller i bought the kit from he told me that this is not normal and he proposed me a partial refund and a temporary solution that was to set in the KT-LCD8 controller settings parameter C5 from 10 to 1 where 10 stands for instant foul power either in PAS or throttle and 1 stands for progressive power (starts with 100w then goes 300w and so on untill it reaches max output!),i could live with this problem until now but day by day it is starting to annoy me pretty bad ,yesterday me and a friend went on i bike ride and on a steep hill (17-20% incline)he with his 250w rear hub and alot of leg power reached the top but mine once the motor went under load and stress it cut of and left me right in the middle of the steepest part of the hill!
I have taken the battery to a technician and he checked the the BMS outputs and all 13s gave same output of 4,15v when almost fully charged,and even dispite this the Chinese seller sent me a new BMS(same 30a BMS)thinking maybe it was the BMS fault causing my problem,so i changed the BMS but still my problem exists !
Can anybody help me with my issue?
Could it be a faulty controller?
Motor?
Battery ?
 
if the display actually shuts off when this happens, it's probably the battery cells being unable to handle the current demand.

if so, then testing their static voltage at (or near) full charge tells you nothing about the battery at all, except that the cells can reach full charge and aren't dead.

it doesn't tell you if the cells are mismatched, which will leave them unbalanced as they discharge, more and more the closer to empty they get, but also that they aren't equally capable of handling current draw.

if one or more groups of cells is considerably lower in capability than the rest (not that uncommon in generic or cheap packs), then under load it will drop in voltage more than the other groups, and under enough load it will drop below the lvc of the bms, and the bms will protect the pack by shutting off it's output.


it's also possible that all of the cells are just poor enough in quality or capability that the entire pack sags in voltage so much that the controller itself shuts off the motor to protect the battery from overload. (this can also happen if the controller lvc is set too high for the pack, but if that's the case, then even under very low power when the battery drops below that voltage, it'll shutdown, and you won't get nearly the range you should out of the battery for it's stated capacity.


you'd have to test the pack under the same load conditions it sees on the bike to see which cell(s) are dropping in voltage enough to do this.
 
amberwolf the display does not shut off, and no error code appears on screen.
The battery i bought was supposed to be 48v 13ah with chinese brand 18650,but after i took it to a battery shop for testing to see if we can find the cut off problem, the guy told me that from his measurements he found the battery to be be 10ah instead of 13,when i asked him how does he do the testing he told me that he fully charges the battery then drains it untill it reaches cut off and found that the battery lasted 66 minutes at 500w constant discharge!
What i cant understand and is confusing to me is how come my battery is as he says 480w (48v 10ah) and not 550w(1hour and 6 minutes of discharge at 500w is or is not a capacity of 550w?).
 
valsam said:
amberwolf the display does not shut off, and no error code appears on screen.
if all that is happening is the battery is dropping below lvc, you probably won't get an error code. if there is a battery meter, and it has an instant response (they don't usually) you might see the icon go down in bars momentarily as the voltage sags under load. but probably it has such a slow response that it can't show such a short blip, even though the controller responds to the blip by shutting down the motor.

to see a voltage drop like that, an old analog voltage meter, the kind with a needle that sweeps across a gauge, works best. cheap digital meters sometimes are not fast enough to catch it; their displays don't refresh fast enough.

a wattmeter will also catch this, as it records vmin, or voltage minimum, as well as amax, or amps maximum, whcih usually occur at the same time when doing a test like this.

anyway, it's probably not the battery shutting off, but it could still be the cause of the problem, due to voltage sag under load, along with the lvc of the system being set too high.

if the system lvc is too high, the battery will not give you the range it should, at any power level, because the system will shut off early. if it doesnt' have a wattmeter to tell you wh or ah, then you don't really ahve a way to know this is happening.

