Stators demystified

I hope!


It seems that a lot of people want more electrical power on their bikes. The standard procedure is to rewind the stator with thicker wire and take a guess about the number of turns. I am an electrical engineer, but had very little knowledge of electromechanics (motors and generators). I did some reading on how stators actually work, made a few measurements on my own bike and came up with some basic formulas to calculate stator output over rpms for various winding configurations.

Here is what I've learned. I think it applies to all permanent magnet generators (both 2 and 3 phase) like in most motorbikes, but not to generators with a field winding like in a car or a BMW R or K bike.

I'm going to start out using the Honda XR as an example, because that's the bike I've done a bunch of measurements on. I'm going to move to the LC4 because I just bought one, and because there is some very useful information already available about it.

I'm going to start with a very basic electrical model of a stator. By model, I mean a mathmatical description of its behavior. The first rule of models is this: "All models are wrong, but some are useful". By making the model more complicated we could get a better idea of how the generator will work, but I think that this model is pretty close to reality and I want to keep things simple. The calculations and graphs are included in this spreadsheet:
<stator_rewind.xls></stator_rewind.xls>stator_rewinding.xls
Start by changing the parameters in pink to see how the results change.


A stator behaves like an ideal voltage source with a series resistance. The voltage is proportional to the engine RPM and the number of turns in the winding, the resistance is proportional to the engine RPM and the square of the number of turns in the winding.
This is equivilent to a circuit consisting of a current source in parallel with a resistance.


<stator_equivilent.png>

In the diagram above, the 'Thevenin equivilent' is the voltage source version and the 'Norton equivilent' is the current source version. The Value of V is A * engine_speed * number_of_windings, R is B * engine_speed * number_of_windings^2, and I is A / ( B * number_of_windings). A and B are constants that depend on the properties of the stator core and the flywheel magnets.
Note that R is not the winding resistance- it actually comes from the resistance in the magnetic circuit that is formed by the flywheel magnets, the stator core, and the little gap of air between them. The winding resistance is small compared to this, so I am lumping it into "regulator and wiring losses" which I will go into later. Remember: "All models are wrong...".


In the case of the XR at idle (1200rpm) V is 30V, I is 6A, and R is 5ohms.
This means that if you let the motor idle with the stator disconnected and attach a voltmeter to the stator connector it should read 30 volts- this is called the open-circuit voltage. If you short circuit the stator output through an ampmeter it should read 6 Amps- this is called the short-circuit current. If you double the rpms to 2400, the oc voltage should increase to 60V and the sc current should stay the same at 6A.

In practice, the current will increase slightly as RPM increases- this is due to the winding resistance. Short circuiting the stator does no harm- that's how the regulator works. The stator also acts funny when it is open-circuited- there are large short voltage spikes that show up at the peak of the waveform. Even a very light load (20 ohms) gets rid of them, but they may confuse some multimeters. The behavior with loads between short and open circuit confims the model, though.

Power delivery:
Basic circuit theory says that we will get the maximum power output from the stator when the load is the same resistance as the source. Since the load isn't quite resistive, let's rephrase that to say we get the most power when the load causes the source's output voltage to drop to half the open-circuit voltage. Unfortunatly the source voltage changes with the engine speed. This means we can only optimize for one engine speed. The normal thing to do is optimize for idle, when the least amount of power is available.

For bikes the load is around 15V. That's 13V to charge the battery plus 2V for wiring and rectifier losses. So the XR stator at idle delivers (30V-15V)/5ohms = 3A at idle. The lights need 35W+5W/13V = 3A. It looks like Honda designed the system perfectly. In fact, the stock system has no rectifier, so there is a margin of error built in that is taken up by the aftermarket rectifier in the dualsport kit.

Above idle, the stator puts out more power. At 2400rpm, Voc is 60V and R is 10ohms, so the output into 15V is (60v-15v)/10ohms or 4.5A. At 7200rpm, Voc is 180v, R is 30 ohms so the output is (180v-15v)/30ohms or 5.5A.

For visual folks, here is a graph of the stator output:

<output graph.png="">

For each engine speed, there is a different line of current vs. voltage. To see how much current is delivered, look at the current value where each line crosses 15V. This is how much current the stator will put out at 15V.



The magic numbers:
So where did the OC Voltage, SC current and Resistance numbers come from? In the case of my XR, I disconnected the stator wires, hooked them up to load resistors, measured the voltage, and made a spreadsheet. To get the three parameters you only need to make two measurements at a single rpm (if you go back to the equivilent circuits, it's realy only two parameters: A and B). I made a lot more, but that was to confirm that the stator was behaving linearly.

