Bafang hall sensor failure - go sensorless ?

[ZapPat]

"Cogging torque" probably isn't the term to use to describe the passive drag that all permanent magnet iron-cored motors have to some extent. Justin uses the term "drag torque" in this post where he actually measures this effect: regen gains vs drag torque

I don't think most people here consider Justin "a god", but he is one of the major contributors to our ebike knowledge base and I find he is generally a cordial, honnest and informative person. I can't see why you are bashing him?

On the subject of the amount of magneticaly-induced drag torque that a motor has --> Squewed pole motors such as the old golden hubs and the xlyte 40X series hubs have reduce the amount of drag torque the motor offers (this may be why affliction can't seem to believe that they do use energy while spining freely?). The 9Cs don't have squewed electro-magnet poles and I know I can feel the drag it offers when I spin it, or try to pedal without throttle.

As for the hall sensor's timing having anything to do with regen, this is not so, affliction. Regen is done using PWM in the controller, and is not linked to the hall-based commutation events. I believe that there may be some confusion here between the passive drag resistance (or torque) all iron cored BLDC motors have and the torque ripple they produce when fed a simple square voltage waveform by the controller. The torque ripple amount/nature *can* be affected by hall timing because it has to do with commutation events, but the drag torque always exists behind the scene because of magnetic losses. Justin posted an example of what kind of torque ripple we have with a 9C hub being fed regular 6-step square waves.

BTW affliction, where did you get that last quote talking about "cogging torque"?

 

[Russell]

Russel, I wasn't referring to starting the motor! I was referring to stopping the motor. My Crystalyte and 9C are generally smooth on takeoff, excepting the occasionally "hammering" on the Aotema controller I use on the Crystalyte. The roughness/"push back" of the motors I only experience on a rapid drop to a no current throttle state when I am near the no load speed. There is noticeable continuous unvariable resistance from the motor - a general roughness too. To me the hammering is some sort of issue with synchronization only felt on startup.

I have not noticed that phenomenon with the ecitypower sensorless controller on my locked-up Bafang. Unless I forget to roll forward from a complete stop the controller running sensorless is indistinguishable from running sensored.

-R

 

[Russell]

Heh Heh, yer both right , and both wrong some. But it's entertaining 8)

Some dd motors cog a lot , some hardly at all. This is with no controller connected. How many different kinds of motor have you owned affliction?

If all I ever owned was my aotema, I'd think cogging was a myth too. With my 5304, it's no myth. Even within a specific motor type, it varies. I do belive Justin has seen a few more motors than most folks.

Dogman,

I don't know who you are referring to as "both wrong" because I clearly stated that the degree of cogging torque varies from motor to motor however it is something that is always there with the direct drive motors we have now. Anyway a no-load spin-down test from a set speed is something anyone can do to find the comparative drag from various motor combinations since the time it takes the wheel to stop is a good indication of the amount of total drag from the magnets, bearings, etc. I don't have my 9C in a rim now so I can't compare it to my locked-up Bafang but since you have three front motors it would be interesting if you could run each one to 20 mph then time how long they take to stop. The Fusin should win of course but it would be interesting to see how the Aetoma and 9C compare.


-R

 

[Russell]

"Cogging torque" probably isn't the term to use to describe the passive drag that all permanent magnet iron-cored motors have to some extent.

Perhaps it's not the best term to use however it is a term which is recognized by the majority of the e-bike community so when somebody says "cogging torque" most folks know what the person is talking about. If they don't know the term generally all it takes is saying "hey you know how nuch harder it is to spin your direct drive motor wheel now compared to your original bike wheel...well that's cogging torque in action" and the person will say "oh so that's what it's called" and that'll be that. Perhaps there's a better term to use which we can all agree on but it will take some effort to spread the word.

-R

 

[chvidgov.bc.ca]

I agree with you Russel as far as the general lingo goes. It just seems that this pushback is worse on sensorless controllers, at least the ones I've tried, compared to the Hall types, when decelerating rapidly from a fast speed, and some roughness is felt. Perhaps some controllers continue to put a bit of juice into the motor when the throttle is disengaged, and the resistance is not noted so much.

 

[ZapPat]

I agree with you Russel as far as the general lingo goes. It just seems that this pushback is worse on sensorless controllers, at least the ones I've tried, compared to the Hall types, when decelerating rapidly from a fast speed, and some roughness is felt. Perhaps some controllers continue to put a bit of juice into the motor when the throttle is disengaged, and the resistance is not noted so much.

I've heard mention of this before somewhere on this board concerning a sensorless controller, but I would guess that not all sensorless controllers do this because they don't really need to. I wonder - do you have a CA to see if it is giving you a slight regen current when it does this?

