How many Hall / Phase Combinsations are there......?

esust

1 mW
Joined
Jan 28, 2018
Messages
18
Does anyone know the definite answer to the number of Hall / phase combinations there are.

I have come across so many different answers online. Some people say there is 1 correct combi out of 36, others say 1 in 6, I have seen 3 in 36 with 3 reverse in 36. This is driving me nuts how many are there??

The way I see it there is only 1 correct combination in 6. Surely you only need to have the correct hall sensor connected to the correct phase on the motor. But i don't see why it matters how the halls are connected to the controller as long the correct phase wire is connected to that hall sensor.

so it if hall C is connected to controller input 2 then when hall C is activated it will activate which ever phase is connected to controller output 2 (which needs to be phase C wire). Then the motor turns clockwise a bit and activates Hall A. When Hall A is connected to controller input 3 then the phase wire connected to controller output 3 will be activated, which needs to be Phase A wire.

With this in mind it doesn't matter which hall sensor is connected to the controller hall input number as long as the corresponding phase is connected to the same controller output number. so both combinations in the picture below should work right? and there are a further 4 which will also work. Making it 6 correct combinations out 36 therefore 1 out of 6.

Pic 1.JPG



Unless I'm getting this wrong and Hall A has to be connected to input 1, Hall B has to be connected to input 2 and Hall C has to be connected to input 3 then then it will be 1 in 36.

Can someone help here!!
 
What about direction@ Out of the 6, three forward and three reverse.

All that only works if the timing is balanced. I've had motors where no combinations worked with KT controllers and the same motor would only run in reverse with self-learning controllers no matter what combination was connected.

Also, out of the three combinations that will give correct forward motion when the timing is balanced, one of the three sometimes runs better than the others.
 
I suspect those are caused by variations in hall sensor placement within the motor, and tolerances of the controller in how it reads those signals and sends the phase currents. Another factor may even be how noisy the signal lines are, and how well a controller rejects that noise.

I expect that if all hall sensors are placed optimally in every motor, then it probably wouldn't matter which of the various "correct" combos is used with what controller, they'd all work equally well.

Would be interesting to test.
 
There are 36 possible combinations. 6 are valid, 3 forward and 3 reverse. For every hall combination there is one valid phase combination, and for every phase combination there is one valid hall combo. This makes things easy. Just keep either the halls or the phases static, and try all 6 combos of the other. If the valid combo is reverse rotation, then just swap 2 (any 2 and only 2) of the set you were keeping static and find the correct combo of the set you were originally swapping. Be careful of "false positive" results, since on many motors some invalid combos can spin the wheel. You want the one with the smoothest most quiet start. On all the motors I've wired up, false positives have always spun the opposite direction from the valid combo, so when I get a false positive with the fwd rotation, I swap a pair of my static set then. The best way to confirm a valid combo is by measuring a low no-load current at WOT.

hints:
Use a small jumper wire to complete the battery connection to make sure invalid combos can only result in low current, so you can't blow anything.
Swap only 2 wires at a time, not all 3. It's the easy way to be systematic in trying 6 combos.
Use small throttle pulses until you're pretty sure you have a correct valid combo.
Take the chain off, so a reverse rotation doesn't cause havoc with the pedals, since the freewheel doesn't work in reverse.


The above method is the easiest most efficient way to wire up a controller that doesn't have a self learning function. Flow charts, spreadsheets, and other nonsense are completely unnecessary. I regularly wire up 6 phase motors, which require dual controllers and always use this method. A 3 phase motor and controller I could do blindfolded, and have on more than one occasion wired up motors and controllers that have no color coded wires.
 
d8veh said:
What about direction@ Out of the 6, three forward and three reverse.

All that only works if the timing is balanced. I've had motors where no combinations worked with KT controllers and the same motor would only run in reverse with self-learning controllers no matter what combination was connected.

Also, out of the three combinations that will give correct forward motion when the timing is balanced, one of the three sometimes runs better than the others.

The old MAC/BMC motors had non-neutral hall sensor positions that would give all sorts of problems when trying to mate to a normal controller. The trick with these was to get the lowest no-load, full throttle current. Ideally a motor like this needs a controller designed for the timing shift or use a sensorless controller.
 
John in CR said:
There are 36 possible combinations. 6 are valid, 3 forward and 3 reverse.

so the valid 3 forward combinations would make it 3 out of 36 = 1 out of 12. but you say below it's 1 out of 6?

John in CR said:
For every hall combination there is one valid phase combination, and for every phase combination there is one valid hall combo.

The above makes perfect sense to me. one valid forward combination for every 6 and one valid backwards combination for every 6.That's totally what i thought. you can have any combination of phase or hall wires in how they are connected to the controller. you just need to change the either the phase or hall wires and leave others alone. There are 6 combinations with 1 forwards and 1 backwards.

Forwards correct combo
Hall A - Phase A
Hall B - Phase B
Hall C - Phase C

Backwards correct combo
Hall A - Phase C
Hall B - Phase A
Hall C - Phase B
 
John always explains this one best. Then why, have I had so much trouble a few times. No reason you should not get the right forward combo in 5 min. At times, Its taken me hours, over a period of days to find it.

What's the deal here? Its bad connections. When you are trying them out, you can appear to have great connections, but you don't. And you pass by the right combo, because one of the 6 critical contacts is not good enough to run smooth. Or maybe all the halls perform shitty, because of a bad connection on the red or black halls wires.

Then there is dumb luck. In I don't know how many times, I always find reverse first. Always!, And then half the time, swapping two gets me nothing.

This taught me two things.

