Re-wind of a turnigy 80/100 (Now-tutorial w/Video)

[johnrobholmes]

Well that is a backwards way of looking at it, but I do see your point. It would be easy enough to force air over an inrunner can too though, and mounting an inrunner in the natural flow of air on a bike is certainly easy enough.

I still need to contemplate mounting the magnets in the "cool zone", instead of trying to cool the windings to keep the magnets cool. It is going to be hard for me to agree on this one :lol:

 

[phyllis]

LFP, do you have a link to your cooling setup?

Thud, you can check out the end of Kingfish's "Is all magnet wire the same?" thread for some calcs about time to reach critical temperature. Just use the same formula but calculate for coolant as well by inserting its specific heat (thermal capacity).

 

[Thud]

wow, this really is turning into a debate, LOL.

Luke, all your points are straight logic, make sense & I totally understand boundry layer fluid dynamics (engine porting )

You know I love the outrunner topology for all the same reasons. But you have to admit, air flow "through" the motor cooling the stator is dismal on a motor with a skirt bearing.

Being the stator is the primary heat source:
Are you saying its worthless to bother cooling the stator because the magnets are the weak link? & that air (althought the least effective conductor of heat) is best as it is the lightest viscosity coolant we can wet our motors with? That also fits into the realm of logic. A tiny pressure(centrifugal/tangential) blower would be far less cumbersome from a fabrication & installation stand point. CPU axial fans & ducted fans are worthless in a pressure applications.

 

[adrian_sm]

Here's a picture of the current Prius transaxle with its two BLDC motors:
View attachment 2

The whole thing is immersed in light oil. ....

Are you sure that the motor is totally immersed? The reference I have seen shows it having a typical oil bath at the bottom that the edge of the rotor dips in to, and they rely on an oil mist, rather than immerse to cool.
Here is the link: http://www.google.com/url?sa=t&sour...fG86yl0Jp_eMMnMcg&sig2=H0hHIkJQa8isOPo2RDNIAg

Figure 9 shows the Prius 2004 rotor. The periphery of the end piece is not perfectly round.
It has eight semi-round grooves that are used to sling the oil to produce oil droplets inside the
motor housing for cooling the motor windings and cores.View attachment Prius fig.9.png

View attachment Prius fig.15 - oil bath.png

 

[adrian_sm]

Found another reference that explores the thermal management of the prius motor. Lots of test results, and computer modelling.
REPORT ON TOYOTA PRIUS MOTOR THERMAL MANAGEMENT

Some good discussion on the max temps, main cooling paths in the motor, and suggestions on how to improve it.

Some of the interesting quote:
p.11

The permissible temperature rise of the motor winding, 170°C, and of the motor cooling oil (or
transaxle fluid), 158°C, of the Toyota Prius motor during tests was determined from the trip conditions of
the motor, 174°C, and the transaxle fluid, 162°C, given on P. 426 of the Toyota Repair Manual, Vol. 1,
2004, No. RM1075U1. The 4° difference between the permissible temperature and the trip temperature
was arbitrarily selected for safe operation during the tests without causing a trip.

p.36

this motor configuration resulted in an oil film forming in the
stator-rotor gap. This film approximates Couette flow and would result in good heat transfer between the
rotor and stator surfaces across the gap. In this model, this gap was approximated by a high-conductivity
solid.

p.37

The main cooling mechanism in the motor is to the casing. Cooling occurs in two ways: via
conduction through the stator to a cooled casing or via convection to the cooling oil and then to the
casing. The latter mechanism results when the slinger splashes oil on the portion of the stator cooling
wires that protrude from the stator iron core.

 

[Jeremy Harris]

You're right, it's not totally immersed statically, but it will effectively be when it's spinning. Having run a motor partially immersed in a jug of oil I can confirm that the motor just picks up the oil and covers itself with it, even at low rpm, so I'd guess that the Prius rotors do the same. The talk of oil droplets and spray is interesting, but having seen the way oil behaves I don't think it's an accurate description of the process. Certainly you're not going to get water-droplet-like behaviour from oil of the viscosity used in the Prius, it will tend to get chucked around as sheets sticking to surfaces I believe.

Jeremy

 

[liveforphysics]

In this picture, I see the guy making the line at the base of the tranny oil fill port.

Not that it matters a whole lot, but I think he has that case clocked/rotated incorrectly.

Reason? No toyota/honda I've ever encountered increases the tranny fluid level above the bottom of the a CV axle seal. It's simply considered to be poor engineering practice, and your car ends up like all the drippy domestic/euro-trash cars. They even sacrifice ground clearance with elaborate tranny case sumps or use internal pumps to ensure they can get the needed volume of oil in the tranny and where it needs to be in the tranny while avoiding having the oil level exceed the axle seal height.

