Adding hall sensors to outrunners

[modern_messiah]

You'll wind the stator ABCABC... which makes it neccessary to put the halls right in the middle of a tooth.

How do you know this? Where is this written? Probably in one of the threads I've read through a million times. *sigh* If you have the time can you explain why this is the case for ABCABC? How is it different to any other winding scheme?

I'm not questioning your response - I'm simply trying to understand why this is the case so I learn something instead of just blindly doing something some one tells me is correct. Basically so I understand the principles and don't have to ask again in the future!

I think it's called learning

[olaf-lampe]

With an AabBCc winding like the usual turnigys have, you can put the halls between the Aa/bB/Cc coils. That's a timing-neutral place and works for both directions. With an ABC winding, the only neutral place is the center of the coil.
-Olaf

[modern_messiah]

Bugger. Putting them dead center of the coils will be a pain. Oh well, another puzzle to solve

[spinningmagnets]

As someone pointed out to me a while back, a hub-motor is an outrunner with the spokes attached to the shell. On the new Crystalyte hubs (HT/HS) the first batch didn't have hall-sensors, but they all had the factory-cut slots for them:
http://endless-sphere.com/forums/viewtopic.php?f=2&t=24593&start=165#p381693


And in the Tucson "Death Race", PaulD's winning Turnigy 80-100 had external halls spaced at 17.1 degrees
http://endless-sphere.com/forums/viewtopic.php?f=3&t=24507&start=330#p390931

[modern_messiah]

Well %$^&.

Using 9 teeth and 12 poles (ABCABCABC) requires the magnets to be placed 20 physical degrees apart from each other. However it is not possible to place them all in the center of a tooth. In fact one will be centered, while the other 2 will be right between the teeth either side of it.

How can this be fixed, if at all? Or is this something that can be fixed by adjusting them a few degrees (i.e. keeping them at 20 degrees to each other but rotating all 3 a few degrees relative to the teeth)?

[TylerDurden]

@ m_m:

Is there a problem physically placing the sensors at the perimeter of your stator?

If so, you can also consider optical sensors. IIRC, Tidalforce motors did .

[modern_messiah]

Is there a problem physically placing the sensors at the perimeter of your stator?

I'm not totally sure what you mean by this? I can place them anywhere I want - internally, externally on a separate jig (behind the backing plate so it will rely on flux leakage) and so on. What I mean is that Olaf stated the best place was the center of the coils as this is 'timing neutral'. However if they are placed at 20 degrees, as required by my motor you get the following:

Looking perpendicular to the rotation of axis - this is an axial flux motor. Blue circles are coils, black dots are hall sensors, the grey is part of the stator holding the coils in place.


Notice the 2 outer sensors are unable to sit directly in the middle of a coil like the center sensor can. No matter how I rearrange the sensors, if I want to keep the 20 degree separation between sensors this will always be the case. Is this a big problem Looking at the picture of the Crystalyte hub above it has a similar problem, though as Olaf hinted to earlier this apparently has something to do with the winding scheme?

I'm so lost lol.

[olaf-lampe]

I don't see a reason, why you couldn't put them 120° apart... anyone?
Just make sure they see the whole diameter of the magnets, otherwise you waste precious dutycycle.
-Olaf

[modern_messiah]

I don't see a reason, why you couldn't put them 120° apart... anyone?
Just make sure they see the whole diameter of the magnets, otherwise you waste precious dutycycle.
-Olaf

1) Can you explain the 120° theory? I'll give it a go now actually...I pieced this together from various sources so I won't be surprised if it's very wrong.

To place Hall sensors properly in a motor, you have to know how many electrical degrees each tooth occupies:

°elec = 360 * p / t where p = number of pole pairs & t = stator slot count. For my motor °elec = 360 * 6 / 9 = 240°

The first sensor is placed anywhere - this becomes zero degrees. I will place mine dead center of a tooth as this is "timing neutral". The second sensor must be °elec ahead of the first sensor. Each tooth is 240 electrical degrees. Begin at the position of sensor A as this is 0°. Add the °elec to each previous position until you reach 120° (this is the desired location of the sensor). Obviously each time you pass 360° you subtract it.

So you end up with:

1) 0 + 240 = 240.

2) 240 + 240 = 480. Subtract 360 = 120. The number of iterations is 2 and seen as you are at 120° the second sensor should be placed 2 slots away from the first.

You now treat the second sensor as the starting point in order to place the 3rd sensor 120 elec° from it. You get the same result for sensor 3. So according to my maths my sensor placement should look like this:



If I was to do as you have suggested Olaf, there would be 2 empty teeth between each sensor.

Can someone clarify this? What's right!?!?!

[olaf-lampe]

Maybe both theories are right.
For the hall sensor it doesn't matter which magnet triggers. As long as the magnet has the right polarity and the right timing. I'm pretty sure, there is a propper magnet in 20°, 120° and 240° distance from the first hall.
You should paint the magnets in your picture red or blue ( north/south) to clearify that.
-Olaf

[modern_messiah]

The blue rings in the above drawings are coils...let's try again.

