500

A milestone or is it a kilometer-stone...

With the odometer working, it is much easier to see total electric distance.  Well today, Jane hit 500 electric-KM.

This is significant since it has been 500KM without a mechanical issue.  Next goal.... 1,000KM

Electric-Regenerative Braking

One of the interesting characteristics of electric cars is their ability to recapture some of the energy that would normally be lost to braking and use it to recharge the batteries.  A key reason to use an AC (PMDC) motor is they are ideal for regenerative braking.  Previously, the regen function was set in the Kelly Controller (KHB 72701) to start when the gas pedal was released.  However, with the desire to be able to shift gears while rolling (and the lack of clutch), I realized that if the motor was braking, it put back-pressure on the drive train which made shifting impossible.  Also, the brake lights would not illuminate while regen was active which may have been confusing for other drivers, since Jane was slowing but no brake lights were on.

The Kelly controller supports having Regen start upon the grounding of the brake-sw signal.  So a simple 12V relay and a few wires later, the brake-sw signal is now grounded when the brake-lights are activated by the brake pedal.   This give the driver complete control.  To coast or shift, simple take your foot off the gas and the motor free-spins.  Shifting from 2nd to 3rd or 4th is easy since the motor has little momentum.  When the brake pedal is pressed the controller turns on regen braking and starts to put charge back in the batteries.  The braking is currently set at a 15% current limit or about 105 amps.  This results in soft braking right as the brake pedal is pressed and before the mechanical brakes start to engage.

The Arduino Mega digital dash reports that initially ~1200 watts are being generated through regen and this decrease as the car slows.  @ 15% the onset is not too jarring yet the rate of deceleration is reasonable for city driving.  This will save the mechanical brakes and provide a little charge back to the batteries.

A small code addition to the dash added a counter to display total watts from regen along with total watts from normal driving.  This will provide some information about of the range benefit from regen.

Efficiency and More

Engine Positioning

This weekend had a number of small enhancements which all came about from replacing the Pot-Joint rubber boots earlier.  The boot replacement required pulling the engine which was straight-forward and this provided the opportunity to tweak the engine position to assure the pot joints were fully engaged.  After about an hour of shifting and sliding the engine around, a new engine mount was attached along with re-positioning of the torque bar attached near the top of the engine by over an inch.  This greatly improved the drive-shaft alignment and thus the pot joint positioning.

Also, which the engine was out, I added rubber insulation between the engine mounts and the frame to absorb some of the vibration from the motor at low-RPM.  This is sometimes called cogging which is typically with BLDC motors.

The new motor mount is shown in the center of the photo.  Please ignore the tie-wrap since it is just a temporary addition. .

Cooling

The custom adapter (Black sheet metal) to the route the 200 CFM fan to the motor cowl was in need of some sealing, so a strip of vinyl weatherstripping placed all around provided a great air-tight seal, assuring all of the blower's air is going into the motor and is not wasted.

The black adapter, shown below the motor here, adapts from the 3" flexible hose to the motor cowl.  Near the center of the image is the cowl screw which is now all the way to left side of the slot since the cowl is rotated  fully clock-wise.

Timing

Many people talk about advancing the timing on electric motors to gain more power and efficiency.  Many motors are shipped for a balanced approach of setting the timing for both low and high RPM while maintaining symmetry for both forward and reverse operation.  Well with Jane, the motor is going to spend most of its time around 3,000 RPM in forward, so there may be some efficiency gains from tweaking the timing.  This involves rotating the fan cowl clockwise since the timing pick-up sensor is mounted in the cowl with the magnets attached to the fan-blank now installed.  The consensus is the motor should run more efficient (less heat, more RPM) at higher currents & RPM.  

New Electronic Dashboard feature

The Arduino proved its value once again.  The ultimate efficiency metric for electric cars is Watt-Hour/Mile.  This is similar to a miles/gallon measure used in the US for gas cars, though with WH/mile, a lower number is better.  Well 4 lines of code and the digital dash is now reporting this stat, allowing the driver to see how efficiently the car is running.  Some early cruising data is showing in the 100-150 WH/Mile.  The calculation was straightforward - = watts / (rpm*RPM_Per_MPH)  - This is equal to Watts/Miles_per_Hour.  It is probably time to add logging of the data to start to gather real efficiency numbers.  The Arduino has an SD Card to hold the boot-up-image of Jane Austen, so there is plenty of storage and a permanent USB cable.

