Ironless axial flux Halbach project - 170mm, 36 pole, 44mm bcd disc brake mounting

[safe]

Go read the laws your preach carefully says nothing about rear wheel measurement distinctly states motor output not rear wheel output.

Motor Output is not controller Input.

1000 watts Input to a motor of 75% efficiency produces 750 watts Output.

------------------------------

I'm 100% correct on this... horsepower in a car is not measured by fuel potential power, but by what mechanical output can be created. Some motors are more efficient than others.

Go read wikipedia:

http://en.wikipedia.org/wiki/Horsepower

...there are various ways to measure output (shaft, rear wheel, or theoretical) but no one measures the Input side.

The 1000 Watt Rule makes sense because it applies to the Input side while the 750 Watt Law applies to the Output side.

It's the perfect compromise between track and street.

 

[madact]

Motor Output is not controller Input..

Correct. But different jusrisdictions have different requirements of where 'power' is measured. In your case, I'm guessing the legislation calls for <750W at the rear wheel?

In my case, the legislation very clearly and specifically calls for <200W at the motor shaft - which in my case is power at the rear wheel + power dissipated by rolling resistance in the tyre (which is pretty negligible).

We both get to feed whatever energy we like into our motor as long as it delivers no more than the specified power. But I'm confident there are places where the law restricts power drawn from the battery - 'the law' doesn't have to be sensible, it just has to be legislated and enforced. Heck, you can even legally define PI to be 3 (or was it 4?) if you can sneak it past both houses without some pesky mathematics professor pointing out the inherent problems of doing do...

I think seeing e-bicycle are not yet popular in Australia we will get away with
flying under the radar if we ride sensibly. The 49cc motors are ilegal yet there is many members i have seen from Australia on this site that have had no issues and their bikes are much more distint than a hub motored e-bicycle.

Common sense prevails, thankfully P

For myself, even if I were so inclined, my e-bike will look as far from a bicycle as a Zonda does from a steam-tractor, anyhow. I fully expect to be chatting (amicably, I do hope) to every traffic cop who passes by on the road, and a few others besides - opportunity for a bit of e-bike-vangelism, but not a position I want to put myself in with something that's even slightly grey from a legal standpoint. My position will be, if asked:
A) Yes, it is under 200W
B) Sure, you can dyno it if you like
C) Would you like a go?
I'd actually consider getting it tested and certified, but I'm not sure where I'd have that done, or how much it would cost.

More importantly how are things with your motor? and have you had a chance to checkout Endless Sphere? i really think you will be impressed by the information available there, it was very helpful in my e-bicycle choice.

Waiting on magnets and brake rotors, playing with octave, trying to fit everything in to my sillily busy schedule (yes, sillily is a word, I just made it up, but I'm proabably not the first to think of it).

Endless Sphere has some good info, not much I didn't already know from Electrical Machines at uni though - a fair bit to wade through, too. Most of it covers retail systems that aren't suitable for my speed range, though chain-drive systems would be applicable, and better for hill climbing than my design (which is limited by both heat and torque in that application). Still going to see how far I get though, nothing comes close to this for efficiency when cruising at low power on the flat.

Some of the pdfs linked on Endless Sphere are directly applicable, too, though I do have a couple of unanswered questions
- Can I afford to use carbon fibre board in my stator, or will the eddy currents (which I've learnt travel elliptically due to the weave topology, but not gow large they are) kill my efficiency and/or cook the motor? (Currently I'm planning on fibreglass, PCB-style board)
- Can I really ignore the end effects at the inner and outer edges of the rotor, or can I apply some known empirical or theoretical formula to this, or do I have to go 3D? (Papers on axial flux seem to ignore this, but don't give good justification AFAIKT)

If anyone has an answer to these two questions, I'm all ears (figuratively speaking )

 

[safe]

Correct. But different jusrisdictions have different requirements of where 'power' is measured. In your case, I'm guessing the legislation calls for <750W at the rear wheel?

In my case, the legislation very clearly and specifically calls for <200W at the motor shaft - which in my case is power at the rear wheel + power dissipated by rolling resistance in the tyre (which is pretty negligible)

I forget the exact language, but it's something like "propel the bike with not more than 750 watts of power". It's clearly following along normal horsepower measurement techniques that are used elsewhere.

Another thing....