Black dots are Hall sensors, Yellow Circles are Coils, Blue and Red are magnets (alternating of course).

[olaf-lampe]

I stand corrected. That picture clearly shows, the 120° doesn't work here.
The way you've drawn the hall's is the way to go.
-Olaf

[modern_messiah]

I stand corrected. That picture clearly shows, the 120° doesn't work here.
The way you've drawn the hall's is the way to go.
-Olaf

Wait...I got something right? I'm learning!!!

Progress feels fantastic

EDIT: Just for clarification; is the reason 120 physical degrees won't work because if you put them at this separation all 3 hall sensors would be over magnets of the same polarity at the same time, and hence the controller would have no idea which way it was going?

[Thud]

in the above illistration the halls are not 120deg apart. they are near 90 deg. 120 will work on any 3-phase machine.

now to the report:
I have gone to internal Hall sensors on a re-wound Turingy 80-100 motor & can say that the 12fet controller is finaly delivering enough performance to keep me interested.

Thanks for the steerage Burtie. I doubt I will ever recomend external placement again....wich sucks as I built tooling to make a Jiffy slip-on Hall bracket for the turnigy motors with the halls cast in resin & idiot proof for instalation & set up....now thats out the window...I am goig to go make a video & kill a few battery packs!

[HumboldtRc]

I'm still trying to figure out which slot to start the placement of the sensors....?
Do I just chose any slot and do the 120 degree spacing??

[Thud]

Humbolt,
I placed the sensors in the centers of a wound pr of teeth....since I wound this motor I had it very easy.

trial & error is actualy easy though....pick a slot then every 4th slot gets a hall....if your motor will run in either direction equily well you hit it on your 1st guess. if not, moving to the next slot will get you on track.

Kurts, trike thread has some good notes & video of his motor set up.

[AussieJester]

trial & error is actualy easy though....pick a slot then every 4th slot gets a hall....if your motor will run in either direction equily well you hit it on your 1st guess. if not, moving to the next slot will get you on track.

Whats the best method for holding the hall sensors in during this 'trial and error' period
seeing epoxying them in would be out of the question.

KiM

[HumboldtRc]

I read here, someone held there's in with tape to test them. I guess I will just give it another try...

[Thud]

a dab of hot glue can be picked out easy enough also. & then re-epoxied

[kfong]

Thud, thanks for the heads up. I was planning external sensors. What halls are you using? Looking forward to pics and vids.

[Thud]

I am using the
Honeywell s411a

I mistakenly bought some 411p that were terible in the xternal set up....maybe they will work internaly?
you prolly have a mutitude of Ellectronics sources K, but I bought my last batch from Mouser.

http://www.mouser.com/ProductDetail/Hon ... AcF98Uk%3d

[Thud]

Alright,

turnigy 80-100 re-wound 6turn 14g dbl wye.
internal halls 120deg spacing
XieChange 12fet irfb3077's
beefed traces with 14g copper wire
1k resitor across c20
18s lipo
programed 60 amp battery/150 amp phase.....the set up delivers very lively performance

after an 8 mile tourture run with speeds peaking at 43mph...cool controller, motor slightly warm (no problem holding a hand on it)
burned 6.5ah out of a 15ah pack.
(started on a freash charge....run & measured the ah to re-peak the pack)

[liveforphysics]

Hooray! You've got an inf infinion setup right on an RC motor! It's a damn force to be reckoned with with you get everything setup correctly (which is not easy).

[Thud]

Thanks Luke,
You have no idea how much I would like the death race to be next weekend ....I pounded this set up tonight...seems absolutly rock solid.
I rolled film tonight but the camea slipped leaving the drive & all you can see is my right hand

[Cheeseboy]

Okay Guys, I've just about had it with this I've gone through the spreadsheet of possible wiring combinations THREE TIMES (108 combinations!) and have so far been unable to find success. I recorded a video, if you have a few spare minutes PLEASE help me troubleshoot
[youtube]http://www.youtube.com/watch?v=3mevtsqnIVA[/youtube]
If I didn't do that right, the direct link is http://www.youtube.com/watch?v=3mevtsqnIVA
So the problem I have is that when I turn the throttle at any speed faster than a tedious constant crawl, the motor cuts out and I have to return the throttle to zero to be able to start it again. Also as you can probably hear it doesn't sound too happy, but no load current is just over 2A, I killed my watt-meter so I don't know what the no load current was when running it sensorless.
The way I have it set up currently makes it spin the wrong way around but IIRC switching one of the hall wires will fix that.
In all of these different wiring combinations I have kept the sensors in the same slots, now from reading this thread, and another very similar one I've become confused as to whether one needs to try different slot locations for each of the sensors? It is a 12 pole motor with the sensors evenly distributed in the slots. (a 3 slot gap between each)

I forgot to say, I am using The honey well S411A's and another interesting thing is when I change the controller to 60 degree timing it appears as if nothing changes.