Also, we attempted to cover the passenger side dash with vinyl.  This was mostly a learning exercise to figure out how to cut, stretch and glue it.  After 3 different glues, we settled on DAP StrongStik as the best to adhere to both wood and vinyl.  It was permanent, cured in about 2 hours and was not too messy, like the spray glue.

LED Lighting

Certainly an electric car should have LED exterior lights.  There are equivalent bulbs available for all of the lights, so a quick swap-out brings Jane into the current decade.  They are both bright, cool and more efficient.  This did require swapping in digital flasher controls for both the blinkers and the hazards.  The Hazards' wiring had been moved around a bit but that was quickly resolved.  The only remaining incandescent bulbs are the interior turn indicators and the illumination for the heater controls.

Bulbs Used

• Single element Yellow (rear turn-indicators) Yellow
• Dual Element front turn indicators + running lights White/Amber
• Dual Element Red - Rear brake lights + running lights Red/Red
• Single Element Small side fender running lights - Red

Small note: The front turn indicator/running lights were wired in reverse of how the LEDs White/Amber LEDs were, so more LEDs were illuminating when the running lights were on and the turn-indicator was difficult to see.  Swapping the two wires, resulted in the running lights being the dimmer mode and the turn-indicator being the brighter.

Results

The best news from all of this is two fold:

1) 200 new RPMs are now available from the timing change.  This translates into 2nd gear now topping out at about 37 MPH (previously around 32 MPH).  Third gear now can cruise at 45 MPH even on a slight incline.

2) Cooler running - With the increase air-flow and efficiency improvements, the motor's over-heating problem is solved.  Under normal driving conditions, the motor is staying below 70C.  On hill climbs, the temp approaches 90C but quickly drops once the climb is complete.  Even while cruising at 35 MPH, the blower is able to reduce the temp of the motor, which in the past, was not happening.  Previously, the motor would only cool while at red-lights and there was always a slight rise in temperature while cruising.  Today the air temp was 83F, so this was a reasonable test.

Next up....

More paint stripping  1 fender and 1 door are down to bare steel.

More Instruments - Temp and KMs

Just some minor enhancements that will really help the driver.

The odometer (though in kilometers) had a feature that forced it to stop advancing when it hit 99,999.9 KM.  After disassembling the speedometer it only took a little push to move the counter past 99,999 back to zero.  Well today it worked just perfectly.  There was still 40 year old grease on the gears but I added some fresh just for good luck.  I guess this starts the official counter of electric miles (or kilometers for those who can't do fractions).

Also, with about 5 lines of code, a 2.2K resistor and a 2N2222 transistor, it was trivial to connect the dashboard Arduino to the Smith Temperature Gauge.  Using the

• Pin 9 PWM to drive the base of  transistor through the resistor
• connecting the emitter to ground
• the collector the gauge's input (the original temp sensor input from the non-existent engine)

... the system worked.  The other connector on the gauge was connected to +12V, once it was discovered the voltage stabilizer was not working well.  The spade was lose which caused intermittent behavior.

Here is a shot of the odometer and temp gauge in full operation.  The gauge is currently set to show 70C as "N" (mid-scale) so that should provide a quick indication for the driver if the motor is heating up.

A voltage gauge is on order to replace the oil pressure gauge on the right side of the speedo.  This will monitor the 12V system to make sure it is operating properly.

25 Miles

A long test drive

Today was a day of test drives.  The first was a speed test.  While cleaning up some wiring I realized that the accelerator cable may not be pulling the controllers sensor fully, so I lowered the effective top of range to 70% of full scale.  Well it made a difference and the motor is now spinning near 4300 rpm in neutral.  Which is very close to the theoretical limit of 50 rpm/V with a 85V battery.  Given, this I wanted to see what Jane would do in 4th gear, so we headed to the one stretch of nearby road with a 45 mph speed-limit which is outside the city limits.  On the level straight-away, Jane hit 53 mph and was humming along.  There may have been a few more mph available, but at about 50 mph, I noticed that the front wheels are out-of-balance and started to shimmy a bit.

Well, after a top-off-charge, we headed out again, this time with a goal of testing range and the fuel gauge.

