Gregory Hildstrom Projects Publications Resume Links Contact About Google+ Facebook Youtube Donate




2009 Buell 1125R Stator Rewind and Rotor Modification

Introduction
Disassembly
Winding Pattern
Wire
Epoxy
Rotor Modification
Rewinding
Installed Stator Test
Hot Idle Testing
First Ride
Short to Ground
Logger (4-channel)
Results (4-channel)
Logger (4-channel + ECM)
Results (4-channel + ECM)
Logger (4-channel + ECM) Death
Idle Plateau
After 5000 Miles
After 7000 Miles
Additional Information

Introduction

The 2009 Buell 1125R has problems with overheating stators right out of the factory. The high electrical load from lights and radiator fans coupled with inadequate stator cooling tends to melt the insulation and short the stator windings out over time. The stator is definitely not bathed or sprayed with oil; it gets a tiny dribble from nearby bearings at best. This means it is cooled mostly by conduction to the case and by air circulation inside the engine. This means that battery charging capability slowly decreases over time until the bike fails to start or run completely. I upgraded the
voltage regulator to the FH012AA, but my stator failed anyway because this was still a maximum-current shunt-type regulator. It failed to keep the battery charged while running and charging voltages dropped into the mid-11V range. With the stator disconnected and the bike running on the battery, I got 18V AC, 18V AC, and 1.5V AC from the three output phases, which confirmed the fried stator. That's pretty poor performance from a 2-year-old bike with only 4000 miles on the odometer. My first order of business was to upgrade the voltage regulator again, but to the CE-605 SB, which is a minimum-current series-type regulator. After reading about rewound stator failure and seeing the high price of 2008 or EBR replacement parts, I decided to rewind it myself to have complete control over quality and materials.

Disassembly



Winding Pattern

The three phases are connected together in a delta configuration. When viewed from the engine side, the poles are wound in a clockwise direction and each pole starts and finishes on the engine side. I sanded the enamel off of the wire and measured it with a micrometer, which showed that it was AWG 16 wire. Each pole had 44 turns of wire, which was about 11 feet of wire when straightened out. So, 150 feet of wire should be plenty to rewind it. The following photo describes my rewind plan.


Wire

I have been unable to determine the insulation class or varnish/epoxy type used on the factory stator. My plan is to use 240C polyimide (ML) coated magnet wire in AWG 16 size with heavy build. Thankfully,
MWS Wire provided me with a decent quote for more wire than I needed in 16 HML. You cannot get much higher temperature wire without moving to ceramic coated wire, but that requires much larger bend radius. If the original stator was wound with polyimide-insulated wire, parts of the windings must have spent significant time at or above 270C/518F for it to fail at 4000 miles/hours. It is also possible that the original stator was wound with different insulation or a bad batch of polyimide that would fail at lower temperatures.


Epoxy

Special thermally conductive epoxy, like
Duralco 132, has viscosity comparable to or greater than J-B Weld and it is expensive at $86/pint. Electrically resistant epoxy, like Duralco 4461, is good at penetrating windings, but expensive at $90/pint and not thermally conductive. I plan to use J-B Weld epoxy, which can withstand 500F/260C continuous, to hold the windings in place and protect against vibration, but I plan to use as little as possible. The thermal conductivity of J-B Weld is 0.59 W/(m*C) or 2.36 BTU-in/(hr*ft^2*F), which is similar to water and a bit less than Cotronics Duralco 4460/4461 at 4 BTU-in/(hr*ft^2*f). I bought the J-B INDUSTRO WELD package that works out to $15/10oz or $24/pint. I tested some J-B Weld thinned with acetone, which would penetrate windings better, heated to 500F for 4 hours and it held up really well. It was still strong and hard as a rock afterward, but I did not like the surface microcracks. The internal color was a lot more uniform than the photo, which over-emphasizes the fracture surface shadows, but I should have probably done a more thorough mixing job.