The battery i bought was supposed to be 48v 13ah with chinese brand 18650,but after i took it to a battery shop for testing to see if we can find the cut off problem, the guy told me that from his measurements he found the battery to be be 10ah instead of 13,when i asked him how does he do the testing he told me that he fully charges the battery then drains it untill it reaches cut off and found that the battery lasted 66 minutes at 500w constant discharge!
how much power does your system actually use, at the level you have trouble with it at?

if it's less than 500w, then the system isn't overloading the battery enough to cause the battery to shutdown, if it worked with a 500w load test.

it's common enough for two things to cause this. the first is that sellers lie, especially about batteries. when they don't lie, they often have no idea what is actually in the stuff they're selling. the second is that such batteries often have mismatched or crappy or even recycled garbage cells in them, so they don't all have the same capacity. if the guy doing the testing logged the voltages on the cells during discharge all the way to empty, this will show which, if any, of the cell groups have cells that are problematic, as those will be lower voltage than the rest. it will also show how much voltage sag there is under that 500w load.

What i cant understand and is confusing to me is how come my battery is as he says 480w (48v 10ah) and not 550w(1hour and 6 minutes of discharge at 500w is or is not a capacity of 550w?).
this distinction may not be obvious, but it is very important. if you are talking about capacity, it is not 480w, it is 480wh (watt-hours). (or 550wh). if you don't use wh when referring to capacity, then people will think you're talking about the load on the system, in watts. ;)


what kind of load did he use? if it's not a form of active electronic load, it isn't a constant 500w, and so the actual watt-hours (and amp-hours) can't be calculated based on that, they must instead be measured by a wattmeter during the entire process.

if you don't know what kind of load he used, and how he measured the wh, you can't know if his measurement is accurate, either. if you prefer, you can assume he knows what he's doing, and that he used a wattmeter to measure it, and that he got the 10ah figure from that. but if he didn't tell you how many wh it is, then any calculation you make from the voltage and ah is only a guesstimate. ;) a good one, usually, but still just a guesstimate.

if it's 1.1 hours, at 500w constant, then yes, it should be 550wh. but if it is not constant, because it's just a fixed resistance load, then it won't be 550wh, because the watts actually goes down during discharge because the voltage drops so the current also drops.
 
amberwolf thank you very much for your help!
Should i play around if it's possible(i have the KT-LCD8 display,does it give me a lvc setting function ?) with the the system lvc if it is too high,and how much is to high and how much is too low?
On my display when fully charged my battery when new showed 54,6v and now shows (i have already done about 25 full cycle charges)when fully charged 54,1 and when riding cutts off because it is empty at 42-43 volts,is this any kind of indication of a too high set lvc?
 
valsam said:
Should i play around if it's possible(i have the KT-LCD8 display,does it give me a lvc setting function ?) with the the system lvc if it is too high,and how much is to high and how much is too low?
if you have access to the lvc setting, then you can calculate it from your battery. typically 3.0v per cell is ok for a deep discharge cutoff, but i'd stick with 3.3v or higher per cell; its not as hard on the battery. if your battery is a "48v" (rather than "52v") it's 13 cells in series, so 13s x 3.3v = 42.9v, for the lvc.

if it's already set a lot higher than that, that could be your problem. for instance, if your controller is set for a 14s (52v) pack, then the lvc is probably around at 46v or so, which for your 48v battery is still at 3.5v/cell.

but even if it is set too high, it also means your battery voltage is sagging way too much under load, if it's doing this even when fully charged. if so, the battery is still a problem, being unable to supply sufficient current without sagging too much in voltage.

On my display when fully charged my battery when new showed 54,6v and now shows (i have already done about 25 full cycle charges)when fully charged 54,1
the most likely problem is the battery is not balanced. if it has a bms with balancing function (many do, but not all) then you can simply leave the charger plugged into it, for at least several hours, overnight. if it is charged up back to the original full voltage by that time, then you're set, but will probably need to keep leaving it plugged in whenever you're not using it. if it is not yet charged back up to the orginal full voltage, you would need to leave it plugged in for longer (days, etc) until it does get back to the original voltage.

but it could also be that the cells simply won't charge up to full anymore, because of the cells aging (and probably not being very good to begin with).

it could also be the charger itself failing to provide enough voltage, but that's less likely. voltmeter on it when not plugged into battery should test that.



and when riding cutts off because it is empty at 42-43 volts,is this any kind of indication of a too high set lvc?
42-43v is good for a controller lvc. so if it shuts off then, and the display is still on, that is the ocntroller lvc kicking in.

if it shuts off then, and the display has turned off, then the battery's bms lvc has kicked in, and shut off the battery to protect it.
 
