How to use a sensorless brushless motor

gmouchawar

100 W
Joined
Mar 28, 2008
Messages
105
Hi,

I have a questions to the techies here. I purchased a brushless hub motor. Not knowing that their are 2 kinds i received a sernsorless motor.
This is a problem because there isn't (at least not that I can find) a source for cheap sensorless brushless controllers.
I have a few ideas that I will bounce around here:
1- Use brushless controller from an RC application but change the output FETS. Anyone done this? I have found some that go up to 70A and 23V.
I need a min of 36V. I am concerned that other components in the controller wont handle the higher votlage like the regulator. Also I wouldn't have
a schematic so it would be trial and error. There is another problem with the throttle input. I can test with an RC radio and receiver, but I would need
another solution when i get it to work for a throttle.
2- Add magnets and sensors to the current motor and still use a regular brushless controller with sensor inputs. 1st I thought to open the motor and see
if there is a place to put them inside. However, the seam is inside the spokes. So my current idea is to put magnet/magnets
outside the case the have a pick hall effect sensors mounted onto the fork like the pickup for a speedo.
3- find a good source for a sensorless controller. My problem is that this will still be a pedal 1st so I need a sensor near the crank also.
4- using a sensored controller as a baseline fool the sensor input??? need help with this one as I don't understand much about motors.
My understanding is that the sensors are used to determine rolling direction and at low speeds only.
5- build something to take the back EMF and convert it to sensor input for 4.
6- cut my losses and buy a whole kit for another $300.
 
Check with Justin on his pedal first controllers. They are sensorless capable controllers. http://ebikes.ca

I am pretty sure his pedal first controllers actually use the back EMF from the motor to sense it's phase relationship, so you shouldn't need a pedal sensor.
 
If you can't exchange it, I think the best thing would be to get a controller from ebikes.ca, as Pete says.

If you need a source for cheap RC HV controller: http://www.unitedhobbies.com/UNITEDHOBBIES/store/uh_viewItem.asp?idProduct=4691

You'll also need a servo tester circuit to drive the throttle function from a pot.

Not sure how well this will work with the slow speeds of a direct drive hub motor.
 
if you are technically able you could also build an amplifier to detect the back EMF and generate hall sensor like signals. the disadvantage is that the motor would have to be moving for it to start properly.

rick

it's Saturday and I'm being a smart a$$
 
As said before, the RC controllers can be had for reasonable, and a servo tester for a throttle could work.

Not much chance of modifying one, unless you have enough skill to have originaly designed it. the RC stuff is matched and computer controlled, and all surface mount. But the do build them up to 37V, soyou could go that route.

Justin's controllers are ment for Ebikes, and probably a better deal.

The problem with non pedal first and sensorless is the startup cysle. to get the motor spinning the right direction, the controller takes a guess, fires current at it, and waits to see what direction the motor spins. it then uses the back EMF off the coils to get orintation and direction before cycling for the direction and timing it needs. The problem is, on startup, the motor can run backwards untill it gets its self organized. Usualy less than 1 full turn, but that could be a bit embarasing and painfull if you were at a dead stop at a trafic light, start flirting with the girl in the car next to you, then gun the throttle on the green light and have the handlebars come back to meet your gonads at full power. :twisted:
 
most of the newer ones fire one good long pulse into 2 of the phase wires to orient the rotor in a starting configuration. once you know where you are starting from you can start the commutation sequence.

but that first pulse may roll backwards a fraction of a turn to the start position.

some of the smartest controllers measure the current during the start up pulse and use the information to set the maximum PWM duty cycle and current limit to match the controller to the motor.

rick
 
Pete said:
Check with Justin on his pedal first controllers. They are sensorless capable controllers. http://ebikes.ca

I am pretty sure his pedal first controllers actually use the back EMF from the motor to sense it's phase relationship, so you shouldn't need a pedal sensor.

From what I can tell this is your only source for pedal first controllers. http://ebikes.ca/store/#Controllers

Keep the motor and buy a pedal first controller.

-Cheers
 
I asked the same question. Justin told me that the pedal first controller will not work with a sensorless motor because they freewheel. :cry:
 
I just checked with ecrazyman. He can get a 36v/350w (pedal first) brushless controller w/ sensor kit for about $60 delivered to your door.
He just wrote to me "the sensor is not connection to motor, it is connection to controller, when the controller received 3 pulse signal, the controller will start the motor 1-3 seconds"

If gmouchawar is still interested he could try this out. It may be the solution.

-Bob
 
I wonder if this would work for my tongxin sensorless motor?- Do you have any more specifications? Do you attach the sensors to the motor? How much is it!
 
Gosh I feel like a translator! I've got crazyman (Hong Kong) on skype in one window and this thread in another.
 
Thanks for the great info. Please keep it comming. I am an EE and techincally capable of wiring something up have access to some digital logic and 1000V IGBTs.
But i admit my specialty is in HV and micro electronics and not EM.
I found a patent that describe what I want. A back EMF to hall sensor input coverter. The title is Virtual hall-effect signal generated for a brushless sensorless electrical rotary machine. It has a no. 5,598,074. I guess I am 10 year late with my idea. It issues in 1997.
The only question is whether it is cost effective to build?
What does freewheel mean? Are we now saying that the only off the shelf solution from ebike.ca would not work?
As for determining the direction of rotation, wouldn't one magnet and 2 hall sensors mounted on the fork (like the speedo pickup) solve that problem.
E.g. if the sensor 1 sees it before 2 it is cw. if sensor 2 sees it before 1 it is ccw. It would mean at least one rotation or more than one magnet.
Why would the sensored controller need 3. I am guessing one per motor phase but need to study motor more.
 