I tested some J-B Weld heated to 500F for 4 hours and it held up great. The embedded popsicle stick turned to charcoal. The very dark gray surface is due to either surface overheating from the radiant heating elements, surface oxidation at elevated temperature, or bread crumb smoke absorption.


Next I tested how J-B Weld flows when heated instead of thinned with a solvent. After I smeared the room-temperature J-B Weld on the 200F 4-layer 16 AWG test coil, it thinned, flowed over the coil, and dripped off the bottom. It flowed better with heat than with the recommended amount of acetone. After curing for 24 hours, I cut it open at an angle to see how well it penetrated. It penetrated the first and second layers on the top side with gravity's help. It barely penetrated the first layer on the bottom side with gravity working against it. I think the J-B Weld pulled a significant amount of heat out of my tiny test coil, which also limited the penetration. The stator has significantly more thermal mass. If I heat it to 200F and keep my heat gun on it, I should get much better penetration. I also plan to wipe any excess off for a very thin layer on the outside.


Rotor Modification

Removing the 8-magnet rotor was easy with my impact wrench and no crankshaft locking tool was necessary. I put the bike on my rear stand and put it in sixth gear with the rear tire off the ground. I used my propane torch to heat the rotor nut along all sides for about 20 seconds total. Then it turns out my DeWALT DW059 cordless 1/2" impact wrench was absolutely perfect for this. It has a maximum torque rating of 300 ft-lbs and it took the nut right off. I'm sure someone will say that this put way too much stress on the transmission, but they would be mistaken because I did not use a breaker bar or torque wrench. The rear wheel was free-floating and it did not move at all. The impact wrench worked against the inertia of everything (rotor, crankshaft, rods, pistons, transmission, etc) to loosen the nut. I imagine that the inertia of the rotor and crankshaft alone are enough for the impact wrench to reach 300 ft-lbs. So, I'll be using the same technique to reinstall the rotor since my impact wrench has a maximum torque rating equal to the revised torque spec. I'll use Permatex High Temperature Threadlocker Red, which is essentially the same as the EBR-recommended Loctite 272 Red High Temperature Threadlocker. Months later, I helped Jsg4dfan with his stator and rotor; he removed and reinstalled his rotor nut with same impact wrench and with the transmission in neutral. Neither one of us used the crankshaft locking tool.

EBR's original fix for the overheating stator problem was a kit that contains a rotor with a precision machined oil jet in addition to a lower power rotor and stator. This oil jet sprays oil on all of the stator poles as the rotor rotates. That's pretty cool, but I'm not willing to try that with my hand tools or spend that much money right now. One thing I noticed is the total lack of forced air cooling for the stator. The stator fits tightly in the rotor and there is nothing forcing air in or out of the rotor. My idea, which is sure to be controversial, is to drill six small air cooling holes in the rotor between the magnet insert and the rotor flange. I did not find any oil in this area of the rotor when I removed the case, so this modification will not affect oil distribution. These holes will be small and will not interfere with the magnets, so I feel risk is minimal. Also, the holes are symmetrical, so balance should not be affected much. As the rotor spins, these six holes will function as a crude impeller / centrifugal air pump and force air from inside the rotor to outside the rotor on the engine side, which must be replaced by air flowing past the stator and into the rotor on the open side. I started by punching the hole location to accurately locate the drill, used an 1/8" drill, and then finished up with a 3/16" drill. The steel rotor is about 1/4" thick where I drilled the holes.



Almost a year after doing this rotor modification and debating its effectiveness on BadWeB, an interested reader informed me of some overheating stator problems on the BMW F800GS, which has a parallel-twin Rotax motor. The official fix from BMW/Rotax is, get this, an air-cooling rotor design. Hopefully this lends a little credibility to my hair-brained hand-drilled idea. Notice the same 6 air holes spaced exactly half way between the 6 bolts. The main difference is that I placed my air-cooling holes at the rotor perimeter and BMW/Rotax chose the rotor flange. These flange holes are bigger, but I obviously thought perimeter holes would make better use of centrifugal force. Anyway, the original rotor construction looks remarkably similar to the 1125, which is no surprise. Both this air-cooling design and my hand-drilled one are easier to manufacture than the precision machined oil jet in EBR's rotor. Very interesting.


