First test drive with K9HV and its new controller

To make sure that the new motor installation works properly I decided to build the car ready for a test drive with its current battery voltage of 76,8V. There was only a few things to finish up before the first meters could be driven with the Kostov K9HV motor installed.

Rebuilding the motor controller caused lack of space in its enclosure so the throttle pedal could not be located there anymore. The solution was to replace Corolla's original throttle pedal with a hall throttle. The hall pedal was mounted in a holder made from steel tube. The throttle pedal cable stays well protected inside the holder.

Now all that was left was to mount and connect the new SR250V500A-V2 motor controller and drive!

After a half year of staying inside the garage the Corolla finally got some fresh air:) It was exciting to test both the motor and also its new controller. The first notices were that the motor mounting had succeeded much better than with the ME1003, I could hear the motor RPM, but I could not feel its vibrations, excellent!

The controller worked also as well as I could have hoped for. The maximum power draw was 15kW just as it should with this battery voltage and the controller's current settings. That maximum power should nearly double after setting Corolla's front and rear battery packs in series, resulting into a battery voltage of 153,6V. There was also plenty of torque available. It was a slippery day and the drive wheels started to slip with any higher than 200A motor current on 2nd gear.

There is always something needing improvement; slight oscillation occurred during a slow acceleration. This is most likely caused by too high gain parameters in the motor controllers current PI-controller. These parameters will be tuned after the Corolla's battery voltage has been doubled. Knowing that the basics are ok with these voltages I can now start making the changes needed to double the voltage being more suitable to the K9HV motor.

Fitting the Kostov K9HV into Corolla's gearbox

I had to build a new flange to mount the K9HV into the gearbox. The flange should be at least 30mm thick to prevent the motor axle from touching the gearbox axle. I had some 8mm thick aluminium scrap sheets available for the job, so I decided to attach two such sheets together forming a 16mm thick sheet. The flange would be constructed from two plates, one 16mm plate would be mounted to the motor and an other 16mm plate into the gearbox allowing me to adjust their alignment. I started by making a model from cardboard for the gearbox plate.

Using the cardboard models I cut the gearbox plate using a jigsaw.

Locating the holes to drill was a challenge. I decided to try pressing the plate against the gearbox with some Blu-Tack in between so the hole locations would get visible. This actually worked out well. I started by making two alignment pins that the original motor had as well. After the locating pins were made it was simple to locate one hole at the time to be drilled.

Once all holes were drilled, the gearbox plate could be mounted into its final location. Now the challenge was to get the gearbox axle aligned with the motor. I decided to build a special divider tool that allowed me to draw an accurate circle matching the K9HV motor's diarmeter. This would help me to make a first rough alignment.

Next I decided to build the motor's plate. I used countersunk bolts to mount the plate on motor allowing the motor plate to be pressed directly on the gearbox plate.

The motor plate was built with the same width as the motor's diameter so I was able to locate the motor plate mounting holes quite accurately on the gearbox plate using the circle drawn earlier with the divider. I made the motor plate's holes 2mm bigger than its bolts allowing me to fine tune the alignment of the motor and gearbox axles. The fine tuning would be completed by adjusting three stop screws forcing the location of the motor plate from its bottom, left and right sides.

The motor plate's mounting bolts threads were reinforced using helicoils. These four bolts will need to carry the motor and its torque so I did not trust bare aluminium threads.

Now I had all the components ready for finishing the assembly!

First the motor plate was mounted into the motor securely using plenty of thread lock.

Then the gearbox plate was mounted roughly on its place to be aligned later.

Next I assembled the flywheel and clutch. I used a torque wrench and thread lock in all places applicable.

Finally I got to fit the motor in its place and bolt the gearbox plate securely. Then it was time to adjust the motor and gearbox axle's alignment. I had no idea how I could confirm a good enough alignment so there was a lot of trial and erroring to do. Pretty soon I noticed that when I rotated the motor by hand I could hear a quiet creak as the clutch disk was forced to move due to misalignment of the two axle's. I then adjusted the motor plates location in respect to the gearbox plate until there was no creaking sound at all. I figured this alignment is as good as I would get and lock screwed the plates together so they would not move anymore.

