12 kw rc motor

[Arlo1]

WHat if we dont split the windings up and still hook 3 sets of controlers up? 3 36 fet controlers wired in 3 different spots with 3 phases each still making it a 9 phase but using all the windings of each phase still so the KV stays the same and the controlers have a chance!

 

[7circle]

Just to keep things in focus of what I'm expecting to see in the Colossus.
Here is another sketch to relate 20 magenets to 18 poles.
The above links didn't show both together.


Trying out ideas with 9 phases.
Triple controllers with 3 sets of Hall Effects is getting complicated.
But feel free to sketch it out on the above image.

 

[markobetti]

Great ideas .. this is todays new test : 1 km ride with colossus . Kelly limited on 44% of power in torque mod . And top speed in second gear was 30kmh. Nothing special , but much better then the first time .

going out of the garage : low rpm on kelly dont work well so the car needs at least a straight meter to accelerate normally

http://www.youtube.com/watch?v=BMvSUpk5ymg

2nd gear test

http://www.youtube.com/watch?v=_Pu0rdpS8QU

going 30 mph

http://www.youtube.com/watch?v=Mzff6BOu_3k

 

[markobetti]

sorry , its 30 kmh.... not mph...But tommorow we will try to go faster with 100percent on kelly

 

[AussieJester]

more 'tests' is it going to be made available any time soon
or will we have to sit through another 4 months of 'tests' and
talk about how it could be improved... I have lost interest myself
cancel my order please Hal.

Best of luck with the continued 'testing' hope you all achieve what your wanting...

KiM

 

[Fligh High]

Still impressive to see such a small motor size-wise powering a complete car ! Couple of years ago I would have had difficulties with believing it is even possible.

It's not easy to get a capable controller for testing , hope it will work out.

 

[markobetti]

Okay aussie i am sorry , we are not trying to be irritating bastards , just want everything properlly done. People who know motors and who have their comcept baseD allready on colossus can order now as it is ; with shipping from China can pm me . Okay? The price with shipping is 500 dollars. Understand that Hal is not going to inspect them this way. Three weeks to one month is production period . And 4 days travel to any country . Once agian ; this is not a regular order inspected by Hal .

 

[liveforphysics]

more 'tests' is it going to be made available any time soon
or will we have to sit through another 4 months of 'tests' and
talk about how it could be improved... I have lost interest myself
cancel my order please Hal.

Best of luck with the continued 'testing' hope you all achieve what your wanting...

KiM

KiM- What you're not realizing, is that the way it is, wired in 3-phase doesn't offer a solution that gives all 6 series sets of poles to be working together. Secondly, the L and R are a little more intense than anything on the market can power for under $1,500 or so in a controller. 7circle has brought up some damn slick ideas that could fix both of these problems by changing the winding to something that can work properly, and share the loads between 3 controllers to enable it to get fed enough power to make sense to use this over the big HXT you're all ready running. The testing/research/discovery/R&D stage on things is a hugely expensive and time consuming part of trying to get a decent product worked out. They are spending their own time and money trying to figure this stuff out, and I'm super thankful for it!

If you just had one of these motors appear on your bench tomorrow, it wouldn't do you a bit more good than a big rock or paper weight.

 

[DTEK]

I'd like a controller and motor package solution tried and tested before Christmas please as that is the target for my build to commence.

 

[markobetti]

No problem , We are reciving controllers in 10 days and we will try motorcycle conversion very soon . Also in 2-3 weeks the esc should be finally finished so we will see.

 

[AussieJester]

Thanks for the pm markobetti much a appreciated..
Best of luck to you and Hal...


KiM

 

[Biff]

What you're not realizing, is that the way it is, wired in 3-phase doesn't offer a solution that gives all 6 series sets of poles to be working together.

That is not correct. With the 3 tooth per group winding, the controller will work fine, as long as the three teeth in each group are wound in series (ex AaAbBbCcC...) Winding them like that may not be as good as winding it as it would be if you wound it as a 9 phase motor, but it is not in any way bad, (see my phase diagrams a couple pages back). Having that pole slot combination is essentially the same as if you had a skew on your stator (or rotor), which is a very common and most effective way to reduce cogging torque.

Secondly, the L and R are a little more intense than anything on the market can power for under $1,500 or so in a controller.

If that is the case, it can easily be fixed by adding a nice large inductance on each phase outside of the motor. This would not hurt efficiency significantly as you could use big fat wire on the inductor, and it shouldn't be too heavy since the inductor shouldn't need too many turns.