People known to have the 1125 air-cooling rotor mod as of 7/5/2015: me (Hildstrom), Jsg4dfan, Mitchell, Zew2888, Mike, and Craig. There may be others, but these are the people I've heard from.

Rewinding

I first tried to hold the laminations in my left hand and wind with my right hand, with no gloves. Big mistake. I could not get the wire tight, my hands hurt, and the ridges in the lamination plastic coating made it very difficult to adjust the first layer of windings. So, I sanded the plastic coating around each of the corners. Then I rigged up some clamps and 2x4s to hold the laminations, which allowed me to wind using two gloved hands to pull the wire tight. I used an old wire brush handle to manipulate the windings and to press the first finished layer flat. The first three layers were all about 13 turns with the fourth layer 5 turns or less; 13*3+5=44. I was happy to complete the four poles of the first phase. I did one phase per day for three days and finally twisted the proper wires together for the delta configuration. Check each phase for continuity and ground isolation as you go.


Next, I heated the stator in our toaster oven on convection bake, which uses an internal air circulation fan, at 200F for one hour to make sure it was heated completely through. I marked the inside of a bathroom cup with a pencil using tablespoons of water as a guideline. I mixed up two tablespoons (1 oz) of J-B Weld using one tablespoon of resin, one tablespoon of hardener, and a Popsicle stick. As soon as I felt it was mixed thoroughly, I removed the stator from the oven and set it on a 2.25" hole saw on top of some newspaper. I applied J-B Weld to the engine side first. It had a thick and almost pasty consistency at room temperature (70F), but it thinned, ran, and dripped after a few seconds on the hot stator coils. The drips did not form J-B Weld stalactites on the bottom of the stator, so I know that was a result of small thermal mass in my penetration test earlier. The larger thermal mass of the stator allowed its temperature to stay elevated for much longer, which allowed the J-B Weld to remain thinner for longer. I did not need my heat gun. I applied two coats on the engine side, but that was probably not necessary. A side effect of the heat was extremely fast cure time. A few minutes after I finished coating the engine side, it was no longer tacky and I flipped the stator over to coat the cover side. After pulling the delta wires through to the cover side, I applied one coat on the cover side and you can see it is thinner. In most places the engine side and cover side flowed enough to run and meet in between poles, but some windings were exposed in several places. I was not looking for perfect coverage, just enough to keep things from vibrating and to help conduct heat between layers. With my rotor modification, any exposed copper between poles will function as cooling fins. The side of a pole shown below with exposed windings was the one with the least side coverage; all the rest had better coverage. Based on my lower-temperature penetration test and the fact that the fourth stator layer was only 5 turns with huge gaps, winding layers 2-4 should be well bonded together at the top and bottom of the poles. The elevated temperature and longer duration may have even penetrated into the first layer, but I'd rather not disassemble it just to find out. After doing both sides of the stator and waiting for it to set up (20 minutes), the J-B Weld in the mixing cup was still usable.


In hindsight, a combination of acetone and heat may have been the best approach for this application with this particular epoxy. That would have thinned the J-B Weld even more for better penetration and it would have lengthened the set-up time at this elevated temperature for better penetration. However, my stator is complete, I'm itching to ride, and I'm done testing for now.

