KDX Lighting Stator Rewind How-To

[SS109]

So, you used 16ga wire?

[Slick_Nick]

Yeah, 16, 17, 18, I honestly can't remember. 16 sticks out, but I didn't keep the old spool. I'd know it if I saw it, sorry.

[SS109]

Looking at all the different magnet wire out there, I have a couple of questions for the electrical engineers.

Which type should be used, single or double (heavy) coating for our purpose?
I see there is different colors, red, green, and natural copper. Does the color matter?

[TWMOODY]

Color is for the artist's or where color coding is
important and otherwise meaningless so use
whatever color you like.
I would use the single coated wire (to gain the most turns or wraps)
then cover in a high temp epoxy which would make the assy
good for about 310 degrees.


EDIT
I would even consider removing 4-6 of the core plates to
gain the extra turns, then space the coil the equal thickness
of the inside plates removed from the backing plate if that makes
any sense to you.
The more turns = more output

[SS109]

So, the core size has no effect on output? If so, how would one go about removing a few of the plates?

[TWMOODY]

I have not had a coil apart on a KDX however by looking at the
posted pictures it looks like the core plates are spot welded together
at the ends and all you would do is drill the depth of the plates you
are removing.

The largest benefit of removing core plates and adding "turns"
will be increasing voltage of the unit and a slight improvement
in output.

For instance on Nick's rewind and his last good tests he shows
11.3volts/7.33amps/83watts at idle.
This tweak very well may bring the idle voltage up to or exceeding
12volts with the same or slightly better amperage.

[chkdx]

Ummm, I'm gonna have to disagree with Moody on this one. Too few core plates will cause the core to "magnetically saturate", and output won't go up, and the wire will run hotter, maybe a lot hotter.

Saturation is more likely to occur with fewer core plates or with more windings. Since we're adding windings with this mod, the existing core is already being pushed harder than Kawasaki's engineers intended. I'm sure they built a safety factor into the core; use of some of that safety factor is what allows this mod to work. The question is how close to saturation is the core getting?

To sum up, I doubt adding windings will cause a problem, I think removing core plates to add more windings (a double whammy towards saturation) definitely could cause a problem.

[Slick_Nick]

While it's a good idea, removing plates will NOT give more room for extra wraps.

The clearance issue we run into is on the "Y" dimension of the stator, when looking at it head on, bolted to the bike. Removing plates would allow more wraps, but the coil will become too thick and hit the inside of the flywheel. Adding plates, however would allow more wraps, and expand the coil in the "Z" direction, where there is a bit more room, and while more wraps might not then be possible, each wrap is then slightly longer.

Trust me when I say it's tight under there! I originally wrapped mine too much, and had to unwind wrap by wrap until it just cleared the flywheel.

[kawagumby]

What a great thread...modifying bikes is great fun, one of the few machines we can truly personalize anymore.

Just for discussion sake, a thought based on the pictures I'm seeing here, it appears that the number of turns hasn't actually increased due to larger diameter wire size used vs stock guage? It looks like the stock unit is coiled with smaller diameter wire, which means it had more turns originally (it appears that the final coil diameter is very similar to the original from the pictures shown). If so, the new coil has less resistance and perhaps less peak voltage without regulation, but more current capacity, no? Since max voltage is regulated anyway, this setup probably matches the real-world output demands of the heavier-draw lighting system desired at higher rpms, but total "unregulated" wattage wouldn't be that different, as the "unregulated" voltage would be higher, no? If the number of turns has actually increased then all the above is BS. LOL.

Also, I was taught that the plates aid flux allignment and are used for more efficient current induction, so wouldn't there be a trade-off, maybe even a loss if more plates were removed?

[TWMOODY]

Ummm, I'm gonna have to disagree with Moody on this one. Too few core plates will cause the core to "magnetically saturate", and output won't go up, and the wire will run hotter, maybe a lot hotter.

Saturation is more likely to occur with fewer core plates or with more windings. Since we're adding windings with this mod, the existing core is already being pushed harder than Kawasaki's engineers intended. I'm sure they built a safety factor into the core; use of some of that safety factor is what allows this mod to work. The question is how close to saturation is the core getting?

To sum up, I doubt adding windings will cause a problem, I think removing core plates to add more windings (a double whammy towards saturation) definitely could cause a problem.

Nick has said there is not clearance between the coil and the flywheel to add more turns which kinda puts the binders on this possibility with
the exception of physically grinding the Y area.

However taking into consideration that with this mod there are several LESS turns due to the increase in wire diameter I'm not so sure saturation would be any issue.

I would agree with more heat as demand increased though.