 

[Russell]

Yes there is resistance from my Bafang with a locked up clutch but as I remember it IS LESS than the cogging torque from a 9C motor. I don't have my 9C laced to a rim right now so I can't compare it against the locked Bafang however someone else can if they like by running one up to 20 mph (no load) then release the throttle and time how long it takes to stop spinning. My locked up Bafang takes 8 seconds to stop from 20 mph where it used to take about 50. On the road since I use the throttle 90% of the time there was only an opportunity to freewheel 10% of the time anyway so the ability for the wheel to freewheel was only a marginal advantage at best.

-R

I just checked how long it takes the freewheeling GM Mini Motor (really a Bafang QSWXH) on my Raleigh to spin down from 20 mph and it takes 27 seconds, not nearly as good as my Bafang QSWXB did when it was freewheeling. Cheaper clutch or bearings????

-R

 

[ZapPat]

Yes there is resistance from my Bafang with a locked up clutch but as I remember it IS LESS than the cogging torque from a 9C motor. I don't have my 9C laced to a rim right now so I can't compare it against the locked Bafang however someone else can if they like by running one up to 20 mph (no load) then release the throttle and time how long it takes to stop spinning. My locked up Bafang takes 8 seconds to stop from 20 mph where it used to take about 50. On the road since I use the throttle 90% of the time there was only an opportunity to freewheel 10% of the time anyway so the ability for the wheel to freewheel was only a marginal advantage at best.

-R

I just checked how long it takes the freewheeling GM Mini Motor (really a Bafang QSWXH) on my Raleigh to spin down from 20 mph and it takes 27 seconds, not nearly as good as my Bafang QSWXB did when it was freewheeling. Cheaper clutch or bearings????

-R

The spin-down time from a certain speed you are measuring depends also on the mass of the spokes+rim+tube+tire, are they the same for each hum motor tested?

I am surprised that your geared hub with glued freewheel offers so little resistance compared to a 9C... I would like to try something like this with some regen.

How do you find that mini golden motor, Russell? I'm thinking of getting one although shipping is crazy expensive.

 

[Russell]

The spin-down time from a certain speed you are measuring depends also on the mass of the spokes+rim+tube+tire, are they the same for each hum motor tested?

I am surprised that your geared hub with glued freewheel offers so little resistance compared to a 9C... I would like to try something like this with some regen.

How do you find that mini golden motor, Russell? I'm thinking of getting one although shipping is crazy expensive.

The spin-down tests are from 20 mph because that's how I originally tested my Bafang motor here;

http://endless-sphere.com/forums/viewtopic.php?f=3&t=7918#p120924

The person who began that thread was complaining that her freewheeling motor spun down in 30 seconds. Mine was 50 seconds from 20 mph when new and that improved to 70 seconds after some break-in.

The rim/tire/tube/spokes are all the same from that original test to the latest test with the GM Mini since I went from the Bafang to a 9C then to a GM Mini on the Raleigh. The Bafang is now on my Kona with a larger tire and tube. My regular bike cone and loose bearings run from 2 to 3 minutes from a hand spin and I have one wheel with sealed bearings that I swear took almost 10 minutes to settle out (I thought I was seeing things ) so while weight may have some impact on how long a wheel spins the overriding factor is the bearing friction or with motors the cogging torque plus in my case the meshing of the gears.

I can't check the spin-down time of my 9C but what I just did was to spin the axle of the 9C and the axle of an extra GM Mini motor (gears and all) with my fingers and the effort is comparable so I would expect the spin-down times to be close.

The GM Mini Motor is just a Bafang QSWXH however it is a newer design than my original Bafang QSWXB with a 1cm small diameter case, a disc mount and it tips the scale at over a pound less. Newer design imo however doesn't mean better and when I lost the hall sensor on my original Bafang after 3,000 miles I replaced the core with a Mini Motor core and promptly melted the internal hall wires on the first ride. I was however pushing the Mini to over 1,000W when I usually kept the Bafang to under 760W. I haven't had a problem with the GM Mini on my Raleigh however I haven't racked up many miles on it yet. The halls and hall wiring on the Bafang and a GM Mini have failed me twice so removing the whole ball of wax from my original Bafang and running it sensorless was a plus in my book.