The polyimide coating is incredibly tough. I tried hand sanding the ends of the wire, but that got me nowhere fast. I used my mini torch to heat the ends of the wire until they glowed orange, which charred and destroyed the polyimide. Then I removed the charred remains of the coating with a wire wheel and my drill. I used self-fusing silicone tape, aka
Rescue Tape, aka F4 Tape, for electrical tape. It is good to 500F, a great electrical insulator, chemical resistant, and very tough. In order to try to replicate the original attachment method, I used a strand of wire to wrap each lead to each pair of delta windings. Then I soldered it, taped it all up, put down a bead of Permatex Ultra Black RTV Silicone, and mounted it back in the ignition cover. Again, check for continuity between phases and isolation from ground.


Installed Stator Test

I installed the stator. At cold idle, the stator produced about 21V AC on each phase. At cold 3000 rpm, the stator produced about 40V AC on each phase.

Hot Idle Testing

I connected it to the new CE-605 SB series rectifier regulator and did some testing. I idled the bike in various states on my rear stand for about an hour. The voltage never dipped below 12.1 with both cooling fans running and the high beams on; even a slight increase in rpm brought the voltage above 13. The voltage never dipped below 12.6 with both cooling fans running and the high beams off. Here are the results:
time coolant temperature (F) gear high beams rpm voltage
12:49 159 n y 1600 14.3
12:54 179 n y 1500 13.6
13:02 193 n y 1300 12.4
13:10 199 1st y 1700 13.6
13:12 200 n y 1300 12.3
13:19 208 n y 1300 12.1
13:24 208 1st y 1700 13.2
13:31 209 n y 1300 12.1
13:33 209 1st y 1700 13.2
13:34 209 1st n 1700 13.8
13:35 209 n n 1300 12.6
13:49 209 n n 1300 12.6


First Ride

Next, I went for a 30 mile cruise. It was mostly 20-45 mph with only a few miles of highway in the middle. I sat through probably 15-20 stop lights. The ride finished with about a mile at 20 mph. Ambient air temperature was about 80 F. The voltage never dipped below 13.8 even at the longest stop light. Most of the time it was between 14.1 and 14.5. The voltages would probably be a bit lower if the air temperature was above 100 F. As soon as I pulled into the garage, I stopped the engine, removed my gloves, and felt the CE-605 SB regulator temperature. It was warm, not hot; I could grab the regulator continuously with my hand or press my inner forearm against it continuously. I think the mounting location I chose for the new regulator is great.

Now only time will tell if this fix is permanent.

Short to Ground

About 200 miles later, on my 12 mile ride home from work, voltage dropped from 13-14 to 11-12 regardless of RPM. It got as low as 11.1 when I had to wait for 3-4 traffic light cycles. It was about 90F in the shade and my coolant temperature got up to 220F thanks to the extremely low fan voltage. I knew something was very wrong. As soon as I got home, I removed the seat, unplugged the stator, and tested it for a short to ground. It was shorted to ground. I disassembled it this past weekend and found the short. One of the solder joints had been pinched between the cable clamp and the case. The silicone tape could only withstand so much pressure and heat before it tore open. This was probably a result of different wire lengths, different insulating tape thickness, and improper centering of the wires under the cable clamp.


I was very curious about the discoloration of the J-B Weld after only about 190 miles of normal operation and 10 miles shorted to ground. I really hoped that the discoloration was not throughout the J-B Weld, which would indicate severe overheating and a need to rewind again. I lightly sanded the surface of the J-B Weld on two poles. The discoloration was only on the surface. The internal color was at least as light as my 500F test sample if not lighter. The discoloration must be a result of the oily mist getting deposited on or sticking to the hot stator.


If you look at the photos, you can see why I did not center the wires properly. There is a sharp raised stiffening rib cast right in the middle of the clamp area where the clearance should be at a maximum. I went crazy with my air die grinder and took care of that.