Manufacturers are allowed to "cheat" a little because they "rate" their motors at a certain rpm and amperage. A motor that is "rated" for 750 watts at their "rated load" might actually in reality produce more than that (by about 25%) at lower rpms. The trick is that heat is the defining charactoristic of electric motors, so sustaining illegal levels of power for a long time would destroy the motor.

There is some "flexibility" in the law in the way it's enforced, but the general idea seems to be that more than one horsepower violates the concept of a "motor assisted" bicycle and pushes it into the moped classification. (I totally support that wisdom and have used it as a guiding philosophy for the racing side of all this)

- Can I afford to use carbon fibre board in my stator, or will the eddy currents (which I've learnt travel elliptically due to the weave topology, but not gow large they are) kill my efficiency and/or cook the motor? (Currently I'm planning on fibreglass, PCB-style board)
- Can I really ignore the end effects at the inner and outer edges of the rotor, or can I apply some known empirical or theoretical formula to this, or do I have to go 3D? (Papers on axial flux seem to ignore this, but don't give good justification AFAIKT)

I've heard that carbon fibre will act like wire and should not be used near the magnets, but I don't have solid information beyond that. My guess is that a small gap of non-metallic material separating the carbon fibre portion from the magnets would be enough to avoid problems. The magnetic flux falls off within about 1/4" so that's all the gap you would need.

End effects?

I'm of the "hope" that the FEMM Simulation properly adjusts for end effects before giving it's result. There is a "Depth" setting in the setup area so before you run the Simulations you need to have that set at 1/2" of Depth. For about a month I was using the default 1" Depth setting in the Simulation and thinking that I was going to get this incredibly awesome motor really easily, but once I realized that error I had to really work to get a motor that looked any good. Hope you have that setting right.

(mine is set to 0.25")

 

[pedalless]

Correct. But different jusrisdictions have different requirements of where 'power' is measured. In your case, I'm guessing the legislation calls for <750W at the rear wheel

There legislation is the same as Australian worded very similar, "an electric motor of less than 750W of power" Is exact ly what Federal US ebike law says. Entire wording-

"Federal Laws and Regulation
[edit] Defined

The U.S. NHTSA Code of Motor Vehicle Safety simply defines low-speed electric bicycles as consumer products and not Motor Vehicles for safety standards.[22] In doing so they vest authority over commercial safety standards to the Consumer Product Safety Commission.

The Consumer Product Safety Commission(CPSC) stipulates that commercially manufactured low-speed electric bicycles, or tricycles, must have fully operable pedals, an electric motor of less than 750W of power and a top motor-powered speed not in excess of 20 miles per hour (32 km/h) with a rider weighing 150 pounds.[23] An electric bike remaining within these specifications will be regarded simply as a bicycle for purposes of safety standards. This supersedes any state law that is more stringent, but only regarding safety equipment required on electric bicycles and the standard of manufacture they must meet.[24] The legislation enacting this amendment to the CPSC is also known as HR 727.[25]

No known federal regulations apply to the manufacture of homebuilt electric bicycles."

safe is the only one i have seen that doesn't accept this and as said before no use debating with this man just look at the other threads here that are his the mans very unique let us say haha. This very irrelevant to your thread though i think we should be sticking to your halbach motor not US laws that mean nothing to us.

Yes Endless Sphere is packed with much information, if you make this worklog there you will get the answers your after.

 

[madact]

This very irrelevant to your thread though i think we should be sticking to your halbach motor not US laws that mean nothing to us.

Eh, he could just be trying to maintain implausible deniability ("but yer 'oner, taint that ways 'tall!")

Yes Endless Sphere is packed with much information, if you make this worklog there you will get the answers your after.

I think I might - we're all members on both anyway, right? I'll get my OctaveFEMM stuff up and running, then make a summary and start a new thread.

I forget the exact language, but it's something like "propel the bike with not more than 750 watts of power".

My reading of the legislation pedalless posted is that the electric motor must be of less than 750W of power (cleverly gleaned from the phrase "an electric motor of less than 750W of power" ). In fact, this is even more open to interpretation by a court than the Aussie law - ours says "delivers a maximum of", yours could be taken by a court to mean "will handle an electrical input of", should it come up and they decide to throw the book at you. Not saying that your interpretation is wrong re. electrical input vs mechanical output, just that a judge applying that law is not bound to see things your way, as the law is not explicit about whether "750W of power" is power used or power delivered, and the appeals process can be expensive.

Manufacturers are allowed to "cheat" a little because they "rate" their motors at a certain rpm and amperage. A motor that is "rated" for 750 watts at their "rated load" might actually in reality produce more than that (by about 25%) at lower rpms.