Sorry, a couple of more questions, - how many amps is this controller? What is the low voltage cut out? I am interested because I have had 2 tongxin controllers blow up. Each only lasted for one ride.
 
Very Sad,

I just asked crazy dude "Do you think that "36v/350w controller +sensor kit" will work on tongxin sensorless motor? Yes or No? Do you have pics maybe I can post to client?"

He just respond ... "no, it can not do" "motor sensor with pedal sensor is different"

gmouchawar; I am curious ... what is make and model of your motor.
 
gmouchawar said:
Thanks for the great info. Please keep it comming. I am an EE and techincally capable of wiring something up have access to some digital logic and 1000V IGBTs.
But i admit my specialty is in HV and micro electronics and not EM.
I found a patent that describe what I want. A back EMF to hall sensor input coverter. The title is Virtual hall-effect signal generated for a brushless sensorless electrical rotary machine. It has a no. 5,598,074. I guess I am 10 year late with my idea. It issues in 1997.
The only question is whether it is cost effective to build?
What does freewheel mean? Are we now saying that the only off the shelf solution from ebike.ca would not work?
As for determining the direction of rotation, wouldn't one magnet and 2 hall sensors mounted on the fork (like the speedo pickup) solve that problem.
E.g. if the sensor 1 sees it before 2 it is cw. if sensor 2 sees it before 1 it is ccw. It would mean at least one rotation or more than one magnet.
Why would the sensored controller need 3. I am guessing one per motor phase but need to study motor more.


Does this thread help http://endless-sphere.com/forums/viewtopic.php?t=340
I guess the TongXin is a geared motor and thus it does not freewheel (spin freely when coasting)
Argh! So much to know! So much to learn ... :roll:
 
gmouchawar said:
Thanks for the great info. Please keep it comming. I am an EE and techincally capable of wiring something up have access to some digital logic and 1000V IGBTs.
But i admit my specialty is in HV and micro electronics and not EM.
I found a patent that describe what I want. A back EMF to hall sensor input coverter. The title is Virtual hall-effect signal generated for a brushless sensorless electrical rotary machine. It has a no. 5,598,074. I guess I am 10 year late with my idea. It issues in 1997.
The only question is whether it is cost effective to build?
What does freewheel mean? Are we now saying that the only off the shelf solution from ebike.ca would not work?
As for determining the direction of rotation, wouldn't one magnet and 2 hall sensors mounted on the fork (like the speedo pickup) solve that problem.
E.g. if the sensor 1 sees it before 2 it is cw. if sensor 2 sees it before 1 it is ccw. It would mean at least one rotation or more than one magnet.
Why would the sensored controller need 3. I am guessing one per motor phase but need to study motor more.

Usually geared motors freewheel. - there is no resistance from the motor when you spin it.
 
Johnbear said:
I asked the same question. Justin told me that the pedal first controller will not work with a sensorless motor because they freewheel. :cry:

Now I am confused here. Brushless hub motors freewheel (even geared motors too). So what is Justin talking about?
 
and of course the obvious solution? Just add hall sensors. they can be glued to the stator with some blobs of epoxy.

the hall sensors are normally centered over a slot. that would be neutral position. since you have the motor apart anyway to glue in the sensors you can use a charged NiCad "c" or "d" cell and a compass to identify the slots.

use the battery to power up ONE of the phase windings to ground. move the magnetic compass around the stator until the "S" needle points at a slot. that would be the first slot for that phase. repeat for the other 3 phases and you know where to place you sensors. route the wires so that they don't interfere with the cover and reassemble the motor. now you are set.

rick
 
Assuming the stator has slots. If so then sure (if not then use a dremel). Go to digikey and order some hall sensors and use some cat-5 network wire for the 5 leads. Heck upgrade the entire harness and you now have a brusless hub motor with hall sensors ready to go! These are the sensors I use btw http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=480-1999-ND
 
Knuckles said:
Johnbear said:
I asked the same question. Justin told me that the pedal first controller will not work with a sensorless motor because they freewheel. :cry:

Now I am confused here. Brushless hub motors freewheel (even geared motors too). So what is Justin talking about?

They don't truely freewheel, the motor spins. With a geared motor the motor is not always spinning. :wink:
 
Thanks Rick,

That was what I tried to do. Add the sensors somewhere. But I can't open the motor without unlacing the rim.
So it would be my last move.
Can't you do the same thing external to the motor? find the correct orientation and add magnets to replace the stator and hall effect sensors.

I can't find any info on the motor. It is made in Italy by ABB servomotor. Nicely made but difficult to use.
 
Johnbear said:
Knuckles said:
Johnbear said:
I asked the same question. Justin told me that the pedal first controller will not work with a sensorless motor because they freewheel. :cry:

Now I am confused here. Brushless hub motors freewheel (even geared motors too). So what is Justin talking about?

They don't truely freewheel, the motor spins. With a geared motor the motor is not always spinning. :wink:

I see. Thus a pedal first controller becomes usless (because a geared motor is not always spinning). A pedal first controller can only function if the motor magnets are moving over the armature (spinning).

I like things simple. I am so happy I got regular old brushless hub motors (no gears - no moving parts except for the bearings and the wheel itself).

So gmouchawar ... rip that motor apart and see how it ticks! Heck you weren't busy this weekend anyway. :shock: Add hall sensors and build a new harness and buy a 72V controller and go :twisted:
 
gmouchawar,

I just had a thought on diagnostics of your motor. I assume your motor only has three wires (if not please post motor wiring). Hook up the three to a standard brushless controller phase wires (you may have to play musical wires as there are 6 possible combinations). Spin the wheel (motor mounted - bike upside down in kitchen - I work on bikes in the kitchen) and see if the controller spits out voltage between the + and - battery leads. If so then a standard pedal first controller should work. Why is this not a good idea? Anybody?
 
Back
Top