Logger (4-channel)

I bought and tested a
thermocouple. I hooked it up to an Arduino Uno using a ProtoShield and MAX6675 IC, but I should have bought a MAX6675 pre-mounted to a breakout board or a SOIC to DIP breakout board because soldering tiny wires to tiny SOIC pins is difficult. I tested the sensor at room temperature and in boiling water; it was within about 5 degrees Fahrenheit for both. Then I tested it in my toaster oven at 300, 400, and 500F. My toaster oven has about +15F -25F range as its thermostat cycles. I used J-B Weld to attach the thermocouple as deep between two poles as I could. I applied several coats, so the thermocouple is embedded under 1/8" to 1/4" of epoxy. After twisting the wire a bit, I shielded it with 1"x1' strips of aluminum foil before taping it along side the stator wires. I paid much more attention to routing the stator wires to avoid pinching them this time and they were taped wide and flat instead of circular.


I wrote up a simple Arduino sketch based on this example and logged 1Hz data to the serial monitor. The J-B Weld had only partially cured, but I had to know if it was going to work properly. I started the bike with the stator disconnected. I got 21VAC at idle just like before. The thermocouple temperature rose steadily and the interference was not bad. Next, I started the bike with the stator connected to the CE-605 SB voltage regulator, temperature rose more quickly, and interference was still not bad at all. Then I stopped the bike and let it cool some. I have three more sensors to add to the data logger before I start riding and logging, but at least I know the stator thermocouple is working well. I'll post photos, schematics, and code once the logger is complete.


I added thermistors and a voltage divider to the ProtoShield. Notice the difficulty I had attaching tiny wires to the SOIC leads; sloppy but functional. That brings the sensor count to 4: stator thermocouple, regulator thermistor, air thermistor, and battery voltage. I finished the board and did some sketch debugging last night. I seemed to be getting the correct numbers from all 4 sensors in the 1Hz live serial monitor and the 1 sample/minute EEPROM log. I used some J-B Kwik to fix the wires going into the SOIC headers. I also used J-B Kwik to attach one of the thermistors to the back of the CE-605 SB regulator in the corner between the heat sink and the epoxy backing. After experimenting with J-B Weld so much, J-B Kwik seems near instant with a very short working time. I used a cable tie to attach the other thermistor to the left bottom license plate bolt hole. It hangs down about 1/2" past the plate, so it should give decent air temperatures near where I mounted the regulator. I put some duct tape on the back of the Arduino to prevent shorts and tucked it and the excess wiring near the battery. I connected a barrel connector to an ignition-switched fuse so the logger will only have power when the key is in the on position or when I connect via USB. The latest Arduino sketch archive, logger-120419a.zip, works with Arduino 1.0 on Linux and Windows. Pin connections are explained in the code comments.


Gnuplot rules, even on Windows. Here is the command I use to quickly and easily recreate plots from the logged data: "load 'gnuplot.txt'". I capture the EEPROM dump in Linux using screen, but you can easily grab it from Arduino's serial monitor window using ctrl-a/ctrl-c and then paste it into a text editor or Excel. Gnuplot handles the LOGDATA strings and the column headers without complaining. The current logdata.txt file must be in gnuplot's current working directory for this to work.

I developed a short on the logger almost immediately. The bike's vibration caused the solder dots on the back of the Arduino to rub through the single layer of protective duct tape and contact the battery box. The SPI pins for the MAX6675 were intermittently shorting out, which is why the stator measurement was dropping out to zero and everything else was ok. I fixed this with five layers of duct tape and a cable tie.

PartSparkfun Part NumberUnit Price
Arduino UnoDEV-11021$29.95
Arduino ProtoShield KitDEV-07914$14.95
Thermocouple Type-KSEN-00251$13.95
Thermocouple Amplifier Digital MAX6675COM-00307$11.95
SOIC to DIP Adapter 8-pinBOB-00494$2.95
Thermistor 10KSEN-00250$1.95
9V to Barrel Jack AdapterPRT-09518$2.95
Total w/o shipping, resistors, & wire $78.65