The image shown above is the screen shot from the GPS Essentials program which shows 25.2 miles covered with a max speed of 35 mph.   The max speed is only captured while the screen is on.  It looks like in 2nd gear, Jane will do about 37 mph, which for city driving is perfect. 

The fuel gauge showed 15% remaining which calculates out to 41 AH consumed (out of the 60 AH available in the battery).   Using the 80% available capacity, this works out to a range of 29.6 miles on a full tank (with a 10% reserve still there to protect the battery).  The batteries were not showing any signs of being depleted.  Only one near the end did the low-voltage alarm trigger during a very steep hill climb.  It cleared as soon as the load was reduced.

The other interesting calculation is 41 AH x 84V / 25.2 miles = 137 WH/mile or 7.2 miles/KWH.  This works out to about 250 miles/gallon (MPGe) using the conversion factor of 1 MPGE = 0.0292 miles/KWH.  This does not take into account the efficiency of the charger which would lower this by about 40%.  You wouldn't think the charger would have to be taken into account....  I have not done the math to see if spending more on an efficient charger would save enough electricity to justify the added cost.

To take this one more step... At $4.00/gallon, Jane with the gas engine hit about 42 mpg or about 9.5 cents per mile for gas.  Based on Oregon electricity prices at 11.5 cents/KWH... the electrical cost comes in at about $0.115/7.2 miles or 1.6 cents/mile.  Even factoring in 60% efficiency 2.25 cents/mile.  The battery pack adds about 4 cents per mile so this brings the total cost of the electric mile to about 6.25 cents (a 30% savings from the gas model of 9.5 cents/mile).

Finally, accurate current measurements

Getting to the "truth"

With a borrowed Fluke 600A DC amp meter, I now have accurate current measurements for the charging system.

1. The charger's current reading is within 0.1 amps of what the Fluke reports for current going into the batteries.
2. The charger pulls 10.1A @ 240V or close to 2,450W.  The fan adds another 0.4 amps.
3. The 12V battery charger pulls about 0.5A @ 240V
4. The 12V power supply draws approximately 0.1A but is probably in the noise.

Impact

1. The current monitor (fuel gauge) Arduino is calibrated to the charger.  This should be now less than 5% off.
2. The fuel gauge is set to never exceed 100%, so it should "zero-out to full" near the end of every charge cycle, as long as the cycle runs to completion.  By ignoring the charging efficiency of the LiFePO4 cells, the fuel gauge should hit 100% about 1%-2% before the batteries are full.
3. The charger's efficiency is about 55%-60% based on power-in (240V x 10.1A = 2425W) divided by power-out (~90V x 16A = 1440W; the charger is rated at 1500W).  The charger spec states that charging should take 1.41x the battery capacity / 16A.  Not a surprise this ties out to the 60% efficiency, though it is spec'ed at 85%-95% efficient.  Maybe at the higher voltages it can achieve this.
4. The best news is the overall energy efficiency of the car is coming in where predicted.
1. 4000 KWH of usable capacity (80% of the pack)
2. Today I drove close to 12 miles using 48% of the pack.  This included several hill climbs and normal traffic driving. 
3. This leads to about a 25 mile range with 20% reserve. or 160 KWH/mile.
5. At work, there is CharePoint charger which reports KWH used, so next I'll try to calibrate against that with the data above.  Mostly it is good just to have an accurate fuel gauge

The Non-Fan - Cooling the ME0913

The Fan or lack there of....

A local machine shop fabricated up a blank disk to replace the fan in the Motenergy ME0913 motor.

This is the original fan in the motor.  Note the blades spin fairly close to the motor housing, which works well when the rotates counterclockwise (CCW, to the left).  In Jane the motor spins clockwise (CW) from the fan-end and thus the blades move no air and only serve to frustrate (restrict) any air that is trying to be pulled through the motor, from the shaft end, which is the direction the original fan attempts to move air..

This is the original fan out of the motor.  The central hub is riveted on and the timing (magnetic) ring is bolted on.  There are no markings on the timing ring, so I marked its position relative to the key slot on the hub.  It took a little persuasion to get the fan off the motor shaft.  There is a 2.5 mm set screw holding the key and hub in place.

The blank aluminum disk is now mounted to the hub using #8 bolts and the timing ring mounted to it with the original bolts.