7circle has brought up some damn slick ideas that could fix both of these problems by changing the winding to something that can work properly

Re-winding the machine to be a 9 phase (using 3, 3 phase controllers) is a task that could be done by someone at home, and they would probably do a better job of it than the manufacturer, and doesn't need any modification to the current design (unless there isn't room to bring that many wires out of the controller, however, rewinding it you would have to use smaller wire, and more turns on each tooth (almost 3x the turns per tooth) to get the same kv on each of the new phases, so in theory the total cross section of wire should be about the same)

The non-equal spaced teeth solution would take some time to get prototyped and would be pretty expensive to prototype as well. The manufacturer would probably laser cut the first run, then would have to make a new stator stamp if the design was shown to work better. From what I gathered by reading this thread, the current design is just a motor which the manufacturer already makes with a modified stator that can be liquid cooled.

and share the loads between 3 controllers to enable it to get fed enough power to make sense to use this over the big HXT you're all ready running.

From what I have read, the ideas that 7 circle has presented (un-equal tooth spacing) would make this a 3 phase motor without skew. As such it would be difficult to make it use 3 controllers, it would be good for 2 controllers (as there are 2 groups of teeth per phase).

The testing/research/discovery/R&D stage on things is a hugely expensive and time consuming part of trying to get a decent product worked out. They are spending their own time and money trying to figure this stuff out, and I'm super thankful for it!

They way they are doing the research is not very expensive, but very time consuming. I am betting that they would have much quicker development time if they sold a couple prototypes to people who are aware of the issues, and were willing to do some development / testing themselves. I hope that some people take Marko up on his offer and buy some models "as is" and post their experiments here. I am currently busy working on a similar (but much higher power) project otherwise I would buy one and get it in my testing lab.

 

[Arlo1]

Im taking Marko up on the offer and I will post all info I find and all tests. I will also involve Luke for what ever tests I can he is only a couple hours away from me!

 

[markobetti]

What you're not realizing, is that the way it is, wired in 3-phase doesn't offer a solution that gives all 6 series sets of poles to be working together.

That is not correct. With the 3 tooth per group winding, the controller will work fine, as long as the three teeth in each group are wound in series (ex AaAbBbCcC...) Winding them like that may not be as good as winding it as it would be if you wound it as a 9 phase motor, but it is not in any way bad, (see my phase diagrams a couple pages back). Having that pole slot combination is essentially the same as if you had a skew on your stator (or rotor), which is a very common and most effective way to reduce cogging torque.

Secondly, the L and R are a little more intense than anything on the market can power for under $1,500 or so in a controller.

If that is the case, it can easily be fixed by adding a nice large inductance on each phase outside of the motor. This would not hurt efficiency significantly as you could use big fat wire on the inductor, and it shouldn't be too heavy since the inductor shouldn't need too many turns.

7circle has brought up some damn slick ideas that could fix both of these problems by changing the winding to something that can work properly

Re-winding the machine to be a 9 phase (using 3, 3 phase controllers) is a task that could be done by someone at home, and they would probably do a better job of it than the manufacturer, and doesn't need any modification to the current design (unless there isn't room to bring that many wires out of the controller, however, rewinding it you would have to use smaller wire, and more turns on each tooth (almost 3x the turns per tooth) to get the same kv on each of the new phases, so in theory the total cross section of wire should be about the same)

The non-equal spaced teeth solution would take some time to get prototyped and would be pretty expensive to prototype as well. The manufacturer would probably laser cut the first run, then would have to make a new stator stamp if the design was shown to work better. From what I gathered by reading this thread, the current design is just a motor which the manufacturer already makes with a modified stator that can be liquid cooled.

and share the loads between 3 controllers to enable it to get fed enough power to make sense to use this over the big HXT you're all ready running.

From what I have read, the ideas that 7 circle has presented (un-equal tooth spacing) would make this a 3 phase motor without skew. As such it would be difficult to make it use 3 controllers, it would be good for 2 controllers (as there are 2 groups of teeth per phase).

The testing/research/discovery/R&D stage on things is a hugely expensive and time consuming part of trying to get a decent product worked out. They are spending their own time and money trying to figure this stuff out, and I'm super thankful for it!

They way they are doing the research is not very expensive, but very time consuming. I am betting that they would have much quicker development time if they sold a couple prototypes to people who are aware of the issues, and were willing to do some development / testing themselves. I hope that some people take Marko up on his offer and buy some models "as is" and post their experiments here. I am currently busy working on a similar (but much higher power) project otherwise I would buy one and get it in my testing lab.