At this time I would not recommend buying a front GM Mini Motor due to yet another redesign that isn't working out for one member of this forum;

http://endless-sphere.com/forums/viewtopic.php?f=7&t=14835

-R

 

[DAND214]

3 PAGES LATER!
All I want to know is did Mir decide what to do! Fix it, replace it or go sensorless. Don't confuse the guy try to help him. That's what we are here for, aren't we?
Dan

 

[Russell]

3 PAGES LATER!
All I want to know is did Mir decide what to do! Fix it, replace it or go sensorless. Don't confuse the guy try to help him. That's what we are here for, aren't we?
Dan

He knew with his first post what his options were; replace the hall sensor, go sensorless or replace the motor, so unless someone has a step by step procedure on how to replace a hall sensor in a Bafang motor it's up to him now. We can still jabber until he comes back. :wink:

-R

 

[johnrobholmes]

As I understand it:

Cogging: Used in the RC world for a bad sensorless start or operation. Turnigy controllers cog on a bike, they don't keep sync.

Cogging torque, drag torque: The magnet to steel attraction inside the motor. If the motor is not built with a skew between the steel and magnet angles, it can have a noticeable resistance to turning that is best described as bumpy.


I am a big fan of sensorless controllers, but also love the immediate and predictable start of sensored. I would try to repair my bafang if the sensors gave out.

 

[DAND214]

Yeah Russell, it's all in fun and I guess that's what it's all about. We can argue what works for me/us and throw it back and forth till one of us gets off there butt and do something. All I know is if we can put a bike together we should be able to fix it. We can take it apart and fix it or some just throw more money at it till it's fixed. I prefer to fix it than to replace it if possible.
As for MIR's problem it sounds like a broken wire as I had on more than one hub brand new. I really enjoy riding but I also get great reward out of fixing it also.
I sure hope we didn't chase him away!
Dan

 

[mer]

Hey guys

thanx for all the answers.. Well I like pedaling a lot so locking the freewheel is not reallu an option for me ... I know the drag wouldn't be that much of a problem.. but I would prefer replace the hall sensors I guess and if I went sensorless I think I would go kinda mad if I was hearing the hummer sound almost everytime I applied some throttle. Can I replace them without having to re-spoke the hub to the wheel ? Respoking the hub to the wheel is a pain in the *ss.. Where to buy the hall sensors ? I know there is a step by step instruction how to replace the hall sensors ( by Knuckles I think ) .. Does maybe anyone know of a sensorless controller that would start the motor from a dead stop without the hummer sound ? ( maybe the golden motor magic controller ?? )

I like the discussion about cogging... don't worry.

 

[dogman dan]

Regarding a previous question, would the Fusin gearmotor work well with a Wilderness Energy type sensorless controller?

No, the thing has a hard time starting. If you really ease into the throttle, then the motor will start to turn, and run ok, but it's a PITA. Most of the time, the motor will just chattter instead of running. It's because of the freewheel inside, and the controller needs just a tiny bit of motion forward to get started. The controller is not pedal first, where you need to get to 5 mph, but it does like a bit of forward motion to get going.

If you locked the freewheel( or is it a clutch?) in the motor though, it should work fine.

Btw Russel, maybe I should have said, both failing to communicate. I'm too ignorant myself to call you wrong some, I apologise.

 

[Russell]

Dogman,

No need to apologize I just wanted to be clear about what I said.

I'm glad you got a chance to try the WE sensorless controller on the Fusin geared motor, I was tempted to buy one and try it for myself so you saved me some money. I'm not surprised you found it unacceptable as I did with my ecitypower dual-mode controller on my GM Mini and Bafang geared motors. Still there is no clear picture if there are ANY sensorless controllers that work well with freewheeling geared motors, some people say one brand works well then someone else comes along with a contrary opinion. My own feeling is that it would be next to impossible to have a freewheeling motor run well with a sensorless controller for the reason you stated which is why I locked the clutch on my Bafang (It's called a clutch on the Bafang web site but I use "clutch" and "freewheel" interchangeably) but if that's so why is Bafang manufacturing sensorless freewheeling geared motors?

-R

 

[johnrobholmes]

I had no problem with the infineon controllers, they will jump start the motor no problems. I also had no problems with Castle controllers getting it started smoothly once rolling. Just have to be careful not to jam open the throttle when you are coasting too fast, or it will hammer the freewheel very hard.

 

[DAND214]

Yes MER, you can take it apart without unlacing the wheel. take the cover off and take the guts out and you will find the broken wire or bad sensor as you do the tests.
As I said I had a broken wire in the axle and it cut out as I moved the cable, try that before you take it apart to see if it's a wire or a sensor. Good luck, it's lots of fun.
Dan

 

[dnmun]

the hall sensor wire can be broken, yet the hall sensor ok. what kills the hall sensor is when it toggles on to sink the current in the hall sensor line from the controller and is shorted out to the phase wire at the same time. this happens when you get the motor twisting itself out of the forks and the phase wire is cut and shorts to the hall sensor lead cut at the same time. but unless you did that or had lotsa water in the motor, i doubt if the hall sensor really is dead. but i'm not the one working on it either. kinda like i said up front.