Miles' DIY motor thread

Miles said:
You're very kind. I was thinking to use 0.25mm thickness material, though - that's 50 different shapes..... :shock: There would still be a limit set by the tool radius.
Click to expand...


0.25mm material should cut like butter with a little carbide end mill. With each sheet holding a hundred+ lams, all the work would just be in creating the program. On my side, it would just be sticking down a sheet to something flat in the machine, loading the program, and leaving for work. :) Might take a few days to cut them all, but I would be glad to help if I can. The last thing I want to have happen would be to NOT get some wear on this machine. :)
 
HAL9000v2.0 said:
I can generate you NC code directly in ProE.
Click to expand...


I don't know jack about it, but if you can write code that I can just click to load in Mach3, tell me where to position the tool to start, and hit run, I will gladly cut what ever you like.

I've got 1 day of experience running my CNC now, so I'm about as newb as it comes still. :oops:
 
liveforphysics said:
Anything on a mill will always be limited by the needed diameter of the end mill, but I think a 1.5-2mm radius on the corners wouldn't be such a bad thing.
Click to expand...

For a round wire winding, maybe - for a tape winding it wouldn't be so good....
 

Attachments

  • Core one piece radiused.jpg
    Core one piece radiused.jpg
    54.8 KB · Views: 2,239
Once that core is made,
I see simple drilling & squareing the opening. That would get you the profile as drawn. (just need more time :p )
We have seen what Miles can do with a proper file.
 
Thud said:
Once that core is made,
I see simple drilling & squareing the opening. That would get you the profile as drawn. (just need more time :p )
We have seen what Miles can do with a proper file.
Click to expand...


I agree Thud, if you're only making 1 unit, it wouldn't be too tough to square off the corners by hand. :)
 
Looks like there may be some clues in these, regarding the winding/rotor configuration decisions....... :|

http://cmrt.centrale-marseille.fr/cpi/ever09/documents/papers/ev2/EVER09-paper-138.pdf

http://www.ornl.gov/~webworks/cppr/y2001/rpt/123938.pdf

http://www.smartmotor.biz/downloads/publications/ICEM2006_Skaar_Nilssen.pdf

http://www.elkraft.ntnu.no/norpie/10956873/Final%20Papers/048%20-%20NORPIE_2004_paper_048.pdf

http://www.ee.kth.se/php/modules/publications/reports/2003/IR-EE-EME_2003_029.pdf

http://www.ee.kth.se/php/modules/publications/reports/2004/IR-EE-EME_2004_005.pdf

http://www.cedrat.com/fileadmin/user_upload/cedrat_groupe/Publications/Publications/2008/11/Magnetic_field_armature_winding_PMSM_w_concentraed_coils_on_statorISEM08.pdf
 
liveforphysics said:
I don't know jack about it, but if you can write code that I can just click to load in Mach3, tell me where to position the tool to start, and hit run, I will gladly cut what ever you like.

I've got 1 day of experience running my CNC now, so I'm about as newb as it comes still. :oops:
Click to expand...
Dude your awesome thats what I would do If I had your money. I just love building stuff.
Have you seen this its prety cool! http://www.youtube.com/watch?v=ggvzcGdZsTc
 
That scanner is brilliant! Attention all fabricators on endless-sphere: watch that video. LFP will buy one! $3000 for the scanner and then the 3D printer is magic. One each of those and a casting expert and we're in business. Far out man!
 
For a one-piece post-machined core, 24 slots is starting to look more suitable... Here's what it might look like (4mm cutter for slot).
 

Attachments

  • Core 24 slot one piece.jpg
    Core 24 slot one piece.jpg
    91.1 KB · Views: 2,071
I have both 3mm and 2mm carbide endmills. It could look even a bit better than that, though that really doesn't seem too bad even at a 2mm fillet. :)
 
Miles said:
The one-piece core would look something like this:
(the insides of the slots should be painted, too..).
Click to expand...

that looks really nice. definitely doable (machine solid lamination). may i suggest:

1. turn a round fixture on the lathe that acts as workholding for laminating and machining
2. wrap and glue flat lams continuously around fixture.
3. before glue has solidified, slit thru lams where the rings start to overlap
4. align and glue the slit edges. now the rings are concentric.
5. move chuck and chucked fixture/workpiece to 4th axis
6. cut slots using end-mill that is same diameter as your slot (to reduce passes)
7. square out corners with 1.5mm end-mill (the one i have reaches about 14mm)
8. deburr and you're golden 8)
 
j3tch1u said:
Miles said:
The one-piece core would look something like this:
(the insides of the slots should be painted, too..).
Click to expand...

that looks really nice. definitely doable (machine solid lamination). may i suggest:

1. turn a round fixture on the lathe that acts as workholding for laminating and machining
2. wrap and glue flat lams continuously around fixture.
3. before glue has solidified, slit thru lams where the rings start to overlap
4. align and glue the slit edges. now the rings are concentric.
5. move chuck and chucked fixture/workpiece to 4th axis
6. cut slots using end-mill that is same diameter as your slot (to reduce passes)
7. square out corners with 1.5mm end-mill (the one i have reaches about 14mm)
8. deburr and you're golden 8)
Click to expand...