I just read the KT-LCD8 manual and it has C12 parameter where C12 is controller minimum operating voltage adjustment setting (tiny adjustment of voltage shortage +- 0,5-2V)4=Default value is 40V ,and my screen shows that i also have it in 4=Default value is 40V ,but motor cuts off at 42-43volts as i said before,is 40v estimated under load(SORRY FOR MY BAD ENGLISH!)whta i am trying to say is when the screen shows 42-43v is it realy 40V when motor draws volts nad cuts off?
Also i forgot to mention that i have noticed that the problem with my power cut off does not happen when battery is below 50% charge,when below 50% charge i can start with throttle and the motor will not cut off and also in PAS i can go up a hill in assist level 5 and motor will not cut off again!
I told this to the Chinese seller that i bought the kit and he told me O.K it's a faulty BMS problem and he sent me a new one,but after installing the new one the problem still existed,so i contacted the SELLER once AGAIN AND HE TOLD ME TO to try to find a 40A bms AND INSTALL IT(the one with the kit and the other he sent me are 30A)!
So could it be a BMS problem,my controller says Maximum current 22±1A the screen shows max output 1180w+- when my battery is fully charged (54v(full battey) X 22A(controller)=1188w)the BMS is 30A so why is the seller telling to try a 40A bms?
 
valsam said:
)whta i am trying to say is when the screen shows 42-43v is it realy 40V when motor draws volts nad cuts off?
you'd need to use a separate voltmeter, on the bike during the shutoff, to determine that, there is no way for someone not actually at your bike to know what voltage anythign is actually at.


Also i forgot to mention that i have noticed that the problem with my power cut off does not happen when battery is below 50% charge,when below 50% charge i can start with throttle and the motor will not cut off and also in PAS i can go up a hill in assist level 5 and motor will not cut off again!
too bad you didn't say that before, it would have saved all of the previous posts, becuase it completely eliminates the battery as the problem. (complete and total detail of every single symptom of a problem, given step by step, is necessary to troubleshoot remotely).

it cannot be voltage sag dropping below lvc, because sag would be worse as the battery discharges, so the prolbem would happen *more*, not less.


what this means is that the controller must be cutting power because of what it sees as too much wattage or too much current, and as the battery drops in voltage it gets below the point at which the wattage or current can be above whatever it's cutoff is.


I told this to the Chinese seller that i bought the kit and he told me O.K it's a faulty BMS problem and he sent me a new one,but after installing the new one the problem still existed,so i contacted the SELLER once AGAIN AND HE TOLD ME TO to try to find a 40A bms AND INSTALL IT(the one with the kit and the other he sent me are 30A)!
So could it be a BMS problem,my controller says Maximum current 22±1A the screen shows max output 1180w+- when my battery is fully charged (54v(full battey) X 22A(controller)=1188w)the BMS is 30A so why is the seller telling to try a 40A bms?
if the display is not shutting off, only the motor is being disabled, then it cannot be the bms, because the bms would cut power to the entire system, and hte display would shut off.

plus, with the fact that the problem does not happen at higher voltages (and thus higher power and current) it must be the controller itself (or the display, wherever the settings are actually stored), that is causing the problem.

it could even just be a bug in the firmware of the controller/display, so that above some voltage it misinterprets the power and thinks it should shut down.

either way, it's not the bms, and even if the battery's cells have other problems (not staying balanced, perhaps), it's not causing the problem you're seeing.
 
THANK YOU!!!!
I have contacted the Chinese seller asking him for a controller under warranty,so let's see how this will go!
 
note that if the controller has settings (from the factory, presumably) that are causing your problem, a new controller won't fix the problem unless it's from a different batch with different settings.

and if it's a problem in the display, instead, a controller won't fix it. (i have no idea where your system's settings are stored).
 