To make sure the motor was properly mounted, I connected the motor into its controller and gave it a spin. I was really happy with the result, there is clearly less alignment and balancing error than I had with the former ME1003 motor's mount:) This motor will drive the Corolla smoothly!
  

Placed the original shields on the motor and now it's ready to be lifted in the car:)

Preparing the Kostov K9HV motor

The K9HV was purchased pre used so I inspected it through before mounting it. The picture above shows my first setup that was able to spin the motor. The controller in the picture was the first prototype of Motoc - SR250V500A that will be developed to feed Corolla's motor. This setup got the motor to spin nicely.

A closer look to the bearings revealed a worn bearing housing. The smaller diameter axle side of the motor had about 0.5mm play between the bearing and the bearing housing. This was going to be fixed.

The motor's rear flange was removed so that the bearing housing could be repaired. The ball bearing seemed ok, but it got replaced with a new one just in case.

Cleaning the bearing housing revealed clear signs of wearing. Reviewing different methods to fix the worn bearing housing, I decided to purchase a bottle of Loctite 660 from eBay and give it a try. First I punched some marks in the worn bearing housing to help aligning the new bearing better in the center. Then I applied plenty of  Loctite 660 into the bearing housing before mounting a new ball bearing into its place. Carefully removed the excess Loctite so it wont get in the bearing, I think I got it fixed :) After a couple of days I checked the result and there is now a new bearing installed without any play on its bearing housing, success!

My next challenge was to mount Corolla's old flywheel into the motor shaft. That turned out to be easy as I only had to purchase a 1210 taper lock bush with a suitable bore to K9HV's shaft. www.bearingboys.co.uk had just what I needed once again!

With a new taper lock bush I got my old flywheel mounted and I decided to check how it was aligned. Measuring from the clutch plate surface I got nearly 0,3mm difference on one full rotation. I concidered this as too much so I gently sanded the taper lock bush so that it would compensate the alignment error. After a few trial and errors I settled with and error of 0,06mm. This step is done:)

Last but not least it was time to balance the motor with its flywheel on the shaft. With first rotations it became clear that there was some imbalance that could provide uncomfortable vibrations. As this motor will be rotated close to 2x higher speeds that its predecessor I was ready to offer some time to get the flywheel balanced.

The balancing method used was very simple trial and error, using neodymium magnets on the flywheel and determining where to drill off some material to improve the balance. Having drilled four quite big holes I could no longer determine any improvement so I considered the balancing step done. While balancing with this method I had some flying magnets in my garage when I rotated the motor too high. Although handy to use, I would not recommend using magnets this way for balancing flywheels due to safety.

Now the motor was rotating nicely up to very high speeds, so I am ready to start fitting the motor into Corolla's gearbox :)

ME1003 Rotor burned - Drivetrain upgrade begins here

Driving with the new KDH12801E controller lasted only for a couple of weeks resulting in a rotor damage in Corolla's ME1003 motor. I was not able to use low Motor current limits of the KDH12801E controller because it caused very strange behaviour in crawling speeds. Having to use the Motor current limit close to 800A obviously burned the motor. Corolla's former controller PM72401B was a better fit to ME1003 in this application but nothing can change the fact that ME1003 is a too small motor for a car like Corolla.

I decided to upgrade Corolla's drivetrain with higher rated components and double its voltage up to 153,6 Vdc nominal (48 series connected Winston LYP60AHA batteries). The upgrade project started by dropping off the motor and gearbox so that a new motor can be fitted easier into the gearbox.

The above photo shows the burned ME1003. Its brushes had melted badly and the rotor has short circuited. It will be scrap unless a new rotor can be sourced and fitted. The other picture above shows Corolla's gearbox drive shaft where the new motor will be fitted.

Here's Corolla's new motor: Kostov Motors K9HV. This used motor fitted my budget for the Corolla while being a good improvement for the whole car. It's a series motor that can handle voltages up to 220Vdc and it can be rotated up to 7000RPM, which means about 90km/h on Corolla's 2nd gear. I am looking forward to be able to drive my usual trips using only the 2nd gear and not needing to change gears as I had to do with the ME1003 motor:) The down side using a series motor is that I'll be giving up regenerative braking at least for now.