Hey Biff thanks, anyway are you also working on outrunner setup ?

 

[Biff]

Hey Biff thanks, anyway are you also working on outrunner setup ?

Yes, it is a 100kW 8000RPM outrunner. The maximum power depends on how quickly you can remove heat. Simulations indicate that the motor will be around 98% efficient, so to get to 100kW we will only need to remove about 2kW of heat.

 

[markobetti]

wow...thats f...crazy.. Now , we will be happy for you to overcome the problems of controllers .... and help with your experiance..

 

[7circle]

After looking over HAL's picture book of collosus images, it's clear that the design is what it is.
http://i3.photobucket.com/albums/y98/HAL9000v2/COLOSSUS/SAM_0189.jpg
And HALs own 18 pole 3D stator sketch for different project though (I think).

[HAL - Hope you don't mind looking over all your cool ideas in your bucket]

Bifs comment about cogging is valid and its only an issue when taking off from stand still and comming to a stop under regen.
If coping with cogging means you get more torque then give me more cogging .
But if the motor is made in a factory setup to wind the motor stator all at 20 degrees, fine. It's ~$400 it's a great buy.
And its looking to be quiet as it has less cogging. Better for Stealthy EV.

What happened to bench testing?
Finding out if the Collosus is a 12KW 6000 RPM collusus.
I thought testing it would involve setting up a Motor-Generator setup with dirty big load.
Like what HALs already done:

But longer. 12KW into a small friction disc setup isn't going to last long or be measureable for output torque easily.
As HAL has designed them to be hooked up end to end like in his car design picture adding another to make 3 in row would be okay.
They will need a strong support frame though.
[Edit :Having 1 for input power and 2 for output so you can really stress the one driving the shaft]
[Edit if the Controllers have regen you can loop the power back]

Ah.. The Controller. The Elec RPM or phase frequency needs to be 6000 x 20/2 = 60,000Hz.
Thats push and a pull of each magnet pole on the rotor past a stator pole.

Kelly is only 16kHz for PWM. So with 4 pulses for pull and push thats 4khz electrical RPM.
What can the Infinion do?
Is the Infinion Schematic online?

If you want to control the current at 6000 RPM the controller needs to pulse the FETS at 4 times 60 kHz = 240KHz.
(Some might be happy with just 2 times)
Then you need to vary the duty cycle (PWM) of the pulses to give some current control at Max RPM.
Ten steps/levels in throttle twist would mean 2.4MHz PWM Clock.
To protect the phase currents you need very fast feed back.
Switching the FETS at 240KHz (2.167 uSec ON the OFF for 2.167 uSec)
and expecting rise and fall times of 100th that, of 20nSec so Switching efficiency is minimal.

Do my expectations of the controller relate with anyone elses?

This is why some people are eager to get there hands on this to try out their Superfast Switching Controllers.
Some of the ESC's that are just logic ciruits to time the FET's from the three Hall zero crossings will work, but you'll need plenty of overhead in the current and voltage to keep them pulsing.

I'd like a controller that used the HALL Effect sensors to measure winding flux to protect against lamination flux saturation.
But at these frequencies/RPMs it'll just hold on to your ears and scream.

I think you'll be shipping several with the BETA stamp on them.

PS - I didn't sugest the multiphase idea, I was concerned it didn't suit the magnet and pole layout.
{edit - why three and controllers with Regen}

 

[7circle]

No problem , We are reciving controllers in 10 days and we will try motorcycle conversion very soon . Also in 2-3 weeks the esc should be finally finished so we will see.

What controllers did you order?
How many?

Can you confirm the Magnet dimensions used and type?
And is the outer edge of the rotor alloy or iron?

 

[Biff]

Bifs comment about cogging is valid and its only an issue when taking off from stand still and comming to a stop under regen.
If coping with cogging means you get more torque then give me more cogging .

I totally agree, It would be interesting to see how much more torque is actually available with the non-equal spaced teeth. I did a design analysis in 2008 when I had the same idea to stagger teeth. The only difference is my design just staggered the tooth heads, but kept the tooth body at equal separation so that each tooth was wound with equal amount of copper, but still interacted with the magnets as though they were not-equally distributed. I just reviewd the FEA I did on that and the actual torque available for either design was about the same. That analysis was done on rough models to gauge what was worth pursuing and what was not, for a 2 and 3 ajacent teeth per phase design, and back then I decided that the complexities of non-evenly distrubted teeth was not worth the effort. Perhaps with more research and simulation my conclusion would be different today.