I'm 100% with you on all of that, except the squaring the corners after the main milling. Wouldn't it be much better to use the 1.5mm endmill to drill down for the needed corner squaring process first, then send in the main cutter?
 
liveforphysics said:
I'm 100% with you on all of that, except the squaring the corners after the main milling. Wouldn't it be much better to use the 1.5mm endmill to drill down for the needed corner squaring process first, then send in the main cutter?
Click to expand...

yes and no. not sure how wide the slots are, but a 1.5mm hole may not be enough to remove all the "rest material". end-mills are not great at centering either (precise holes are usually center-drilled, then drilled). probably don't want more than 2 tool changes in there and also try run the 4th axis in one direction--the worm gears in rotary tables have pretty bad backlash. in my cam software (solidcam), rest material removal is an option. i'm relatively new to cnc myself--it's nice to discuss these things.

one thing though miles: you probably don't want to cut where the ring edges are glued together (though it would be tempting for aesthetics). that is the weakest link in the structure. maybe make the slit the a-axis ref origin. looking forward to seeing the progress. if it works, i will definitely copy your design ad verbatim and call it my own :p
 
The 4th axis I'm getting claims to have zero backlash from using a pretty slick no-gears design. It seems like it works similar to clever motors for auto-focus on a highend camera, only on a much larger scale.

http://www.doughtydrive.com/

It's a neat design for small mill's.

You can see a pair of them here doing some frilly stuff. :)

[youtube]O3d0fueNr30[/youtube]


Something like a 3-fluke 45deg face 1.5mm bit would likely be just the ticket for getting minimal run-out on locating the corners. The the major slot done with the larger bit, and if there is a residual "rest-material", I would assume with the corner all ready handily located, it wouldn't be much more than a few strokes with a good file in Miles capable hands to clean it up nice and neat. :)

But, this is comming from a guy with 4 days total CNC experience now. :oops: :oops: :oops: So, take it worth a grain of salt! I do learn a lot from all you guys though when I share a flawed idea, and you guys point out the mistakes. :) Messing up, or being wrong is likely the best way to learn. :)
 
Ben & Luke,

Thanks for the input!

As modelled the slot is 4mm wide so, as you say, straight passes with a 4mm cutter.

It may, or may not, be worth reducing the corner radius further, depending on the gauge of wire used.

I'm not sure that it would matter that the rings weren't concentric? The lam material would only be 0.25mm thick.... Interesting idea, though.....

I was thinking to make a winding/mounting inner boss to match the spiral form.
 
I think I'll pursue both designs. The original one, using copper tape, in a concentrated winding scheme and this one, using wire, in a distributed winding scheme.
 
Why an axial flux single stator/single rotor?

Advantages:

Good form factor for ebike use.

Torque levels should be similar to an RF outrunner of the same case diameter.

Good thermal conductive path to dissipate heat from armature.

Potential convective path around/through armature for passive cooling using airflow from movement of bike.

Potential to add a fan to the shaft to push air radially through armature.

Exchanging rotors, for experimentation, would be easy.

Control of field strength is possible, by axial movement of the rotor.


Disadvantages:

Axial load has to be taken by the bearings.

Rotor structure needs to be very rigid.

A lot of work to fabricate core.

Less torque density than is possible with twin rotor AF designs.
 
I did some rough CADoodling to get an idea of the performance effect it would have. Looks like using 0.5mm strands, and doing all 4 corners of each slot might be good for 2-3 more laps of wire. Doesn't seem like much of a deal winner or breaker either way you look at it, but perhaps a tiny bit better efficiency. That 0.3mm humps left behind should be able to be cleaned up with a stroke or two of a file.

cornercut.png
 
You must log in or register to reply here.

Similar threads

Mihai_F
Motor Winding Calculations & Cooling Design
Replies
9
Views
1,189
neptronix
neptronix
Mihai_F
Selling 22Kw motor
Replies
1
Views
463
Mihai_F
Mihai_F
B
DD alternative to mac
Replies
9
Views
544
E-HP
E-HP
H
Boost converter recommendations, 120VDC
Replies
2
Views
194
harrisonpatm
H
E-HP
    • Like
Statorade maintenance, time vs miles
Replies
14
Views
777
fechter
fechter
Back
Top