Adding hall sensors to outrunners

I mounted the sensors in the way I suggested in the previous post.
Was very pleased that it started up using an 18 fet Lyen esc (movie).
But I ran into a problem at higher rpms (no load). At about 70-80% throttle the motor stops.
When I'm carefull with the throttle I can get it passed this point but then the motor sound changes a bit and the phase wires start to heat up.
So I think there is a timing issue? Could it be that I hit the rpm limit of this controller? Can't seem to find any information about max rpm.
The motor has 12 pole pairs and should run 6000 rpm at 40V, so 72000 erpm.
The signals coming from the hall sensors still look good on the scope when this happens.
 
the controller should be good upto at least 10k. Yes it do sound like you have a timing issue, one thing I can suggest is to shift the hall sensors by 1 slot and see if this helps: ref threads : controller cutting out: http://endless-sphere.com/forums/viewtopic.php?f=2&t=16910
hall placement inside motor: http://endless-sphere.com/forums/viewtopic.php?f=30&t=15686&start=210#p358949
 
Thanks for the information!
It looks like delta config needs another approuch.
I took the following approuch:
The hall sensors are spaced 120 degrees. But I did not look at which phase they where placed.
The phase wires are also not marked.
So I just hooked one scope channel up to a phase wire with ground to another.
Then by manually cranking the motor in the direction I want it to turn I found a matching hall sensor.
I called this first phase wire and hall sensor A. I continued this for the other phases.
They all seemed to match. I don't know if this is the right method?
 
Did some further testing. The no load current seems to be too high. It ramps up to 20A close to the point when the esc turns off.
I made a scope image of one of the halls and corresponding phase when turning it by hand (all three look like this):
Rotomax%20100CC%20Hall%20Test1.jpg

The phase of the hall sensors look to be a bit behind the zero crossing of the phase voltage.
Could this be the problem?
 
NickB said:
Did some further testing. The no load current seems to be too high. It ramps up to 20A close to the point when the esc turns off.
I made a scope image of one of the halls and corresponding phase when turning it by hand (all three look like this):
Rotomax%20100CC%20Hall%20Test1.jpg

The phase of the hall sensors look to be a bit behind the zero crossing of the phase voltage.
Could this be the problem?

yes this can be the problem, did you try moving the the hall sensors by 1 slot ?
 
gwhy! said:
yes this can be the problem, did you try moving the the hall sensors by 1 slot ?
I'm going to try that tomorrow with other sensors. The current hall sensors are already epoxied.
Thought it was ok if it ran :).
Thanks for the help!
 
NickB said:
gwhy! said:
yes this can be the problem, did you try moving the the hall sensors by 1 slot ?
I'm going to try that tomorrow with other sensors. The current hall sensors are already epoxied.
Thought it was ok if it ran :).
Thanks for the help!

Its worth a try, I have always used external hall sensors up to now so have never encountered these sort of problems. hopefully the move one slot will work out..
 
Scope image with the sensor shifted one slot:
Rotomax%20100CC%20Hall%20Test2.jpg

The problem seems to be shifted to the other side.
Seems like neutral timing is between these two.
 
when burtie made the small video of the sensors in the different slots this did not effect the running of the motor that much with just the 30 degrees out so there may be also something else going on, I just relised that the motor has 24 slots so how many magnets has it got ? Im am wondering now if you have come to the max erpm of the controller, can anyone confirm that these 24 slot motors run on 40+v with sensor using a crazyman controller ? . yes the current will be a bit higher than it should be initially but ( and this is what can be causing the controller to cut out ) it should not get to a point in the rpm that the current then shots through the roof and things start getting very hot very quickly. It may be a good idea to rig up some external hall setup that you can adjust just to test unless someone chips in about the max erpm limit. The motors that I use are 12 slot 14 mags motors and I have run these upto and slightly beyond 10000 rpm without problems (72000erpm) wondering if the erpm limit on these controllers is around 70k'ish now wouldnt your motor be nearlly 2 x the erpm of a my motors :? what about testing your motor on 24v to see if it still cuts out and misfires at full throttle?
 
The motor has 20 Magnets.
I think I'm going to test it with the sensors shifted one slot.
The movie shows the timing advanced by 30 degrees.
I think the timing of the sensors now lags behind the neutral timing.
With the sensors shifted one slot the timing seems to be advanced by a small amount (15 degrees?).
Will test it on 24V first.
 
NickB said:
Scope image with the sensor shifted one slot:
Rotomax%20100CC%20Hall%20Test2.jpg

The problem seems to be shifted to the other side.
Seems like neutral timing is between these two.
If I have to guess from my computer screen this looks like about 6 deg advanced. Which might be perfact for a little higher rpm.
 
Did some further testing. The hall setup advanced one stator slot didnt seem to work.
So I did another test with the original hall setup.
But now I noticed that when the current started to rise, the rpm stayed the same.
I could advance the throttle and watch the current go up, but the rpm stayed the same.
I think it might be the controller being at its rpm limit.
Movie (turn the volume down a bit :), its a bit noisy without the rear bearing)
 
a couple thoughts & observations.

this is a standard lyen 18fet controller? no shunt modifications or hacker voodoo performed on it?

if so, I can say it will do better than 70k e-rpm....I have spun 7 pole pr, 80mm motors to 14k rpm with the same board & fet's (not recomended btw)

I will recomend the C20 mod for anyone atempting to run an high rpm motor with a xiechang based controller...I have had the throttle "cut out" issues on every style controller I have tested with...6fet,12fet & 18fet. it works on all of them.

