The Halbach Disc Motor Ebike Project

[safe]

Condensed Version of this Thread

Over on this site:

http://www.ebikehub.com/forum/viewtopic.php?f=4&t=2452&sid=c425f06c530db93310021b40a3990153

...I will take the best of this thread and produce a "condensed version" without the endless banter that clogs up these threads. That way people that are frustrated by wading through pages and pages of troll material can get a "clean read" of the topic.

This will allow me to relax a little more here because I will know that the "clean read" will not get messed up.

 

[spad4me]

Safe I just want to present some power supplies for project dither and wait.

AKA your second linking to a third or fourth rehash.of your Halbach post.

Here are a few ideas for you.


Hans Choler was a early pioneer of a unique power supply.
http://users.rcn.com/zap.dnai/coler.htm
This device requires a radio isotope to run.
The rules for the ERBB may have to be amended to allow it..
The second device uses negative ion energy to push the magnetic flux past your one inch rotor.gap.
http://jnaudin.free.fr/html/vta.htm

And finally a how to for the VTA.

http://www.hyiq.org/Library/Floyd_Sweet-Archive.html

 

[safe]

On the other site the owner takes a dim view of trollers and spammers so if people post they can be deleted. I've had stuff deleted there before.

Not true here... which is a mistake in policy I think.

All I can do to defend myself is continue to report offensive posts.

It's sad really that this happens... yet another thread being destroyed by trolls...

You do get the message right?

Spad4me it would be fine with me if you NEVER post on my threads ever again.

 

[safe]

Back On Topic

...the thread is about Halbach motors being designed and built for ebikes.

The project will take 3-4 months (at least) to finish.

Those are the facts...

 

[safe]

Correcting The Alignment

There was a 50-50 chance that I'd get the wrong alignment for the Halbach array when in an axial configuration. Sure enough I got it backwards. The results having to do with the magnetic flux intensity are the same, but the lines are different.

Better to get it wrong in the simulation than on the bike.

(this reinforces the need to carefully study the topic before doing anything)

 

[safe]

"One Layer" or "Overlapped?"

I've been scratching my head trying to understand how the coils were supposed to be set up. Apparently there is more than one answer to the problem. You can either overlap the phase wires or not, but if you don't overlap it means that things are more spread out and would effect the motors base speed.

I'm finally able to understand how these flux lines are actually supposed to work. Seems that you orient yourself so that everything (from a top of the disc view) is oriented from the North (up) or South (down) direction. The coils are also oriented so that they are a sequence of loops that generate forces in the same direction. (N-S) The timing of the phases is such that you are constantly trying to get things to line up and it's the degree of difference from alignment that defines the torque.

Not sure which way to go yet...

"One Layer" certainly looks easier to do and would allow more cooper in each bundle. Laying this out on a piece of plywood would be easy because you just stick nails into the wood and then loop around them. Once complete you would cover with fiberglass resin.

"Overlapped" would likely mean that the motor would spin more slowly and that might effect low end torque.

We can summarize by saying that:

"Overlapped" spacing goes N (skip S) N (skip S)...

"One Layer" spacing goes N (skip S-N-S) N (skip S-N-S)...

...so you are in effect doubling the speed with "One Layer".

 

[safe]

Nice Weather Allows "Action" (people seem to like that)

Chainring mounted into drill with special tool I made:

Chainring drilled out and teeth removed:

First layer of fiberglass:

Just getting started on building up the shape: (much still to do)

...yesterday it was 40 degrees, but today it was 60 degrees and tomorrow looks good again before it goes back to cold. We will see what can be done tomorrow.

180 mm Disc "Just As Good"

If anyone is thinking of doing this you might just use a 180 mm brake disc instead of the 48 tooth chainring. The net result and amount of work is the same as far as the fiberglass and you get to skip the building of the spider, though that's not a big deal. My hope is that this idea works well enough that I can convert my old bike to the same thing and I have a badly worn 180 mm disc that is just perfect for the task. A "freebie".

I liked the idea of the aluminum chainring because it eliminates any worries about magnetic interactions with the motor, but I suspect that worry is not actually a big deal. (but I don't know)

 

[safe]

More Progress

Flat base has been added:

Beginnings of the sides:

...next "weather window" is looking to be tuesday, but at this time of year any good day is a gift. (it's supposed to be 40 degrees)

 

[johnrobholmes]

If your magnets are going to be on the aluminum piece, no worries. If the coils are going to be on this disc, you will get eddy currents in it from the polarity switching.

 

[safe]

If your magnets are going to be on the aluminum piece, no worries. If the coils are going to be on this disc, you will get eddy currents in it from the polarity switching.

The disc is actually used as if it were two discs:

It's a unusual design because "normally" the rotor is in the middle.

For this one the rotor is on the "outside" as attaches to the frame.

(and will cover less than half of the circumference of the disc)

Litz wire reduces eddy currents to nearly nothing... the CSIRO motor gets 97% efficiency so it seems to work pretty well. But eddy currents in the copper are small anyway... the CSIRO motor estimates that they account for something like 1% at most. Compared to iron losses (which tend to be massive) the copper losses due to eddy currents are almost non-existant.

But Litz wire does make sense... "why not?" right?