The K9HV Motor will be driven with a custom made motor controller inspired by the awesome Paul & Sabrinas open source DC-motor controller and Evnetics Soliton controller. In the new controller I'll be using Infineon's XMC controller and IGBT Chopper module togheter with a DC-Link with only Film capacitors. I am hoping to also publish this new desing later as open source:)

Circuit diagram update 25.4.2017

Corolla's circuit diagram got an update while having to change the motor controller. The KDH12801E cannot be powered directly from the battery pack so a new 12V 5A isolated power supply P3 was added. Also the 400A main contactor had to be changed into one with a 12V coil voltage. Also added the K5 brake signal relay which is powered directly from Corolla's brake light signal. When the brake lights are lit, the K5 relay is activated, and the motor controller starts to regenerate with a constant motor current of about 50A. This way the regen is not strong but it is often effective enough so that Corolla's mechanical brakes are not needed.

Replacing a broken motor controller

On a typical lunch break me my four collegues had just started driving back to work when all torque stopped without any warning. The Kelly Controller PM72401B motor controller flashed for "internal voltage fault", so we pushed the Corolla to a nearby parking space and took a walk back to work.

After towing the car back to my garage I figured out that some mosfets from the controllers other half bridge had failed. The PM72401B main circuit seems to have 7 IRFB4110 mosfets in each leg. As tough as the IRFB4110 is, it's still only a 100V mosfet. The PM72401B operates at up to 90V battery voltage which leaves very little margin for overvoltage spikes. Corolla's ME1003 is a high current motor specified to have a low inductance of only 93uH @ 120Hz which leads to a high current ripple and switching high currents lead to high voltage spikes which the mosfets have to deal with. 100V mosfets are not enough with Corolla's 76,8V battery pack and ME1003 motor in my opinion. I am actually positively suprised for how long the PM72401B managed to survive; 3,5 years and 28000km of driving.

As also the current handling capability of the electrolytics inside PM72401B seemed too low for my application I was determined to replace the controller with a controller having at least 150V mosfets and higher current capability.

Sadly there is not much to choose from when it comes to suitable permanent magnet motor controllers and there are even less alternatives if regenerative braking is wanted. I ended up making a decision between a larger Kelly Controller and Paul & Sabrina's 144v 500amp DC EV Motor Controller Kit. It was a difficult decision between a well tried open source controller and a commercial black box. Should there have been used P&S controllers on sale I would have gone that way, but as there weren't, I decided to go with the shorter lead time solution to get the Corolla back on road faster.

I purchased a Kelly Controller KDH12801E for the job. Its advatages over the P&S were the ability to regenerate and its shorter lead time due to not needing to build the controller. Down sides are that I don't know what's inside the black box and I also have very limited possibilities to configure its behaviour.

The KDH12801E is a 2-quadrant controller being able to motor and regenerate into one rotating direction of the motor. It's rated up to a 120V battery pack nominal voltage and up to 400A continuous current, making it a few times higher power specified than the PM72401B. The controller was delightfully shipped with a USB-to-RS232 converter, a pre-charge resistor and some diodes needed for contactor controlling. The pre wired J2 cable assembly also proved to be very handy.

I had to make a rebuild for my controller assembly box as the dimensions of the new controller varied a few cm. A bigger heatsink was also mounted and machined to fit. After a couple of evenings the Corolla was ready for test drives.

The first test drive impression were a dissappointment. I had limited the motor and battery currents to 50% meaning 400A each. Driving like this was simply slow and the controller gave only around 12kW of power at max throttle meaning around 180A of battery current. I had to increase the motor current up to 87% to achieve the same power levels that I had used daily with the PM72401B. This tells me that the current configuring affects not only to average currents but also to peak currents. Other down sides were that the regen stops at remarkably higher RPMs than the PM72401B did. But the worst thing is that a smooth acceleration is at the moment not possible as the controller gives an uncontrollable current peak when trying to accelerate smoothly from stand still. I hope that I can figure out a configuration that would help it even a little. Anyway I am glad to have the Corolla back on the road again!