Measuring the open circuit voltage is not recommended due to the strangeness of the coil's behavior without any load. It is hard to find a current meter that can handle the high currents of a short circuit. Use two different value resistors (Say 1 ohm and 10 ohm) and measure the voltage across them
at a single rpm. Plug the voltage and ohm values into the spreadsheet and it will give you OC voltage and SC current. R is just OCV/SCC.


<stator_ac_estimate.png>

What if you can't run the bike with the stator disconnected? Measure the rectifier output at several different rpms and then choose parameters that would give a similar output. The LC4 iginition is powered by one phase of the lighting coil- unhooking the regulator will result in high voltage hitting the CDI box with potentially bad results. Flanny contacted KTM and got a table showing regulator output vs. rpm. Plugging those numbers into the spreadsheet gives us a good guess of the regulator parameters. The good news is that a curve fits to within 1/2A at every point using reasonable numbers for parameters. The table from KTM was rounded to the nearest 1 amp.

<stator_dc_estimate.png>


Improvements:
With the XR, improving the stator is easy- pull the windings off and rewind using the same number of turns as stock (300) using all 10 of the stator poles instead of just 4. This keeps the OC voltage the same, increases the SC current by 2.5x and reduces the resistance by 2.5x. So now the SC current is 15V and R is 2 ohms at idle. As a bonus, there are now only 30 turns of wire on each pole instead of 75. This allows the use of much fatter wire, lowering the resistance of the windings.

Suppose there weren't an extra 6 poles on the stator waiting for us to use. We could try increasing the number of turns in the windings. How about 20% more, say 360 total. Our OC voltage goes up, now it's 36V at idle. Unfortunatly our SC current goes down to 5A and our resistance goes up to 7.2ohms. So now our output power at idle is (36V-15V)/7.2ohms or 2.9Amps. Oops. At 2400rpm we get 4A. Also no good.

Let's go the other way to the extreme. Cut the number of windings in half to 150. Now we get absolutely nothing at idle. But, we get 6A at 2400 rpm, 8A at 3600rpm. Hello. Stick a battery on this thing and we can now run a 60W bulb (just don't get stuck in traffic).

What are we doing here? By changing the number of windings, we are changing the engine speed where the output power is optimized. With our +20% example, the power is optimized at 1000rpm, where we never run. With our -50% example, the power is optimized at 2400rpm. At lower engine speeds we give up power, but at higher speeds we get much more.

If you look at the 'Estimating output' tab of the spreadsheet, you can adjust the OCV and SCC and see what changing the number of windings will do for output over the rpm range.


<stator_output_estimate.png>

On my XR I underwound the stator slightly- 280 turns total on 10 poles. This reduced the OC voltage a little but gave extra power everywhere above 1500 rpm. It's kinda pointless because I don't need the power at the moment. It is also putting a little extra stress on the regulator. By total coincidence, the output curve for my stator is a dead match for the stock LC4 output. Go figure.

The LC4 stator is 3-phase, which may lend itself to a big change with very little effort. By changing the connection to the windings- without changing the windings themselves it is possible to change the output of the windings. According to the illustration in the manual, the LC4 windings are Y-connected. This means that all three windings have one end connected to a common point and power is taken off the other terminals. The other possibility is a delta connection where both ends of each winding are connected to another winding.


<stator_delta_y.png>

The Delta connection will produce <strike>70%</strike> 60% of the voltage output and </stator_delta_y.png></stator_output_estimate.png></stator_dc_estimate.png></stator_ac_estimate.png></output></stator_equivilent.png><strike>140%</strike> 173%<stator_equivilent.png><output graph.png=""><stator_ac_estimate.png><stator_dc_estimate.png><stator_output_estimate.png><stator_delta_y.png> of the current output of a Y connection with the same windings. Compare the two output lines in the 'Estimated output' tab of the spreadsheet. EDIT: fixed the numbers, thanks Feelers.

Before you think about reconnecting your stators, look at the short-circuit current. That is what the regulator has to shunt to keep the voltage down to a safe level. It goes up from 16A to <strike>22A</strike> 28A. Apparently the LC4 stock regulator can't handle much more current than the stock windings deliver.


To make a long story short, if you plan on rewinding your stator this tool should help you choose the number of turns you want. It will also give an idea of the stress that the new windings will put on the regulator.

One other improvement that I haven't really gone into here: if you upgrade the wiring harness on the bike with heavier wire between the stator and the battery it will effectively lower the load voltage, which will also improve the output of the stator. The effect is greatest at lower rpms. It doesn't give you a lot of current, but it's a relatively easy upgrade.

Questions? Comments?

</stator_delta_y.png></stator_output_estimate.png></stator_dc_estimate.png></stator_ac_estimate.png></output></stator_equivilent.png>

Good gawd ! Dude, if I knew half of what you know, my head would explode !

