Evader EV1000 BLDC motor and hall sensor replacement

Hello,

I have a electric scooter which I've been riding for a year now. It was sold here as "Leone Ecopower" and I think it is/was sold as "Evader" elsewhere. I bought it used and previous owner had replaced its original controller. It was obviously some cheap chinese controller but so was the replacement. It gave jerky starts and some electrical noise coming from motor. I got myself a Phaserunner controller from eBikes.ca. It doesn't want to tune to this motor properly and throws me hall sensor errors. I can run it sensorless but would love to get it working sensored. Tried swapping hall sensor order while phase leads were left untouched. This did lead me opening the motor and inspecting insides. Too bad I cannot fully identify the motor type as text on it has partially worn out. I thought I would post pictures and questions here if somebody else happens to have similar motor.

It has 40 magnets around the rotor and stator has 36 teeth. I have no idea about actual winding scheme but it looks like three stator teeth are wound as one magnetic pole, wire skips next 6 slots and there is another pole starting from there. It looks delta wound as I see no wye (star) point.

It goes like this (imo): phase A slots:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

I've read various hall effect related threads here for days. Does winding pattern matter in this situation or is it only magnet position related thing? I read that delta windings might offset timing by 30 degrees.

360 degrees / 36 teeth = 10 degrees mechanical for every tooth. So halls are spaced 15 degrees mechanical apart from each other.

360 degrees / 20 magnet poles = 18 electrical degrees per one full mechanical rotation. Hall sensors should be placed 120 electrical degrees apart from each other, right? How come 3 hall sensors are spaced like in pictures (every 1.5 teeth or 2nd sensor in slot and two other embedded to stator teeth)? Should I re-arrange hall sensors?

Now I've seen similar hub motors with similar hall spacing. They seem to work. I got a 3000W MXUS motor with hall sensors that works fine with Phaserunner. It tunes properly and works in fully sensored mode. Pulls nicely from standstill even with a 120kg rider (80 phase amps, pack at 48V). Even this Evader motor pulls okay from standstill when using cheap chinese controllers.

I've ordered new hall sensors already (Honeywell SS41G) and some thicker and more flexible phase wire (12AWG).

Main question is: should I alter hall setup? If not, are current hall positions 120 electrical degrees apart from each other?

Thank you for all the help.

Mora said:

It has 40 magnets around the rotor

Click to expand...

That makes it a 20-pole(pair) motor, as far as settings go in controllers that need that.

I have no idea about actual winding scheme but it looks like three stator teeth are wound as one magnetic pole, wire skips next 6 slots and there is another pole starting from there. It looks delta wound as I see no wye (star) point

Click to expand...

.

The pics don't show the whole winding-connection side, but from what I can see in them, I don't see the star connection either.

Does winding pattern matter in this situation or is it only magnet position related thing? I read that delta windings might offset timing by 30 degrees.

Click to expand...

It's really about the magnets; all it is detecting is the passing of magnets in the rotor.

I suppose a different winding scheme might change the timing of the pulses needed to drive the magnets, but the actual hall positions are determined by the magnets.

360 degrees / 36 teeth = 10 degrees mechanical for every tooth. So halls are spaced 15 degrees mechanical apart from each other.

Click to expand...

It's the "electrical" spacing that matters. Most likely it is 120 degrees, but it could be 60, which is the other common standard.


360 degrees / 20 magnet poles = 18 electrical degrees per one full mechanical rotation. Hall sensors should be placed 120 electrical degrees apart from each other, right? How come 3 hall sensors are spaced like in pictures (every 1.5 teeth or 2nd sensor in slot and two other embedded to stator teeth)? Should I re-arrange hall sensors?
No, no reason to change it; that's a very common way of doing it.

I don't know the reasoning behind it, but pretty much every "generic chinese hub motor" (regardless of "brand name") with halls I've seen has been built that way. The UltraMotor I have off a Stromer is not--it's actually got them physically 120 degrees apart.

I've ordered new hall sensors already (Honeywell SS41G) and some thicker and more flexible phase wire (12AWG).

Click to expand...

Just make sure you install the new sensors the same way as the old ones. If they have one facing up and the others down, do it like that. If htey're all the same way, do that.

You can test the halls for operation before replacing, if you like, per the directions at ebikes.ca 's learn-troubleshooting section.


That said, that looks like a monster torque hub; looks like 60-80mm width for the magnets, doubled up end to end like that?

Thank you for the reply.

Here are pictures of the winding side. No star point there either.

I was thinking of trying that sensors 120 degrees physically apart thing. I got few sets of sensors coming so I could try CA'ing them in place, wire them up with separate wires and see what motor controller thinks. I'll replace those original sensors too. Originals work to some degree as there is 5V output when magnet is present and no output when there is no magnet nearby.