The disks is now on the motor.  A little high temperature grease was put on the shaft which eased the mounting process and hopefully removal if that is ever necessary in the future.  You will see there is now a nice gap between the disk and the vent holes in the motor housing.  The theory is air will move more freely through the motor and around the spinning disk to the exhaust fan.  The exhaust fan was able to cool the motor quickly when the motor was not spinning but when it was operating, the cooling efficiency really seemed to drop.

Good Results

Hill driving is the best test since it forces the highest current through the motor for an extended time.  I took Jane over to the steepest hill in town and attempted a climb at 20 MPH in 3rd gear.   The motor was already warmed to about 60C.  In the past the temperature would rocket up to 120C about half-way up the hill.   This time under a good load (15KW) it only got to 105C about the time Jane crested the top.  The second test was whether the cooling fan would be able to cool the motor while moving.   This was something that did not happen in the past.  The next 3-4 minutes were driven at lower power (3KW) and the temperature dropped quickly to 85C.  Just the fact the temperature dropped is a vast improvement.  In the past, the only way to see a decrease in temperature was to stop and let the cooling fan run for a few minutes.  Tomorrow, I'll try a more extended run to see if there is a steady-state temperature under normal driving conditions.  Also, the next enhancement is to connect the Arduino that reports the temperature from the Kelly controller to the original Mini engine temperature gauge.  

Controller Supply Voltage and Fan Update

Surging

During the last few test drives, I noticed that the motor tends to have a slight periodic surge when approaching maximum rpm.  It is subtle when cruising at a constant speed, but a slight pulsing or surging could be felt.  This led the following investigation:

1. The Kelly KBH 72710 controller is rated for a supply voltage of 8V-30V.  This is the supply for the controller's functions, not the main drive power which is 85V in Jane, with the controller rated at 90V.
2. The system currently has an adjustable voltage booster to covert the 12V standard car supply upward.    TI LM2577 Link
3. Previously, the booster was set to output 18V to supply the controller.
4. I can't remember why I landed on this but there were  published reports that 12V was insufficient.
5. While reviewing the Kelly datasheet again for the controller, it says that 24V is "preferred".
6. So, a few small turns of the screwdriver and the booster is now set to output 24V.
7. During a test drive today, the surging seems to have stopped.  But it will take a bit more testing to confirm.

Not sure why this was not noted previously but it can only help.  

Cooling

In the next attempt to cool the motor, a local machine shop is fabricating a blank disc to replace the fan in the Motenergy ME0913.  John from Motenergy provided the mechanical drawing for the fan assembly.  The fan carries the hall phase magnets so it is a critical component of the motor and can just be simply removed.  The current theory is the fan, though running in reverse (CCW from the fan end)  from its designed direction (CW from the fan end), is frustrating (impeding) the air flow through the motor.  With the blank disc, there should be no interference or obstruction for the air.  Currently, the cooling system is just on the verge of being sufficient, so if by removing the fan, there can be a 10%-20% gain in flow, the results should provide sufficient headroom.

Dashboard

Work continues on the electronic dashboard.  Here is the latest design:

Boot Up Screen  (Yes, that is an image of Jane Austen)

Sitting Idle

After pulling out from a stop sign

A few highlights:

1. The graph shows power consumption and regeneration in half second intervals.  The bars are color coded to make for quick interpretation.  It will show about the last 30 seconds of data and continuously rotates
2. Watts appear to be the best indicator of system strain and is closest to RPMs for a gas engine.  Thus the real-time motor watt consumption is now shown with the large numerals.  They are set to change color based on the following:
1. Regeneration in effect = Green
2. Less than 12KW (Continuous power limit for the motor) = White
3. 12KW - 18KW =Yellow
4. >18KW = Red - The motor's 1 minute rating is 30KW
5. This is a computed value based on values reported by the controller - Motor Amps ( % of controller rating - 700A) x Battery voltage at the controller x PWM % (pulse width modulator duty cycle)
3. The other values are also color active to help the driver quickly spot an issue.  Much like gas cars, the driver is responsible for managing the system.  This is no different than red-lining a gas engine or ignoring a check-engine light.   The car will let you do it but should warn the driver of the situation.
4. The large "Forward" indicates whether the motor is in Forward or Reverse
5. Warnings and error messages are displayed across the bottom of the screen when needed.