I guess if I had the extra time to try this out it would be worth a shot.
Trial and error is what it's all about right ?

[chkdx]

Although fewer turns implies less saturation risk, it really comes down to output power.

Greater output power = greater magnetic flux density = closer to core saturation.

Since Nick's tests showed greater output power, the flux density has to higher.

[chkdx]

BTW Moody, you're right. The only way to really optimize this is by trial and error, ideally using an oscilloscope to monitor the output waveform under varying loads and rpm to check for saturation. That'd be a bunch of work!

Even with a bunch of time spent optimizing it, I doubt the end result would be more than a few percent better than what Nick's done. I'd be inclined to copy what he did and go riding!

[TWMOODY]

BTW Moody, you're right. The only way to really optimize this is by trial and error, ideally using an oscilloscope to monitor the output waveform under varying loads and rpm to check for saturation. That'd be a bunch of work!

Even with a bunch of time spent optimizing it, I doubt the end result would be more than a few percent better than what Nick's done. I'd be inclined to copy what he did and go riding!

I'm with you on that one!
I never ride at night when a light is necessary anyhow.

[Slick_Nick]

I did notice a big difference in the magnetic field strength compared to the stocker. Loaded up with 110w of bulbs, flick the switch and listen to the idle drop by a few hundred rpm. That's how much drag was placed on the engine by the strength of the field. The stocker didn't do that.

While the number of turns was reduced, using a thicker core wire will output more available current, but less voltage. I don't know that the 16awg I used is the majic number, it might be 18 or 20awg, depends what you want. Voltage (which is regulated down anyway) or current capacity. However, more turns with a finer wire would output higher voltage at a lower rpm than my coil, at sufficient rpm, it would have nowhere near the current capacity.

[chkdx]

I did notice a big difference in the magnetic field strength compared to the stocker. Loaded up with 110w of bulbs, flick the switch and listen to the idle drop by a few hundred rpm. That's how much drag was placed on the engine by the strength of the field. The stocker didn't do that.

While the number of turns was reduced, using a thicker core wire will output more available current, but less voltage. I don't know that the 16awg I used is the majic number, it might be 18 or 20awg, depends what you want. Voltage (which is regulated down anyway) or current capacity. However, more turns with a finer wire would output higher voltage at a lower rpm than my coil, at sufficient rpm, it would have nowhere near the current capacity.

While the number of turns was reduced, using a thicker core wire will output more available current, but less voltage. I don't know that the 16awg I used is the majic number, it might be 18 or 20awg, depends what you want. Voltage (which is regulated down anyway) or current capacity. However, more turns with a finer wire would output higher voltage at a lower rpm than my coil, at sufficient rpm, it would have nowhere near the current capacity.

You're 100% right, Nick

[chkdx]

Oops, sorry for the double quote thing there.

[SS109]

This is some great info. Thanks, Nick, for taking the plunge and doing the write up and the subsequent testing to verify your results. Thanks to all that have contributed to one of the best technical threads ever here on the forum. I'm sure this thread will be referenced for many years to come!

Anyways, I think I'm armed with just enough knowledge now to be dangerous so I'm going to give it a shot myself!

[rbates9]

BTW Moody, you're right. The only way to really optimize this is by trial and error, ideally using an oscilloscope to monitor the output waveform under varying loads and rpm to check for saturation. That'd be a bunch of work!

Even with a bunch of time spent optimizing it, I doubt the end result would be more than a few percent better than what Nick's done. I'd be inclined to copy what he did and go riding!

I'm with you on that one!
I never ride at night when a light is necessary anyhow.

Then why did you need exact tenth of a watt output?

[SS109]

Just so everyone knows for sure what Nick used, I found this while reading all the old threads on stator rewinding.

Anyway, I simply removed all the stock wire, and rewound with 18GA high quality magnet wire that I got at an electronics store. Keep all the winds tight, and uniform, put as many as you can fit under the flywheel. Scrape some coating off the ends before you solder them onto the stock location. Check for continuity with a multi-meter before re-assembly. Not sure if the direction of the wind makes a difference, I went the same way as stock to be safe.

This entire mod cost me only $13! Mr. Green All that's needed was the high quality spool of magnet wire. Wire quality does make a difference. Each "layer" used roughly 8-9ft of wire, and a 1/4lb spool did the entire coil with lots left over. Took me about an hour to unwind, rewind, and solder everything back up. If you choose to epoxy it, make sure there's room under the flywheel.

Found here http://kdxrider.net/forums/viewtopic.ph ... &start=120 about half way down the page. So, I'll be ordering some 18ga wire!

[Slick_Nick]

Ah ha! There it is! Good find! I never even thought to go back to my original post! :D