Here in Australia, the law is very clear that it is peak power output, not continuous or rated power output, that is limited. I was hoping for a higher 'peak' allowance, but apparently that's not the case.

The magnetic flux falls off within about 1/4" so that's all the gap you would need.

Well, I was talking about the stator core (that is, the rigid disk to which the coils are glued / wound around), which by definition is smack in the middle of the rotor gap, where the field is still > 0.8T. I guess one way to test it would be to but a peice of carbon fibre, and when I have my rotor assembled, chuck it in the airgap, spin it with a model motor, and measure the load on the motor with a true RMS power meter - if the losses are << 2W, I might consider it. I've discovered that the big boys in the academic world run these kind of machines at up to 85,000 rpm, and they use litz wire, not just to allow higher frequency driving currents, but to reduce eddy current losses in their stator windings (!). The operation range of this motor means eddy currents in my conductors aren't a problem, but eddy currents in the stator core are an unanswered question.

End effects?

I'm of the "hope" that the FEMM Simulation properly adjusts for end effects before giving it's result. There is a "Depth" setting in the setup area so before you run the Simulations you need to have that set at 1/2" of Depth. For about a month I was using the default 1" Depth setting in the Simulation and thinking that I was going to get this incredibly awesome motor really easily, but once I realized that error I had to really work to get a motor that looked any good. Hope you have that setting right. (mine is set to 0.25")

Sorry to break it to you, but unfortunately it doesn't. Try a simple magnetic flux measurement with 'Depth' set to 0.001, and then repeat with depth set to 1000 - you get identical flux density results in the gap (everywhere, in fact). This is clearly not adjusted for end-effects, which should be significantly different between an 83 foot deep scenario and a 1-thou-thick 'magnetic foil' scenario

This could be a bit of a downer for you with the cube magnets... I'm fudging in a 20% average drop in flux density for mine in at the moment, but that's just a number I pulled out of my... creative imagination

 

[Toddmoto]

physical measurment of Gauze strenght

Back OT. (laws schmause)
I have been following along looking & waiting for some real world data regarding gauze strength of the halbauch array using off the shelf magnets to configure the effect.

Specificly in comparison to a standard magnet with an iron backing to controll flux leakage.

as an amature motor builder I researched heavily into the halbach array & never found a senario that looked good enough (better said simply enough) to attempt.(personaly)

I am an advocate of the axial flux topology & am very anxious to see one of these running.
My current build is a bit smaller scale than this example.

There is a challange issued for hobby motor builders at ES:

Goal:
- Over 4Nm continuous torque per kg of motor weight.
Rules:
- Less than 3kg in weight.
- No energy input other than that to the motor itself.
- Capable of practical use on an electric bike.
the grand prize is a.... kudo
looking foward to your next update.

 

[pedalless]

madact i have found a thread with a pdf file to the proposed new 250watt law in Australia you can download it from this Endless Sphere link i think?

http://www.endless-sphere.com/forums/download/file.php?id=38237

I also thought of one other thing about measurement, there is a company making legal 200watt china motor here in Australia guess where they measure the power output safe? the crankshaft of the motor not the rear wheel. There is dyno printout sent with motor to verify this fact.
It would be very easy to dyno and electric motor the motorcycle engine dynos
would work pretty much as they are.

 

[madact]

The most annoying thing for me in the proposed revision to the law is the 25kph assist limit, which is based on your average joe riding a 10-speed a lot further than most people drive to work - for reference, my best time riding to work was 14.8km in 25 minutes, including at least five sets of traffic lights and 3 annoyingly long waits to do right-handers on busy main roads in peak traffic, and entirely uphill, if only at a small gradient. My average time on that run is 35-40 when I'm out of shape, and for me "in shape" doesn't mean I'm that fit anyway. They don't take into account the breather you get at traffic lights, which makes their estimate way out IMHO, and their 'oh noes this isn't safe' speed is way slower than any cyclist I know rides given the slightest downhill slope (my mum included, and she's over 65).

This would effectively kill any attempts to build a viable electric assisted velomobile for long trips stone-dead (an athlete can keep a velo going at an average speed of over 45 kph for 24 continuous hours - even a couch potato like me could not be expected to drop under 40kph on the flat on a 4-hour ride - the only scenario the motor could be used for on a velo or streamliner under that legislation would be hill-climbing, and 250W isn't much help for that (nor is my motor designed for efficiency at low speeds). While some have lauded the extra 50W as a great advance, it basically means you can't use it unless you aren't moving, or are sitting on a hill that will just make your motor twitch like a dying fly

Thanks for the encouragement Toddmoto.