This design can support three more analog inputs, for a total of 6 analog inputs. It can also support at least three more MAX6675 thermocouple ICs for a total of 4 thermocouples. If you share the clock and output pins between MAX6675 ICs, you could connect 10 of them to D2-D13 in addition to the 6 analog inputs A0-A5. These additional measurements would certainly reduce the time capacity of the EEPROM or you could reduce the sample rate, but you could also log via USB or alter the design to use a microSD shield. However, I do not need many more measurements and the current EEPROM design seems to be working well for me. The capacity at 1 scan/minute with 4 8-bit samples is over four hours. Here is a pseudo schematic of what I built:


Results (4-channel)

I went for an 80-minute ride to test it out. My ride included highway, 20-30mph, and a few traffic lights; almost all of it was below 5000rpm. Here are the results from
logdata01.txt:


Everything looks good to me. The maximum stator temperature (282F) is well below the maximum wire insulation temperature (464F) and the maximum voltage regulator temperature (115F) is well below the maximum safe operating temperature of most electronic components. However, the average air temperature was about 88F and I did not have to sit through consecutive light cycles. Things will probably look a bit different when I have to sit through four consecutive light cycles in over 100F.

I commuted for a week with daytime highs in the upper 80s and low 90s. Here are the results from logdata02.txt:

The log data revolved and you can only see a portion of my first 80-minute test ride remaining around minute 170. My commute is usually 20-30 minutes. The lower peaks at 40, 95, 138, 205, and 248 are the end of my morning commutes. The higher peaks at 20, 65, 120, 156, and 230 are the end of my afternoon commutes. The dips in battery voltage at 10, 55, 115, and 220 correspond to sitting through 3-4 consecutive traffic lights toward the end of my evening commute. Both radiator fans draw enough current to reduce idle voltage to the low 12s. What is really interesting is the amount of time required to get the stator up to full temperature; at least with these modifications and air temperatures. My 30-minute afternoon commute with several traffic lights was not long enough to get the stator up to the same temperature recorded during my long test ride, which is a good thing. The small peak at 75 was a quick test after fixing my ignition cover gasket situation. The small peak at 98 was a quick 3-mile ride from my office to my wife's office.

Logger (4-channel + ECM)

Numerous people on
BadWeB have suggested that it would be nice to have RPM and speed data to help make sense of the temperature and voltage measurements. Sniffing the CAN bus would be cool, but that is a research project in itself. Thankfully, EcmSpy can log run-time data and they have published a lot of information about serial communication with the Buell 1125 DDFI-3 ECM. Since I'm not concerned with ECM data faster than 1Hz or even 1/60Hz, polling the ECM via serial should work fine for this temperature logging application. While I was searching for more information on the checksum byte in serial requests and responses, I came across the Buell ECM to Arduino project, which aims to display instantaneous horsepower, torque, and shift point in addition to logging ECM data. The information posted in that blog is extremely helpful even though it is for a Buell XB9 DDFI-2 ECM. I combined the information from the EcmSpy website and the Buell ECM to Arduino blog to add some ECM logging to my 4-channel logger. I made the serial ECM cable, connected ECM TX to D8, and connected ECM RX to D9. I did not run the ground wire since the two controllers already share a common ground. The latest Arduino sketch, loggerecm-120502c.zip, logs 4 sensor values (stator_thermocouple, regulator_thermistor, air_thermistor, battery_voltage*255/15) and 4 ECM values (RPM*255/12000, VSS, CLT, Fan_Duty) to the EEPROM. With 1 scan/minute and 8 1-byte channels, the EERPOM can hold just over 2 hours of data. The gnuplot script I use to plot this data is gnuplotecm.txt.

On a side note, if you require higher performance or more features, the TI LM4F232 USB+CAN Evaluation Kit, the BeagleBone, or the Raspberry Pi may be good alternatives to the Arduino platform.