Uhh, anybody seen my Cheeze Whiz?

I'm all ears for the second installment.

A thousand thanks on behalf of all of us for the effort you put into this post.

Dirty

Superb article, Luke. Thanks so much.

Norm Kern

Nicely done, Luke. Happen to have a tutorial on how you rewound your XR stator w/ jpg's? Or maybe the KTM? I've wondered about this but looked into it before.

Also, would you mind touching on why the factory wouldn't have initially gone w/ a wye configuration rather than delta wrt the KTM's stator? Particularly if such a simple reconfiguration seems to make sense from an upgrade standpoint.

Also, what would you estimate the voltage drop to be on a typical circuit w/ typical existing conductor and how much drop would you get on the same circuit if you dropped a gauge or two (say, 16 to 14 or even 12)?

Hi Luke ,
Thank you for taking the time to explain rewinds so well. Quality info is Gold when undertaking this rework.
<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>
My XR rewind is giving me a bit less power (although adequate) than expected.
I am 100% mechanical (designer) and have limited multimeter skills.
<o></o>
I have two windings, each are 5 poles with a total of 330 turns of 18ga copper. This was supposed to produce 125Watts per winding at ? rpm.
<o></o>
OUTPUT 1 just replaces the stock windings and runs to the stock XR regulator. From there it goes to the headlight with an H3 90watt bulb. The light output while not exceptional is much better than stock, (this is an AC circuit)
<o></o>
OUTPUT 2 goes to a Baja Designs regulator-rectifier and a 1.2 Ah battery. This circuit powers the brake light, tail light and turn signals (all LED).
This output is also connected to two 55 watt lights . When only one light is on all is good. When the second one is on the bulbs quickly go “dark orange”
I know I am over loading the circuit.
<o></o>
My question is: Is the stator expected to output 2X 125Watts at the present winding?
Should I upgrade the stock XR regulator and or the Baja designs Reg/rect for performance or reliability?
<o></o>
Thanks again for enlightening the collective!
<o></o>
Cheers!
Eduardo
XR650R

ps. wire gauges are generous and all connections are soldered or gold plated Amp connectors)

I didn't take any pictures, unfortunately. Here is one web article that shows the mechanical side of rewinding. I seem to remember one better one, but I can't find it now. I would recommend against adding to the stock windings- just pull them off and start fresh. The most important thing is that you don't mess up the ignition coil windings. Next most important is that you remember to alternate the winding direction on alternate poles. If you don't do this, the windings will cancel each other instead of adding.

I haven't looked into the mechanics of rewinding a 3-phase stator, and I won't until my bike runs out of power. I am sure that a more powerful stator on the 640 will require a better regulator. My first attempt at building a regulator (for the XR) was a learning experience.

I don't actually know that the ktm uses a Y winding. The picture in the manual shows a Y winding, but that doesn't mean it's so. Starting with a Y winding does make for an easier upgrade path for us hacks, but for a factory upgrade they can just specify a different number of turns.

KTM seems to have chosen a really high open circuit voltage, but it actually makes sense. The stator puts out enough power at low rpms but the power doesn't increase much at higher rpms, keeping stress off the regulator and battery. It also provides higher voltage when starting, which will make the CDI happy.

I don't know, but here's a start: The resistances of 20,16,12 ga wire are 10,4,1.5 milliohms/foot, respectively. So for a 3 foot run at 10 amps, the voltage drop is 3ft x 2ways x r x 10amps or .6,.24,.09 volts, respectively.

EduardoMas;

Based on the numbers you gave me I figure that your stator windings have an OC voltage of 33@1200rpm and SC current of 6.8A. Here's how I got these numbers: Stock windings are 300 turns and 30 OCV, so it's 10 turns per volt- 330 turns / 10tpv = 33volts. Current is a bit trickier. If you had 300 turns and rewound on 5 poles instead of the stock 4, the SC current would go to 6A * 5 / 4, or 7.5A. Because you went to 330 turns, the SC current drops to 7.5A/1.1, or 6.8.

Here is a graph of the stock output (orange), 300 turns on 5 poles (blue) 330 turns on 5 poles (red), and 270 turns on 5 poles (grey)



So according to the tea lea...uh...spreadsheet, your stator should give you in the neighborhood of 5 or 6 amps (70ish Watts) of lighting. This seems to agree with what you are seeing. You can get another .5 or 1 amp out by rewinding the stator with 300 or 270 turns.

The fact that the lights go dim indicates that the battery can't provide enough current. It's not just whether or not it is charged- 10 amps from a 1.2Ah battery is just too much current.

To get the full 125W / 10A that Bajadesigns is claiming this model says that you need to go to 150 turns on your windings. This will produce absolutely no power at idle, but get you up to full power somewhere around 3500rpm. If your battery can't handle the current then the extra power does you no good at low rpms when you need your brakelights and turn signals to work.