One, make more sure those contacts make good connection. If they pull apart easy, snug them up. (referring to the typically unreliable square white scooter type plugs.) And have a good hard look at those phase plugs too. One loose one, and you try all 36 combos with no joy.

Two, buy controllers that self learn. So easy with self learn.
 
Dogman,

You're just not being systematic, probably swapping more than 2 wires at a time, and/or not keeping either the halls or phases static. To be systematic, I choose one anchor wire of the set I'm swapping, eg motor phase green. So starting with color on color, if that doesn't work, swap yellow and blue, so you've tried both possible with green on green. Next swap motor green and one of the other two. Now swap motor blue and yellow again, so you've tried motor green on both possible with green on that color. Next swap is with motor green to the remaining color you haven't tried, and next swap yellow and blue. That's 6 of 6. If at any point you get a valid reverse, stop and swap 2 and only 2 of the wires you were keeping static and your good forward will be one of the five remaining of the set you were swapping (it's not the one that was just a reverse).

The motor spins in a repeating circle, so A-B-C firing order is the same as B-C-A and C-A-B, so I simply don't believe there are exceptions due to misalignment, only 60° or 120° phase angle. This method has worked for me every single time with many different motors and controllers as long as the phase angle matches. Kollmorgen, Bafang, Xlyte, Wilderness Energy, 9C, Revolt, Zhejiang Haoren, HubMonster, QS motor, Fusin along with a bunch of no-name motors (both DD and geared hubbies) as well as my poorly aligned halls added to a Turnigy outrunner. The method even works with the wrong set of halls on my 6 phase motors that results in a 3° timing advance.

11 years ago I burned up a little Bafang motor and controller in frustration after trying a slew of combinations haphazardly swapping halls and phases at the same time. I goosed and held the throttle for too long with a bad combo to show it who was boss. Then I looked online for how to wire a 3 phase bldc motor, and read that sensorless you just swap 2 phase wires to change direction. Then after a bit of drawing and thinking it through, I came up with the method I've used ever since. Only once in the past 10 years did I get the most unlucky and tried 11 combos before getting it right...the last of the 1st 6 was a valid reverse and the last of the 5 was the good forward. It typically takes less than 11 or 12 total to get both controllers done on my 6 phase motors. Confidence and taking my time to be systematic makes it quick and easy.

Sine wave controllers make the false positives harder to identify, so I always use a current meter to look at the no-load current.
 
fechter said:
d8veh said:
What about direction@ Out of the 6, three forward and three reverse.

All that only works if the timing is balanced. I've had motors where no combinations worked with KT controllers and the same motor would only run in reverse with self-learning controllers no matter what combination was connected.

Also, out of the three combinations that will give correct forward motion when the timing is balanced, one of the three sometimes runs better than the others.

The old MAC/BMC motors had non-neutral hall sensor positions that would give all sorts of problems when trying to mate to a normal controller. The trick with these was to get the lowest no-load, full throttle current. Ideally a motor like this needs a controller designed for the timing shift or use a sensorless controller.

That's right. I tried to update an old Ezee bike with a KT controller, but no combination would work. I tried three different KT controllers, but none would work. When I tried another controller with auto phase and hall detection, itran perfectly backwards every time, regardless of conection sequence. There must therefore be some asymmetry in the Ezee motor too.
 
FWIW, no, not changing two at a time. Not that dumb. But looks connected sometimes is not. And that can drive you absolutely crazy. I can't be the only one this happens to.

Why I always seem to find reverse first is just the same as how you can flip a coin ten times and get ten heads. Shouldn't happen, but it can. Cracks me up I always find reverse first.
 
The thread above was very helpful, coming from my perspective as a beginner. I wasn't totally clear at a couple of points though, so here's my attempt (...as a beginner) to write out beginner-friendly steps for finding a working combination of phase wires and Hall wires between the motor and controller.

This process requires you to try up to either 6 or 12 combinations of phase wires, and either 1 or 2 combinations of Hall wires.

1. Connect the Hall wires on the motor side with the Hall wires on the controller side. Use any convenient combination, e.g., like colours to like colours

2. Connect the phase wires on the motor side with the phase wires on the controller side. Use any convenient combination, e.g., like colours to like colours

3. Squeeze the throttle
- If the motor turns smoothly and in the correct direction, you have probably found a working combination. However, if you're a beginner and aren't quite sure what counts as smooth, then take note of this combination of phase wires and proceed to step 4 to see if you can find a smoother-running combination
- If the motor does not turn, or turns roughly, proceed to step 4
- If the motor turns smoothly but in the wrong direction, skip ahead to step 5

4. Try another combination of phase wires by swapping two (and only two) of the pairs of phase wires. E.g., if the first combination of phase wires was blue, yellow, green on the motor side to blue, yellow, green, on the controller side, then the second combination you try can be blue, yellow, green on the motor side to blue, green, yellow on the controller side
- If the motor turns smoothly and in the correct direction, this is a working combination
- If the motor does not turn, or turns roughly, repeat step 4, trying a new combination of phase wires (there are 6 possible combinations)
- If the motor turns smoothly but in the wrong direction, proceed to step 5

5. Swap two (and only two) of the pairs of Hall wires (but leave the red and black Hall wires connected with each other). Next, go back to step 3

If you prefer, you can modify the above process so you instead try up to either 6 or 12 combinations of Hall wires, and either 1 or 2 combinations of phase wires.
 
Word of caution when trying the combo flowchart. Make sure you are clear of anything that can spin. I thought I’d broken my ankle when the motor spun in reverse and the pedal hit the inside of my knee with full motor torque. Lucky it was only a grade 2 sprain but still took six months to come good.
 
Back
Top