Results (4-channel + ECM)

Here is my first commute with ECM data being logged. I took a route home with less highway and a lot more traffic lights. Temperatures are still in a safe range, but the most interesting thing is that the fan duty cycle field exceeded 100%. My best guess is that this value is calculated based on what the ECM thinks is needed and that values above 100 indicate you are at the mercy of ambient air temperatures and ambient air flow. Here are the results from
logdata03.txt:



Here are many more commutes from logdata04.txt:



Here is a 90-minute ride from logdata05.txt:



Logger (4-channel + ECM) Death

I set out to capture the idle stator temperature plateau. I went on a long ride and then idled the bike for an hour. Afterward, when I downloaded the data, I had hardly any stator thermocouple measurements; they were mostly zero. This is probably from my shoddy SMD workmanship around the MAX6675 chip. I removed the Arduino from the ProtoShield and then soldered up one of Adafruit's MAX6675 breakout boards, which provides much better connections to the IC and the thermocouple. It works great.


Idle Plateau

Ambient air temperature was 100F. I went for a 1-hour ride around town and then pulled into the garage. I put the bike on a rear stand and put a 24" fan behind the rear tire blowing out of the garage. Both doors were open. Ambient air temperature rose from 100F to 104F during the test. I connected my Arduino and MAX6675 board to my laptop and to the stator thermocouple leads. I idled for 40 minutes and observed the plateau of 163C or 325F; maximum coolant temperature was 225F. At minute 47, I disconnected the rear fuel injector and later observed a higher peak of 167C or 333F. At minute 50, coolant temperature was 211F. At minute 66 coolant temperature was 194F. I think stator temperature rose 4 degrees when the rear injector was disconnected because of the slight drop in idle speed and the rough running on one cylinder with the high fan current draw. It took a while, 20+ minutes, but coolant temperature dropped enough so the ECM could reduce fan duty cycle and the combination eventually reduced stator temperature. At minute 71, stator temperature was 159C or 318F and I turned off the engine. I am pleased with these results. My stator should survive this kind of torture indefinitely. Even a stator wound with 200C insulated wire instead of 240C insulated wire should survive this torture. These numbers lead me to believe that the air-cooling rotor is an improvement over stock, but I do not know by how much. I am also pleased with the rear fuel injector disconnect modification because it works. If I had started the idle test with the rear injector disconnected, like shifting into neutral with the mod at a light, coolant and stator temperatures would never have climbed to the plateau levels I observed. However, there has been some debate about whether the temperature actually plateaued during 2-cylinder operation; people want to make plots with large dots and ignore the obvious kink in the data when I switched to 1-cylinder operation. I should have let it run longer on 2 cylinders to silence the skeptics, but temperatures were in a safe range either way. The results from
logdata07.txt are plotted here:

Also, fuel started boiling at some point during the test.



After 5000 Miles

I had a little oil mist leak around where the additional thermocouple wire exited the ignition cover. It had also been about 5000 miles since I rewound the stator, so I figured it would be a good chance to fix the leak and inspect the stator. My charging system is still working great. The voltage doesn't dip quite as low at long traffic lights since I've had my BatteryMINDer 12248 charger/maintainer/desulfator, so I think a weak battery was related to that. As you can see in the following photos, the J-B Weld is still discolored about the same as it was when I had the short to ground around 200 miles. I lightly sanded it again to verify that the discoloration is still only on the surface of the epoxy. The J-B Weld certainly wasn't designed for this application, but aside from the surface discoloration, it seems to be holding up great. The exposed polyimide insulation I can see still looks like new with no discoloration. I now have 1000 more miles on the rewind, air-cooling rotor, and series regulator than I had on the original charging system.



After 7000 Miles

I sold my 1125R right around 11000 miles on 10/22/2014. My rewind, rotor mod, and series-style regulator had about 7000 miles and almost 3 years on them when I sold it. Everything was still working great. I sold it because I was ready for a change, not because of any problem with the bike. I was a bit sad to see it go, but it went to a good home and I really enjoyed my time with it.

Additional Information