I hope this helps.

Eduardo;

In answer to your actual questions: No you won't get 2x125W from your current windings- more like 2x70W

After making these measurements I have to admit that I'm a bit skeptical of the 2x125W claim. Either they really are designed to produce zero useful power at idle or something else is going on that I don't know about.

All I know about reliability is that my BD regulator hasn't failed, and it is connected to a 170W coil. It even survived when my headlight blew out and had no help sinking the load.

The AC regulator you have is the best as far as efficiency- there is no rectifier drop. I also doubt you can improve on the efficiency of the BD regulator- diodes going to be about the same in every regulator

:huh

Teacher, my brain is full. May I be excused? I have to go drain it.

I have self rerwound my stator, I left the ignition poles alone, and wound the other 10 poles with 32 wraps each of 16 ga wire, I also soldered to that a section of 14 ga wire to feed the system, and I purchased a TrailTech DC reg/rec.

my problem is this, I hook up the recifire and get 2.6 volts at idle which drops to 1.2 volts at 4k. Trailtech keeps saying it is my stator and they never see any regs fail. I have no DC shorts or opens in the system where they are not suposed to be. In other words my stator is floated, and is not grounded any place.

Also I can put my stock AC reg back in and it works fine, in fact I can run two 55 watt drive lights and one 65 watt head light at full brightness. I have not gotten a voltage level for the AC.

any thoughts would help.

Thanks Luke.

So, if I was to do about 150 turns (will study the curves to prolly get more juice at low rpm) and use a heavier battery I would be better off. It is all statring to make sense now.

Just one more question. Should I stay with 18 Ga for the windings?

Thanks Luke!

Eduardo
XR650R

My first thought is that if it works with the AC regulator that it's the regulator's fault, but here is another possibility. Do you have any sort of energy storage device on the DC side of the regulator? Either a battery or a big capacitor? The DC regulator won't work properly without one.

We said the same thing in university...... :huh

Would you mind telling me where and how you got lost? Did some part just make no sense? Was it just a general 'my eyes glazed over'?

I'm not much of a writer and it's difficult material but I'd like as many people as possible to be able to understand it.


Regarding the GS, it's because of the motor layout. The stator would add an extra 3 inches or so to the length of the motor which would make the motor even longer than it is already. There's also the issue of where you'd put the points- a stator should be bathed in oil, the points and clutch shouldn't be. It would be really hard to attach the points (or hall sensor) to the clutch side of the motor.

Nowadays, the point problem is solved but the automotive stators can also put out a lot more power much more reliably (the regulators don't burn up)

On th other hand, regular bikes already have a flywheel sitting in oil, not doing anything else- so it's easy to put some magnets on it and bolt a coil next to it. No extra oil seals, no belts to break, etc.

You're all welcome, it was fun to do and I figure I owe you guys for the lc4 oil change thread, your HID thread, the BMW hall sensor fix, etc, etc....

Um, I don't have a second installment... if you mean the LC4, that's mashed in the bottom of this one. I only have theory for that- I haven't hacked on mine yet.

I'd like to say a bit about how regulators work, but my first try at building one didn't work so well so I figure it should wait until I think I know what I'm talking about.

Luke, My experience (30 years ago) was that dirt bike permanint magnet alternators acted like a constant current source in the voltages used. When you take one of the old 6V 35W systems that were just load regulated and add a 12 V shunt regulator you could get over 60 watts out. That is they were just 5.5 A current sources.
Jerry

For 150 turns on 5 poles you should use 15 for OCV and 15 for SCC. Do you understand how I got these numbers? One other thing to note is that the Load Voltage is probably down around 13V for your AC side because there is no rectifier loss.

Is it possible to run the 90W light on DC and the two 55s on AC? That might work a little better- or at least the lights will all be equally bad :)

Regarding the wire gauge- the short answer is yes- go with 16 or 14 gauge. 12 ga. might be too thick to wind properly.

If you can, measure the resistance of the windings with an ohmmeter (motor off, winding disconnected from the load).Add 50% to that number (the resistance of copper goes up when it gets hot) Multiply that resistance by the load current to get a voltage. Add that voltage to the load voltage to see the change in output. The table on this page will show the relative resistance of each wire gauge- use it to figure out the difference different wire will make.

So for example, the stock stator uses 22ga wire and has a resistance of .6ohms. The load is 3 amps. .6ohms*150%*3amps = 2.7V. Changing to 18ga wire (with no other changes) will change the resistance of the coil to .6*6.385/16.14 or around .2 ohms, for a voltage drop of .9V. Take a look at what the change in output is if you change the load voltage by 1.8V