This also enables us to swap out a broken front axle on something like a Civic/Accord/Prelude/Camary/Corolla/Integra/CRV etc in about 5 minutes at the track, because there is no fooling with oil needed (if you keep the car level when you jack it at least).

 

[Ludo91]

As we can see from the main page pictures it is possibile to run the motore "without the front plate", just bolting the motore with some sort of spacer using the holes in the front bearing support where the front plate is bolted too!

 

[Arlo1]

awesome

 

[Ludo91]

ya cuz it allowos us to have a lot more space for air to flow out of the motor if we puta a fan on the other side

 

[Arlo1]

Hey thud you can rewire it as a 6 phase and run 2 controlers which should up the efficiency a little and make the controlers happier

 

[Thud]

If any one missed the twin wound /twin controller motor testing...........
http://endless-sphere.com/forums/viewtopic.php?f=30&t=22091&start=15

I couldn't watch another HV160 go up in a ball of plasma the way that one did....It must have been a cumulative effect as the set up worked for awhile. I stopped to adjust the brakes & when I rolled the throttle back on...snap crackle, POP, POOFFFFF!!!!

I really hate having that mutch plasma being liberated so close to "the boys"....the controllers were an inch below the seat. :lol:

 

[Miles]

:lol:

 

[Arlo1]

If any one missed the twin wound /twin controller motor testing...........
http://endless-sphere.com/forums/viewtopic.php?f=30&t=22091&start=15

I couldn't watch another HV160 go up in a ball of plasma the way that one did....It must have been a cumlitive effect as the set up workd for awhile. I stoped to adjust the brakes & when i rolled the throttle back on...snap cracke, POP, POOFFFFF!!!!

I really hate having tha mutch plasma being liberated so close to the boys......the controllers were an inch below the seat. :lol:

So...... Still with the RC toy controlers. I am going to use 2 12 fet controlers for this,
It would have 6 hall sensors and two of burties timing bords.

You dont think it was because of it being wound as a twin wind do you?
And how was it split ???? Was it parrellel on each phase or each controler powered its own seprate teath?

 

[Thud]

Arlo,
I am not sure why the one went up in smoke...It never made the "lost sync" screetch sound.

Using sensored controllers may be the way to accomplish it as there could be no arguments.
that motor was wound as a dual LRK set up, terminated in wye. One controller for 2 coils on each phase.
1x2x3=6 per controller.

I have looked at the geometry & if you go to 8 magnet pole-pairs you can run 2 controllers for a single set of halls.
Either 8 or 16 pairs on that colossus motor...& unless your winding for some lower Kv & reasonable expectations get ready to start vaporizing the xie-chang controllers.

It really is all about the motors.
I will prolly run the Big CA-120 with my 18-FET unit...it is a 1/4 colossus motor from my observations.

Also running back-to-back with stock 80mm motors...the CC HV-I60 DOMINATED even the 18-FET xie chang controllers. I am making all my motors more E-vehicle friendly & I think the bike controllers now have a chance.

Need time to get the JH Lobo controller built & tested...it should dominate the xie-chang for motor controlling.

 

[Arlo1]

Arlo,
I am not sure why the one went up n smoke....It never made the lost sync screetch sound.

Using sensord controllers may be the way to accomplish it as there could be no arguments.
that motor was wound as a dual LRK set up, terminated in wye. 1 controller for 2 coils on each phase.
1x2x3=6 per controller.

I have looked at the geometry & if you go to 8 magnet pole pairs you can run 2 controllers for a single set of halls.
Either 8 or 16 prs on that collosus motor....& unless your winding for some lower kv & reasonable expectations get ready to start vaporizing the xie-chang controllers.

It really is all about the motors.
I will prolly run the Big CA-120 with my 18fet unit.....it is a 1/4 collosus motor from my observations.

Also running back to back with stock 80mm motors....the CC hvI60 DOMINATEd even the 18fet xie chang controllers. I am making all my motors more E-vehicle friendly & I think the Bike controllers now have a chance.

Need time to get the JH Lobo controller built & tested..it should dominate the xie-chang for motor controlling.

A sensorless controler uses the back emf to figure out rotor posistion so if two controlers are pulsing the same stator tooth then that would lead to major issues with a sensorless system!
I have the small collossus motor as well and its just right for a 6 phase motor then you will have 2 the resistance in the windings if you aim for the same copper fill and KV!!! And it will be more efficient because of the better timing.

 

[John in CR]

I love the idea of a 6 phase, but each controller seeing only half the coils is going to be half the resistance instead of double, won't it? We need a guru to let us know if split windings would be problematic. I think parallel coils may be, because don't parallel coils interfere with each other?