I measured the magnets too. They are 12mm x 70mm x 3mm. 70mm is total width so one single magnet is propably close to 35mm. This thing produces enough torque to push the scooter and its driver uphill at 45km/h at 48V. This is the legal limit here for these things. Any extra torque from standstill is fine for me

Pulled the endplate off to get better view of windings. Now I think it is wye/star winding after all. Only one bundle of copper wires enter the stator slot when phase winding starts. Haven't cut the wire that is holding fourth protective sleeve in place but I suspect the star point is hidden (and protected) there.

I'm still puzzled why current hall sensors are spaced like that. If 360 eDegrees equals 18 mechanical degrees then 120 eDegrees should be 6 degrees, right? Now it seems halls are spaced 15 mechanical degrees apart from each other. It doesn't seem its anywhere near 120 eDegrees. I drew a picture in sketchup to illustrate everything better.

New sensors haven't arrived yet. Gah.

Maybe you missed the part of my other post where 60 degrees is another "standard" spacing. I don't know if that's what these are setup at, either, but you could try that.

Also note the middle sensor is probably flipped over relative to the other two, which would probably make a difference in the "spacing".

All sensors are facing the same way. I took them off few minutes ago. I will definitely try 60 degree spacing too if none of previous approaches work.

I would just stick with the original locations since you know those work.

Since the rotor is made from alternating polarities of magnets, there will be many locations where the hall sensors will work. All that matters is the gap between magnet sections passes over the sensor at the right time. It doesn't matter where on the ring this happens.

Main problem with hall sensors was they were not working as Phaserunner controller wants. I bet original controller this scooter had did nothing with hall sensors. I can setup Phaserunner totally sensorless and it works as well as previous controller. I'm after smooth torque control. Well tests should be relatively quickly done whenever I receive new sensors.

I believe you will find an explanation to the position here:
http://mitrocketscience.blogspot.com/2011/08/hall-effect-sensor-placement-for.html?m=1
Apparently multiplication of the spacing with integer values work also, ie 360/(20*6)=3deg mech spacing for 60 deg spacing, multiplied by 5 = 15 mechanical spacing, so you probably have spacing for 60 degree hall setup.

So.. Do you have a 60 degree spacing controller? I'd try that first. If you have a 120 degree controller you could place new halls at 6*5 mech spacing, that's at two slots between each

As far as the delta offset goes.. 30 degrees electrical offset is 1,5 mechanical..don't know how to handle that without grinding new hall slots in the teeth.

I think your best bet would be the 60 deg hall controller and keeping the positions in the slots as they are.

New sensors arrived, finally. Tried something simple first: flipped the sensor in the middle upside down. Had to dremel the slot a bit to make it accept sensor that way. Replaced phase wires with thicker ones at the same time. Phaserunner suite autotuned everything nicely at first try. Oh the smoothness now! I read from some MXUS 3k motor thread that Chinese motors might have hall sensors installed wrong right at factory. Cheap controllers don't seem to read hall signals either so it doesn't matter if they're all wrong Now to fix sensors more permanently to stator. I used hot glue to keep them in place for testing.

Congrats!

Isn't the phaserunner configurable between 120&60 spacing?

Thank you. I believe it figures everything out during autotune. It also notified me that hall timing offset was something like +7.9 degrees. This motor might have 60 degree hall spacing after all.

I had to make a quick'n'dirty forced air cooling for the controller by using two 80mm computer case fans I had lying around. Thermal management reduces phase current really quickly if forced air cooling is not used. I made airduct out of cardboard and pointed it at controller heatsink. Now I can sustain 50km/h and climb hills without losing power. 96A phase current is still not enough for this thing in my opinion. It pulls nice and smooth all the time on flat road until max rpm is achieved. Feels like it would go beyond that if I played with field weakening settings a bit. Well that speed is enough for city driving, I'd just like to get sharper response from standstill and especially when stop'n'go happens on uphill.

I swore I would not even touch Kellys anymore. There are no 200A+ phase current capable BLDC controllers around? Real FOC is a must and Kellys don't even have torque controlled throttle. But do I have any options other than Kellys?

I bought a Votol em150s from QS motor. Haven't started it yet but exterior quality and program looks good. Nearly 400A phase..
If it does what it says then it's great for the size and price.

Real FOC is a must

Click to expand...

Why? It's not a showstopper in noticeable performance difference.

ASI 8kW controller would be my FOC bet just from the specs but i haven't seen anyone use it here yet and the price is $$$

860 maximum phase amps Now that would make bad contacts weld together quickly. Or burn down the whole thing. Price difference between 4000W model and 8000W isn't that big so might as well get the 8000W unit. Winter is here soon so it is a project for next spring. Price seems reasonable regarding maximum phase amps.

I have Paul & Sabrinas AC-controller PCBs lying around too. It can drive BLDC motors if I remember right but it doesn't accept hall sensor input. Or does it? Will have to check how it was.

If motor has hall sensors installed then why not to utilize them properly and feed motor at the very right time? Sensorless starting hasn't been good in any controller I've seen/used. You are right, it is not mandatory.