Ah.. The Controller. The Elec RPM or phase frequency needs to be 6000 x 20/2 = 60,000Hz.
Thats push and a pull of each magnet pole on the rotor past a stator pole.

A slight correction on that. It is actually 60,000 Cycles per minute, not Cycles / Second (Hz), the calculation (with all the conversions so eveyone can see how the units end up being cycles / second would be

6000 Rotations / Minute * 20 magnets / rotation * 1 cycle / 2 magnets * 1 minute / 60 seconds = 1000Hz electrical frequency

Kelly is only 16kHz for PWM. So with 4 pulses for pull and push thats 4khz electrical RPM.
What can the Infinion do?
Is the Infinion Schematic online?

I am also interested in what the infinion can do. Kelly advertizes a 40,000 electrical RPM limit on their standard controller, which is 667Hz, which I mentioned will only get the 20pole motor up to 4,000 RPM (I mentioned somewhere around page 30 of this thread I think, so I don;t blame anyone for missing that)

If you want to control the current at 6000 RPM the controller needs to pulse the FETS at 4 times 60 kHz = 240KHz.
(Some might be happy with just 2 times)
Then you need to vary the duty cycle (PWM) of the pulses to give some current control at Max RPM.
Ten steps/levels in throttle twist would mean 2.4MHz PWM Clock.
To protect the phase currents you need very fast feed back.
Switching the FETS at 240KHz (2.167 uSec ON the OFF for 2.167 uSec)
and expecting rise and fall times of 100th that, of 20nSec so Switching efficiency is minimal.

Do my expectations of the controller relate with anyone elses?

I think when you put in the factor of 60 error in your initial calculation the numbers won't be so discouraging.
The FET pulse frequency comes down to 4kHz rather than 240KHz etc.

I still think the Kelly should be a good match for this motor, since I really don't think the motor will perform that well above 4,000 RPM with that skirt bearing and the core loss inherent at such high frequencies. Once the tests have been run I would be glad to be wrong, but until then I will remain skeptical that it will be suitable to operate in that range. If it does go above 4,000 RPM, then there is the high-frequency kelly option availble which in theory could spin this up to 7,000 RPM (if the voltage at that speed is still lower than around 144V) for an extra $80, if I remember the price correctly. The Infinion might be a better / cheapper match for the motor, but as far as I know, it is just sold in kit form so it needs to be assembled and therefore isn't suitable for people who don't have the time / equipment required to put together a project like that. I imagine most people will just want to open the box, connect it and go, which is what the Kelly should do.

-ryan

 

[rolf_w]

...Ah.. The Controller. The Elec RPM or phase frequency needs to be 6000 x 20/2 = 60,000Hz...

isn't ?:
6000rpmx10pole pairs = 60'000rpm_electrical
1Hz=60rpm
=> 60'000rpm_el=60krpm_el=1kHz (PWM)
?

 

[markobetti]

Hey guys, kelly is a good option but not for low rpm, i think infineon can do the same as kelly and in 36 fet version even more. The kit will include finished controller with best mosfet- and thats all that we can do with sensored version . But , today i recived the picture of first (there are two) finished prototye of 75v esc with water cooling, it will be sent today from China . I will post detail sepcs of this esc and then 7circle and Burtie - you can check out what we ve made. I would be happy with your opinion

 

[Arlo1]

...Ah.. The Controller. The Elec RPM or phase frequency needs to be 6000 x 20/2 = 60,000Hz...

isn't ?:
6000rpmx10pole pairs = 60'000rpm_electrical
1Hz=60rpm
=> 60'000rpm_el=60krpm_el=1kHz (PWM)
?

1Hz= 60rps times per second not minute

 

[markobetti]

the esc comes with usb linker and has High rate adjustable switching (PWM:8KHz/12KHz/16KHz) while we are at khz

 

[rolf_w]

1Hz= 60rps times per second not minute

1Hz (=1/s) = 1rps = 60rpm (=60/60s = 60/1min)

 

[liveforphysics]

Hey guys, kelly is a good option but not for low rpm, i think infineon can do the same as kelly and in 36 fet version even more. The kit will include finished controller with best mosfet- and thats all that we can do with sensored version . But , today i recived the picture of first (there are two) finished prototye of 75v esc with water cooling, it will be sent today from China . I will post detail sepcs of this esc and then 7circle and Burtie - you can check out what we ve made. I would be happy with your opinion

Make sure to try and see how a sample performs under-load before buying more than a sample or two.