novel
charging approach. It intends using the motor controller in recuperation mode for charging of the batteries,
with external power fed into the controller, simulating power generated by the motor.
There is a planned test of the Luka EV electric vehicle novel
charging approach. It intends using the motor controller in recuperation mode for charging of the batteries,
with external power fed into the controller, simulating power generated by the motor.
For this purpose, frequent rewiring of the motor-controller assembly will be needed, together with current and voltage and possibly waveform measurements. As-is, the accessibility of the wiring is suboptimal, with risk of accidental contact with powered batteries. The 100V DC kick is rather unpleasant to touch, and the current can draw quite an arc.
A jig to facilitate more comfortable and therefore safer work is needed.
A wooden board was chosen as the base, for the sake of availability and mechanical and electrical properties. Wood has advantages over plastic in price (important for limited-use temporary jig), thermal behavior (at elevated temperatures it will not soften like plastics but instead will brown and smoke, visually indicating overheating, and then char, while still keeping enough strength to hold the clamps in place), and electrical behavior (insulator).
There is a fire hazard in using wood, but this will be mitigated by not running the setup unattended for prolonged periods.
A 720x150 mm door-step was chosen as a prefab part due to easy availability in a local hardware store, and due to being made from good quality hard wood.
Four pairs of 6mm holes were drilled in positions matching the structural beams in the car's back, to facilitate temporary attachment by zip ties.
 Mechanical dimensions |
M8 bolts are used on the controller, the contactor and the rest of the equipment. 45mm long M8 bolts were therefore chosen for wiring posts. Large washers were used over the wood, to spread the load over the fairly soft material and allow more firm tightening. Wires were soldered to the top washers to avoid use of cable lugs that were at the moment in short supply.
The wiring posts are separated to two main groups; the main group, with the motor phases, power, and one auxiliary pair, and the aux group without assigned functions.
Fixed connections are done with half-height locknuts. Wingnuts are used for attaching the patch cables.
A pair of ground/chassis posts was added as an afterthought, located just over one of the beams. The electrical contact with bare screw heads with the chassis is not detrimental here as they will be connected to the chassis/ground anyway.
A terminal strip was attached to the end of the board. Every wiring post of the main group is attached to one of the terminals, to facilitate easier attachment of measuring instruments. The terminals are arranged in pairs.
 Assembled board |  Assembled board |  Testpoints and indicators closeup |  Testpoints and indicators closeup |
To facilitate safety, debugging and operator awareness, a board with indicators was assembled on a piece of protoboard. Due to relatively high voltages (100V) between the terminals, the unused strips of the board were ground away.
Antiparallel pairs of white 0603 LEDs were used to indicate both voltage presence and polarity. As LED diodes have rather low reverse breakdown voltage, the other LED is acting as both a protection against spikes and their indication. 0603 size was chosen as it allows placing two LEDs with comfort next to each other on a single 0.1" square protoboard pad.
A 47 kiloohm series resistor was chosen as current limiter. The white LEDs (blue LEDs in phosphor disguise) are very sensitive and start shining at even minuscule currents. At 100 volts the current is limited to 2 milliamps. The diodes produce visible light at well below a milliamp, though chip-to-chip differences are showing there. 4 volts are enough to visibly light up the LED, at 12V they are fairly bright. Even stronger mains hum capacitively coupled from a soldering iron is sufficient to dimly but visibly light the LEDs.
Gallium nitride LEDs (blue, white, "true green") are handy for applications where wide range of input voltage has to be indicated without a current-regulating active element.
A triangle of LED pairs was added to the board, with jumper pins to connect them to the A or B side of the motor group. Other LEDs are indicating voltage against the chassis, these show the relative voltage between the motor coils or motor driver outputs.
Visual indication of voltage present is important to avoid unintended contact of the operator or equipment with the exposed terminals, to prevent shock and arc and other surprises.
 LED indicator board schematics |  Indicators closeup |  Indicators closeup |
The patch cables are a necessary part of the device. Originally intended to be made of 16mm2 copper cable, but 20mm2 was chosen as the thinner one was not available at the vendor.
The cable was cut to five 25 cm long pieces for patch cables, three 65 cm long pieces for motor attachment, and one 50 cm long piece for power attachment.
The insulation at the ends was cut using a hot knife attached to a soldering iron. The cable lugs were soldered to the cables, using a propane torch as an electrical soldering iron was not powerful enough.
 Cables |  Cables, closeup |  Cables, closeup |
 Final assembly |  Final assembly closeup |  Final assembly closeup |
The board will be mounted to the back of the car, over the motor controllers. There are two struts, both of hollow square steel; one is a 40x40 mm square, one is 30x30, about 60mm higher and 365mm towards back.
The board will be attached via four zip ties to these two struts, placed in between them. The zip ties have to be tight enough to prevent the board shifting and the heads of the screws touching the struts. If this can not be satisfied, a layer of electrical insulation has to be placed over the heads.
 Proposed mounting |  Jig installation |  Jig installation |  Jig installation |
 Jig installation |  Jig installation |  Jig installation |  Jig installation |
 Jig installation |  Jig installation |  Back of the jig |  Back of the jig |
The high power lightbulb limiter
was first tried
as a test load for the battery. The lamps are rated to 230 volts, the battery has
100 volts only, so the lamps are underheated. Current was not measured yet.
 Brief discharge testing |  Brief discharge testing |  Brief discharge testing |  Brief discharge testing |
 Brief discharge testing |  Brief discharge testing |  Brief discharge testing |
As the alligator clip cables were too short, a mains power cable was used. Its IEC end was cut off and a pair of M8 lugs was attached to the L and N wires. The cable was attached to the jig using wing nuts, to the U and V lines of the motor controller.
 Power cable attachment |  Power cable attachment |
The variac was used as a voltage stepdown. The series
lightbulb limiter, first the low power and then the
high power one, was connected in series to limit the
current flowing into the assembly and protect the controller and battery.
 Test with variac |  Test with variac |  Test with variac |  Test with variac |
It was decided to skip the variac, and connect the car directly to the mains, with only the current limiter. The low-power unit was used first, to check the wiring, then was replaced with the high-power one. A kilowatt lamp was bought for pushing the test further.
The low-power limiter was kept in series as the high-power one lacked attachment points for the ammeter. The diode bridge and the ammeter shunt are rated for one amp only, so the high current tests were kept short. The low-power limiter expressed its displeasure with such gross overload with a funny smell but did not lose functionality.
 Test with limiter only |  Test with limiter only |  Test with limiter only |
It was found that at lower voltages the current is minuscule, then increases fast at about 90 volts and the voltage after the current limiter tapers off at 108 volts, the voltage on the battery, and does not increase past that value even with the variac maxed out.
The car in essence behaves like one big fat Zener diode.
With two kilowatts worth of halogen lamps in series with the car, it takes five amps of current.
Turns out that the controllers do not take more current in regenerative braking mode, against expectations. In effect the controllers behave like glorified diode bridges.
The car chassis was found to be floating at about 140 volts AC. This is a safety risk, acceptable for testing or emergency operations but not for actual use. An isolating transformer or disconnecting the battery would be needed for actual deployment.
 Measured values |  Measured values |