The seller has 14 reviews of the same kit i bought and nobody has mentioned a simular problem.
This is the controller:
https://www.aliexpress.com/item/32968530915.html?spm=a2g0o.productlist.0.0.4da84643bvnddH&algo_pvid=174cb56e-f627-4930-ad68-e23f0db85803&algo_expid=174cb56e-f627-4930-ad68-e23f0db85803-0&btsid=0ab6f82c15825747962494554eb362&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

It is the intergated one in the battery base,this is the axact one and from who i bought the kit from:
https://www.aliexpress.com/item/4000332284381.html?spm=a2g0o.productlist.0.0.4da84643bvnddH&algo_pvid=174cb56e-f627-4930-ad68-e23f0db85803&algo_expid=174cb56e-f627-4930-ad68-e23f0db85803-1&btsid=0ab6f82c15825747962494554eb362&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_
And my display is this one( KT LCD8H):
https://www.aliexpress.com/item/32917013157.html?spm=a2g0o.detail.1000060.1.6c0935c6wLs92G&gps-id=pcDetailBottomMoreThisSeller&scm=1007.13339.146401.0&scm_id=1007.13339.146401.0&scm-url=1007.13339.146401.0&pvid=220abb90-c40e-4459-a41f-bd8731e6f39c
 
valsam said:
The seller has 14 reviews of the same kit i bought and nobody has mentioned a simular problem.

unfortunatley that doesnt' mean it's not happening. those could all be reviews of a different batch, with different firmware or different factory settings programmed in. or they may never use their kits like you do. or they coudl even be "fake" reviews. none of those things is uncommon. :(

it could even be that the seller has deleted all the bad reviews (either themselves, or gotten them removed by whoever can do that for them). that kind of thing also happens on various sites. (for example, lyen, here on es, used to try to get any form of negative post about his stuff deleted, by the users if they could get them to, by the mods if they couldn't, whether or not it was ever resolved, as have other vendors like luna, etc).

but you could be lucky and all the reviews are real, and are the only people that have purchased the kit (that there are not others that purchased and had problems but didn't post about them for whatever reason). and just unlucky enough to get a bad part.



i tried to check the links you give, but unfortunatley google's webpage translate has changed from earlier today and will no longer translate any of the actual item information, only the extraneous info on the page that is about everything except the item itself (like ads for other items, etc), and i do not read french. :(

(google translate seems to be changed quite often, where sometimes it won't even take a url and translate the page it's for, it will only try to translate the url itself...sometimes it will fully translate the page...sometimes it will display the page with parts of it translated, and sometimes it will only show the untranslated page, or even a page that is blank except for the graphics, with no text at all. every few days to couple of weeks, it goes thru these stages, presumably as they change the software at their server end at google. it's quite frustrating).
 
The links i sent were in English .it's aliexpress and sometimes changes the language automotically to another!
Just look at the top right corner,there is the language drop down menu next to the ship to/currency drop down menu!
 
that's very bizarre. it says it's already in english / usd...but it's defnitely in french. :/ but i selected a different language entirely, then changed back to english, and now it correctly shows english words...but it expanded everything so large that a single letter was about the size of my thumb (on a 32"-ish screen). :roll: had to use the browser's zoom out feature, all the way out, to make it normal sized again. what a wierd website. :?


anyway:

i can't tell from the "specs" provided whether or not the controller and/or display might have a built-in specifciation/limit causing the problem you see, so i guess the best you can hope for atm is that the seller will swap yours out if they don't have any other ideas on settings to try (my guess is they have no clue), and that the replacement will work as expected.


the "reviews" are not useful; they can't help you determine if anyone had problems like yours or not. on the controller reviews (of which there are only 8, not 14, and only 21 have even been ordered), one of them says it works well, but the cable was too short and they ahd to get an extension, another says they haven't even tried it yet, another says "super", and the rest don't say anything at all (which doesnt' mean somethign wasn't wrong, just that they didn't post anything, positive or negative).

the display reviews don't have any useful information either, and there are only 5 of them, not 14. there have only even been 8 orders for them.

with so few orders or reviews for either part, there isn't enough of a sample size to know if there are problems with them or not.

if you have not yet tested the actual battery voltage under load while riding, with a voltmeter, to compare directly (at the same time preferably) with the meter on the display, you won't know if the display is correct or not, showing you the 42-43v just before the battery shuts off (which will also shut off the display, preventing you from reading it after that point).

if what you meant by:
)whta i am trying to say is when the screen shows 42-43v is it realy 40V when motor draws volts nad cuts off?
was instead that the display stays on, but shows you only 42-43v when the motor shuts off under load, then again you'd stil have to test that to see what voltage it really is to know if it's right....and then if that is correct, it means the battery is sagging so much it's below the controller lvc, so the controller shuts off to protect the battery.
 
Back
Top