Magnets and rotors arrived today, thought I might post a pic:

Won't have time for much else this week (will be moving a summary of the OT part of this thread to Endless Sphere on the weekend, but that's about it). I did, however, stop to confirm that the pull force from magnet to steel exceeded the pull between the magnets, which I tested at a 1mm gap - all good .

I also confirmed that lining all the magnets up in a Halbach array will be fun times. I'm still liking the aluminium-spacer idea:having an aluminium ring around the outside edge of the magnets, with little spikes pointing in to "keep em separated" and stop the pole magnets from flipping round and latching onto the Halbach magnets. Might even get away with machining it on the PCB prototyping router at work , it's well within the capabilities of the machine, just have to do it in a few passes... if not, the boss o' the back building will know where I can get it done.

Oh yes, and I spent some time dropping magnets down the gap between a couple of pieces of aluminium angle ... and then out of curiosity down the space between a set of carbon fibre arrow shafts. Aluminium angle: 1, carbon arrows: 0. From this brief test, it looks like carbon fibre might be OK eddy-current-wise, certainly worth me getting a sheet cut to size and measuring it's drag in the gap after I have my rotor together.

Also had a cunning brainwave for stator construction and copper fill - radial slots in the stator would allow for 100% surface coverage plus perhaps an extra 40-50% again inside the stator. The tradeoff here is copper fill vs. strength (bear in mind the whole thing gets slabbed in JB-Weld or similar), and as a bonus, the slots would make any eddy current effects even less noticable, if using carbon fibre.

 

[safe]

Something like this?

 

[madact]

Something like this?
QUOTE]

Indeed, my original plan (which may still be the one I use) is very much like that - with windings crossing to the other side where the pointy bits on that one are. In fact, I considered using ultra-thick copper PCBs and just routing or etching the windings. But I couldn't find anyone selling blank fiberglass PCBs with 1.5mm copper, dagnammit (or at least not in non-production-quantity batches), and I don't like my chances of making my own using fiber-reinforced-plastic and copper sheet - or at least, I don't like my chances of homebuilt ultra-copper not delaminating on me. So wire it is... at least the adhesive can flow around the wires easily, and I don't have to worry too much about bubbles...

The other way in which my design differs from this is that it is dual-layer in the active region, and uses two sets of 3-phase windings, so that for each phase, the winding on the 'top' surface has the current going in the same direction as the winding on the 'bottom' surface. In a CW direction (for instance - 'handedness' is arbitrary) bottom always goes to top on the inner edge, and top always goes to bottom in the outer edge.

So for a single phase, we have the following arrangement with two wave-windings, electrically 180 degrees apart (oops, first revision was wrong way round, corrected now):

[FONT="Courier New"]
--->---. ,---<---. ,--->---. ,---<--- Phase A1
       | |       | |       | |    
       | |       | |       | |    
       | |       | |       | |    
       | |       | |       | |    
       V V       A A       V V    
       | |       | |       | |    
       | |       | |       | |    
       | |       | |       | |    
.--<--._X_.-->--._X_.--<--._X_.-->--. Phase A2
[/FONT]

where V , A , < , > show current direction, X is where the wires cross over, and the rest is wires. Imagine that the neighbouring pairs vertical wires are directly above/below eachother, and you have it - they are drawn slightly offset here for illustrative purposes. This arrangement confines the generated flux to the area where the magnets are generating flux, i.e. the axle and bearings of the motor do not take part in the magnetic field generated by the windings.

Now as for the slots, this is what I mean by slots and more copper fill, comparing my original idea with the 'slotted core' idea:

[FONT="Courier New"]

(Unslotted stator core:)
   ____   ____   ____
 _|_A1_|_|_B1_|_|_C1_|_
          Cor               <- fiberglass or carbon fibre board core
 ______________________
  |_A2_| |_B2_| |_C2_| 

(Slotted stator core:)
   ____   ____   ____
 _| A1 |_| B1 |_| C1 |_
Cor|__|Cor|__|Cor|__|Cor <- fiberglass or carbon fibre board core
 __|  |___|  |___|  |__
  |_A2_| |_B2_| |_C2_|

[/FONT]