Hey Thud, have you tried a sensored rig using a low current setting on the ebike controller and applied your torture test? I ask because the Block Time thing keeps nagging at me, since the programming interface doesn't allow zero. It blocks the current limits for that period of time, and the minimum is 1 second. My concern is that sure the ebike controller limits the current in regular operation just fine, but with it looking down the barrel of these low resistance motors, allowing a full second of unlimited current every time you hammer the throttle would shoot the current limiting in the foot.

 

[Arlo1]

I love the idea of a 6 phase, but each controller seeing only half the coils is going to be half the resistance instead of double, won't it? We need a guru to let us know if split windings would be problematic. I think parallel coils may be, because don't parallel coils interfere with each other?

Hey Thud, have you tried a sensored rig using a low current setting on the ebike controller and applied your torture test? I ask because the Block Time thing keeps nagging at me, since the programming interface doesn't allow zero. It blocks the current limits for that period of time, and the minimum is 1 second. My concern is that sure the ebike controller limits the current in regular operation just fine, but with it looking down the barrel of these low resistance motors, allowing a full second of unlimited current every time you hammer the throttle would shoot the current limiting in the foot.

No i talked to biff on this.

The way i am sugesting is 6 teath are 100% seprate from the other six
and each six have their own 3 hall sensors and their own controler!

Think of it this way if you have it as a 3 phase that meens for a 12 tooth stator you have 4 teath wound with x number of turns to get x KV. But if you change it to 6 phase you still need the same x number of turns but only on two teath so the amount of parrellel turns gets cut in 1/2 thus raising the resistance and making it 2 times as easy on a controler!

 

[Thud]

Arlo1 said:

The way I am suggesting is 6 teeth are 100% seprate from the other six
and each six have their own 3 hall sensors and their own controller!

That is exactly my set up that lunched the HV-160.

John, I haven't ridden my e-bikes since last october...but the sun is out & the snow is almost gone from the parking lot test track. (its a 1/2 mile of muddy dirt road to get there )
I am mounting the little colossus onto my death race bike today...just finished the adapter to bolt it to my new tranny....the Kv is too low now lol! (nothing a little gearing can't fix!)

I stayed home from work today...feeling a bit under the weather as it is. I am anxious to test with the 12-FET controller with the 3077's in it. I am limited to 75V on that but the 18-FET unit has 4110's & will take a full 100V. Not that I need that much.

 

[John in CR]

October OUCH! I'd be beating my kids and wife if I had to go months.

If there's an option I'd always go with higher voltage and less current for the same power, and gearing variability gives you that option. I can't imagine a 12fet working for power riding. I hope those are super fets. I'd strongly urge you to forced ventilate your controllers with a small centrifugal blower. It's not only the fets that generate heat, and it almost doubles your heat sink area for the fets, and cools the other fet surfaces along with all the board components.

My Bosch charges have neat little centrifugal fans that I scavenged for my controllers. They fit in the end plate, so I use it to suck warm air out. At the front end I use two 1" pieces of .5" nylon tubing angled forward and down with a small balloon of AL screen so debris can't enter and water can't get sucked up hill into the controller. One piece of tubing is too restrictive for the fans, which is easy to tell by the sound.

My hubbies torture controllers similar to RC motors, and these little 2W blowers make a huge difference in controller temperature. I was able to push double the power past the failure point of the stock controllers that came with the motors, because the fan replaces the air in the controller almost twice per second. Don't even bother with a regular radial fan, because they can't handle any kind of flow restriction without using big power like a ducted fan.

 

[Arlo1]

Arlo1 said:

The way i am sugesting is 6 teath are 100% seprate from the other six
and each six have their own 3 hall sensors and their own controler!

That is exactly my set up that lunched the hv160

not exactly you tried it with no sensors at all. I am confident 6 seprate hall sensors 3 for each sensored controler will work!!!

 

[phyllis]

spectra melting yet? supposed to take 100c max.

 

[Thud]

Have not lost a re-wound motor yet, p...(touch wood) I have only "had" to use it on 2 motors so far. (significant damage to the stator ends)
I think the spectra melts but still insulates with a bit of goo... :lol: I have also used cotton based upholstry thread...That should take a bit more heat.

 

[E-racer]

Just ordered my wire from mcmaster http://www.mcmaster.com/?orderview=new#7588k24/=bg7lib

I'm going to give this a shot on the 80-100 that's on my ESR scoot. If it works out well I'm going to do the pair of them that will be on my E-bike. This allows me to run my Kelly at 14s without running into the E-rpm limit. Thanks, Thud!

 

[phyllis]

Have not lost a re-wound motor yet p.....(touch wood) I have only "had" to use it on 2 motors so far. (signifigant damage to the stator ends)
I think the spectra melts but still insulates with a bit of goo.... :lol: I have also used cotton based upholstry thread...That should take a bit more heat.

So you're running phase change cooling (solid to liquid) - pretty advanced stuff... Cotton would prolly be ok, and glass roving.