How to find number of poles?

[NoSmoke]

Hello, I have a BLDC motor that I am trying to configure the controller for. I need to know the number of poles in the motor but the motor itself has no useful markings on it. I understand one can find the number of poles by counting the number of cogs the motor does when rotated by hand through one revolution. In my case the number I count is 60.

Does anyone know how to convert that to number of poles?

TIA

 

[amberwolf]

Depends on how the controller defines poles. Some of them actually use the number of magnets. Some of them use the number of poles, which is usually half the number of magnets (typically the magnets reverse polarity every other one).

If the motor has the common UVW hall sensors (three phase, often yellow, blue, green for signals) then you can monitor one of those for how many pulses you get out of it per rotation; this should be the number of poles as many (most?) of the common motor UVW hall sensors only switch states when the magnetic field reverses.

 

[SerjeoWN]

easy peasy
[youtube]https://www.youtube.com/watch?v=Coi4JO5aLCk&t=55s[/youtube]
you have to count how many pulses per one revolution

 

[NoSmoke]

Thanks to you both for replying.

amberwolf, if I understand this correctly, number of poles = number of pulses (state switches) from one of the Hall sensors per rotation.

SerjeoWN, from the YouTube it appears, number of poles = 2 * number of pulses (voltage) from one phase per revolution. I'm assuming the Honda motor shown there is a 3 phase motor.

Is that right or am I missing something?

 

[amberwolf]

amberwolf, if I understand this correctly, number of poles = number of pulses (state switches) from one of the Hall sensors per rotation.

If the hall sensors used in your motor are the common latching type, then yes.

If they are the uncommon non-latching type, then they may change state everytime a magnet passes them, regardless of which polarity it is, so you get twice as many state changes.

If I were setting up a motor based on the sensor pulses, I'd first try the number of pulses directly, as the number of poles. If that didnt' result in expected system behavior, I'd divide that in half and retest.

 

[NoSmoke]

Hey amberwolf, I hooked up the Hall sensors to +5V and measured 10 state changes (0v to 3V to 0 to 3 to 0 to 3 to 0 to 3 to 0 to 3 to 0) in one rotation. Sounds like 10 poles as it could not be 10/2 = 5 which is not an even number(?)

I think that also might tie in to the 60 cogs per rotation (60 cogs/3 phases/2 = 10)??

Thanks much for the suggestion - do the results sound plausible to you?

 

[amberwolf]

Hey amberwolf, I hooked up the Hall sensors to +5V and measured 10 state changes (0v to 3V to 0 to 3 to 0 to 3 to 0 to 3 to 0 to 3 to 0) in one rotation. Sounds like 10 poles as it could not be 10/2 = 5 which is not an even number(?)

I think that also might tie in to the 60 cogs per rotation (60 cogs/3 phases/2 = 10)??

Thanks much for the suggestion - do the results sound plausible to you?

I don't know how number of cogs would calculate out like that, since it doesn't cog by phase, it cogs by magnet passing stator tooth (even if you had no windings on it); there are points in rotation where some magnets line up with some stator teeth so that the attraction is highest; this is where the "cog" feel comes from.

(if you short the phase wires together, the cogging is stronger because now there is an induced current in a coil loop creating a field opposing or attracting those magnets aligned with them. )

Regarding odd/even, pole count for many common DD hubs is 23 poles (46 magnets).

I'd just try it with the direct hall measurement, then if that doesn't fit expected behavior, try it with half of that.

 

[NoSmoke]

OK, I guess my cogging arithmetic was a bit of a stretch...

But I am confused as to how 46 magnets have 23 poles or is that 23 pole pairs (thought BLDC motors always had an even number of poles)?

And, does DD mean Direct drive?

OK, I'll start with 10 and see if the angle routine passes.

 

[amberwolf]

OK, I guess my cogging arithmetic was a bit of a stretch...

It may well be "correct" (don't know)...but it shouldnt' be necessary.

But I am confused as to how 46 magnets have 23 poles or is that 23 pole pairs (thought BLDC motors always had an even number of poles)?

To clarify, it is 23 pole pairs...but that is the number usually referred to when a system asks for the number of poles (in my experiences so far). (sometimes they actually ask for pole pairs)

Technically, to be precise, if your system is asking for the number of poles, you would put the actual number of magnets that are in the motor.