Solder a 1k resitor across C20 & find the ballance between phase & battery program limits.(sounds easyer than it is!)

Have you re-terminated for wye? if so, I would recomend finding your phase pr of teeth & centering your hall between them.

its a 24tooth stator? if so AabBCcaABbcCAabBCcaAabBCc (drlk & you have 10 pole prs in the can?)

my 1st guess is you are on a fasle hall combination. There are a couple that will run a motor....just not very well.
in your video I can hear the controllers high frequency ringing in protest...try a few more hall combo's before getting too disgusted.

last try to sheild your hall wires from your power wires if posible... I braid my phase wires to help counteract traveling inductance...& that has worked on a few bench set ups that were giving me fits.

good luck & keep trying....the answer will present itself soon.
 
It's indeed a standard Lyen without any modifications. Got one from a forum member for testing.
The motor does 167kv according to HK. The batteries are 45V so thats about 90k erpm?
I will look into the C20 mod. But the controller seems to cut out at an alarmingly large no load current.
Below this current it seems to behave normal without any cutouts.
The motor is delta terminated. I read that with delta you need to place the sensors in between the same phase wires which are wound in the same direction?
The stator poles are numbered (photo). I have put the sensors in between same numbered poles.
The phase wires coming out of the motor are also not marked. So just scoped one of the hall sensors and found the matching pole wires and called them both A.
I did the same thing for the other poles. Don't know certain if this is the right method?

I just tried another hall setup. This time I again put them in between poles with the same number but these where shifted 60 degrees from the first hall setup.
The same problem occured. Maybe next time I should try putting them in between poles with different numbers.
But thats something for another day :)

The ringing could be a problem with the sampling frequency of the camera mic. Can only hear it in the movie.

Thanks a lot for the help!
 
Arlo1 said:
NickB said:
Scope image with the sensor shifted one slot:
Rotomax%20100CC%20Hall%20Test2.jpg

The problem seems to be shifted to the other side.
Seems like neutral timing is between these two.
If I have to guess from my computer screen this looks like about 6 deg advanced. Which might be perfact for a little higher rpm.

looks more like 30 degrees to me.

maybe you have access to a rc sensorless esc this will give you the actual no load current at WOT, :?
 
gwhy! said:
Arlo1 said:
NickB said:
Scope image with the sensor shifted one slot:
Rotomax%20100CC%20Hall%20Test2.jpg

The problem seems to be shifted to the other side.
Seems like neutral timing is between these two.
If I have to guess from my computer screen this looks like about 6 deg advanced. Which might be perfact for a little higher rpm.

looks more like 30 degrees to me.

maybe you have access to a rc sensorless esc this will give you the actual no load current at WOT, :?
Looking again it looks like 15 e deg. From positive peak to peak is 360 e deg then each big square is 90 e deg and there is 6 dots across a square...=15 e deg
Post edited to clarify e-deg
 
Tried another hall location, even tilted the sensors to get the timing as good as possible but the problem stays the same.
At a certain point the rpm stops rising and I can controll the extra no load current with the throttle.
The no load current before the problem arises looks normal (about 4A).
Doesnt the Back EMF of a phase have a period of 120 electrical degrees?
 
a pair of magnets represents 360 e-degrees so each of the hall sensors is switching state every 180 e-degrees. Each back emf full cycle is 360 e-degrees.

Do you know the speed it gets to before the current starts going mad ?
 
gwhy! said:
a pair of magnets represents 360 e-degrees so each of the hall sensors is switching state every 180 e-degrees. Each back emf full cycle is 360 e-degrees.

Do you know the speed it gets to before the current starts going mad ?
I edited my post to say E-deg
 
Just tested the rpm. The hall freq stops at 1.07kHz.
So that gives 1070*60/10 = 6420rpm.
This equals to 77k erpm?
 
if Thud have run his motors upto 14k rpm thats around 98k e-rpm so the controller should be able to spin your motor without problem.. the only thing I can suggest is to try different slots and hall/phase combinations .. as Thud have said ( It can be upto a 2 slot difference) or try a external hall setup to test with.
 
Scope all 3 sensors and label the phase wires and sensor wires. It's almost like the timing for 1 sensor is wrong . It is different for the third sensor on WYE vs Delta on almost all motors.
 
I think its solved now.
Only don't have an explanation for it :).
Was trying another sequence with my epoxied in hall sensors:
Controller A - Motor B
Controller B - Motor A
Controller C - Motor C
And the same for the hall sensors. When i tried it it didnt start up.
So I tried different combinations with the phase wires only.
With one combination it started up and had noticably more torque.
When I went to full throttle the problem was gone. It runs at full rpm now with normal no load current.
The combination of the phase wires is:
Controller A - Motor C
Controller B - Motor B
Controller C - Motor A
And the combination of the hall sensors is:
Controller A - Hall B
Controller B - Hall A
Controller C - Hall C
Anyone have an explanation for this? :D
Movie
 
you had a false positive.
& now you have a good set up.
 
Back
Top