If your system is asking for the number of pole pairs, you put the number of pairs, which typically equals the number of hall pulses for any of the three UVW hall sensors.

But that is only for systems that have accurate terminology, which is not necessarily the case.

And, does DD mean Direct drive?

Yes. The "9c", etc., are examples of 23-pole(pair) motors like that.

OK, I'll start with 10 and see of the angle routine passes.

It's worth a shot. It shouldn't hurt anything if it's wrong, it just wont' work as expected.

 

[NoSmoke]

It may well be "correct" (don't know)...but it shouldnt' be necessary.

I think I am finally getting that - I've been trying to relate cogs to poles and it may well be a futile exercise as there are two unknowns - number of poles and number of slots in the stator(?). Anyhow, thanks for pointing out it is not necessary.

But I am confused as to how 46 magnets have 23 poles or is that 23 pole pairs (thought BLDC motors always had an even number of poles)?

To clarify, it is 23 pole pairs...but that is the number usually referred to when a system asks for the number of poles (in my experiences so far). (sometimes they actually ask for pole pairs)

OK, understood. According to Kelly, they want the number of poles in their program.

Technically, to be precise, if your system is asking for the number of poles, you would put the actual number of magnets that are in the motor.

Yeah, that sounds right - if I read 5 state changes, it means 10 magnets as a Hall device typically only responds to a south pole as I understand it.

If your system is asking for the number of pole pairs, you put the number of pairs, which typically equals the number of hall pulses for any of the three UVW hall sensors.

Understood.

But that is only for systems that have accurate terminology, which is not necessarily the case.

I hear that - the Kelly user manual is quite voluminous but also hard to decipher in many places. For example they refer to "real number of poles". I had to ask for clarification as to what they meant by that (it means number of poles, not pole pairs).

OK, I'll start with 10 and see of the angle routine passes.

It's worth a shot. It shouldn't hurt anything if it's wrong, it just wont' work as expected.

Yeah, that's what Kelly says - try it and adjust till it works... (if I don't smoke anything in the process )

 

[amberwolf]

Hall device typically only responds to a south pole as I understand it.

The typical latching hall used in UVW-sensored motors respond to the change in polarity, not just one specific polarity. There could be some that do just respond to one polarity, but I haven't had one in my hands. An example out of the Allegro Micro documentation:

https://www.allegromicro.com/-/media/files/application-notes/an296067-latching-switch-hall-effect-ic-basics.ashx

Latching Hall-effect sensor ICs, often referred to as “latches,” are digital output Hall-effect switches that latch output states. Latches are similar to bipolar switches, having a positive BOP and negative BRP , but provide tight control over switching behavior. Latches require both positive and negative magnetic fields to operate. A magnet presenting a south polarity (positive) magnetic field of sufficient strength (magnetic flux density) will cause the device to switch to its on state. When the device is turned-on it latches the state and remains turned-on, even if the magnetic field is removed, until a north polarity (negative) magnetic field of sufficient strength is presented. When the negative field is presented, the device is turned-off. It latches the changed state and remains turned-off, even if the magnetic field is removed, until a south polarity (positive) magnetic field of sufficient strength is again presented.

Not all motors use this kind of digital-output hall sensor, some of them use a type that responds to any field with an "on" (grounding the output). These can be more sensitive to magnetic "noise", interference from the changing stator phase fields, etc., especially in motors being used at high currents; I suspect they may be part of the problem some people have with hgh-power bikes that use that type of sensor in the motor (I don't know any specific ones, but it would be worse with ones like Revolt that may bury the sensors deeper in the laminations rather than having them right up at the surface near the rotor magnets like most hubmotors do.

So your motor may not use latching types...but it's a fair bet that it does.

Yeah, that's what Kelly says - try it and adjust till it works... (if I don't smoke anything in the process )

Unless the controller is defective or broken, it should be smart enough to protect things. I have a broken Kelly here that is not, that began smoking a motor (perhaps only the external phase wires; never got around to opening it up to check) during the setup while connected to the program via serial, even with current/etc turned down to minimums on all fields. But I already knew it was probably broken when I got it (had just hoped I might fix it; was really cheap).

 

[NoSmoke]

Thanks amberwolf for the interesting info.

Was notified yesterday my controller has shipped and should be here Monday so we will see what happens.

BTW, took a look at the Kelly forum - surprisingly little activity there and a lot of queries with no reply. Also tried to sign up but could find no way to do that on their web site.