As an automotive technician specializing in advanced driver assistance systems (ADAS), I recently encountered a challenging case involving a BYD EV, specifically a 2021 BYD Song PLUS DM-i hybrid vehicle. The owner reported that after accident repairs, the instrument cluster displayed a “Lane Departure Function Limited” warning, and during driving, features like lane keeping and adaptive cruise control were inoperative. This BYD car had accumulated approximately 20,641 km, and the issue arose post-collision, where components such as the front bumper and front millimeter-wave radar were replaced with non-genuine parts without proper programming or calibration. In this article, I will detail the diagnostic and repair process from my first-person perspective, emphasizing the complexities of ADAS in modern BYD EV models. I will incorporate tables and formulas to summarize key concepts, ensuring a comprehensive understanding of the system’s operation and fault resolution. Throughout, I will frequently reference BYD EV and BYD car to highlight the brand’s technological advancements and common issues.
The ADAS in this BYD car is supplied by Bosch and comprises a front millimeter-wave radar and a multi-function video controller. These components work in tandem to enable functions such as lane departure warning (LDWS), lane keeping (LKS), traffic sign recognition (TSR), high-beam assist (HMA), and intelligent cruise control (ICC). The system relies on sensors, including cameras and radars, to perceive the environment, collect data, and perform real-time analyses. For instance, the radar detects objects and measures distances, while the camera captures visual data for lane detection. The integration of these elements is governed by complex algorithms that can be represented mathematically. For example, the sensor fusion process for object tracking might involve a Kalman filter, which can be expressed as:
$$ \hat{x}_{k|k-1} = F_k \hat{x}_{k-1|k-1} + B_k u_k $$
$$ P_{k|k-1} = F_k P_{k-1|k-1} F_k^T + Q_k $$
where \( \hat{x} \) is the state estimate, \( F \) is the state transition matrix, \( B \) is the control input matrix, \( u \) is the control vector, \( P \) is the error covariance, and \( Q \) is the process noise covariance. This mathematical framework ensures accurate predictions and updates in the BYD EV’s ADAS, but any misalignment or fault can lead to malfunctions, as observed in this case.
Upon initial inspection, I used a diagnostic tool (VDS2100) to scan the vehicle’s systems and found a historical fault code U014086 in the multi-function video controller, indicating a BCM data fault. This code typically relates to communication errors between modules in the network. The BYD car employs a domain-controlled network architecture, with controllers for left, right, and rear body domains managing various low-voltage electrical functions. The fault could stem from software issues, hardware failures, or wiring problems. To systematically analyze the potential causes, I considered factors like incorrect software versions, radar calibration errors, or damaged wiring harnesses. The table below summarizes the primary fault hypotheses and their probabilities based on my experience with BYD EV models:
| Potential Cause | Description | Probability |
|---|---|---|
| Software Mismatch | Incorrect coding or version in ADAS modules | High |
| Hardware Failure | Defective radar or video controller | Medium |
| Wiring Issues | Damaged or repaired harnesses affecting communication | Medium |
| Calibration Errors | Improper dynamic calibration of sensors | Low |
Beginning the diagnostic process, I first addressed the front millimeter-wave radar, which had been replaced with a non-genuine part. Using the VDS2100 tool, I wrote the correct vehicle configuration and performed a dynamic calibration on the road. This involved driving the BYD car for about 10 minutes until the calibration progress reached 100%. Post-calibration, the adaptive cruise control functioned normally, but the lane departure features remained inactive. This partial success indicated that the radar itself was operational, but other factors were at play. The calibration process can be modeled using optimization formulas, such as minimizing the error between sensor outputs and expected values:
$$ \min_{\theta} \sum_{i=1}^{n} (y_i – f(x_i, \theta))^2 $$
where \( \theta \) represents calibration parameters, \( y_i \) is the observed data, and \( f(x_i, \theta) \) is the sensor model. In this BYD EV, the radar’s alignment was correct, but the persistent fault suggested deeper issues.
Next, I examined the wiring harness connected to the front millimeter-wave radar. The harness showed signs of previous repairs, with multiple splices and potential weak points. Referring to the circuit diagram, I identified six wires: one power, one ground, and four CAN bus lines (two for radar subnet and two for chassis network). Using a multimeter, I measured voltages: the power supply was normal at around 12 V, ground was secure, and CAN bus voltages were within specifications (e.g., CAN_H at 3.1 V and CAN_L at 1.91 V for the radar subnet, and CAN_H at 2.75 V and CAN_L at 2.25 V for the chassis network). The CAN bus communication can be described by differential signaling equations:
$$ V_{diff} = V_{CAN_H} – V_{CAN_L} $$
where \( V_{diff} \) should typically be around 2 V for proper operation. In this BYD car, the measurements were normal, ruling out obvious wiring faults. However, to be thorough, I replaced the entire front bumper harness and conducted a road test. Unfortunately, the lane departure malfunction persisted, leading me to reconsider the fault code U014086.
This fault code, starting with ‘U’, points to network communication issues, often involving data exchange errors between modules. In the BYD EV’s domain-controlled system, the right body domain controller integrates the BCM functions. I used the VDS2100 to reconfigure the left and right domain controllers, but this did not resolve the issue. The communication network in such BYD car models relies on CAN protocol, where data frames are transmitted with identifiers and data fields. The error detection in CAN involves cyclic redundancy check (CRC), calculated as:
$$ CRC = \text{remainder} \left( \frac{M(x) \cdot x^n}{G(x)} \right) $$
where \( M(x) \) is the message polynomial, \( G(x) \) is the generator polynomial, and \( n \) is the degree. Faults like U014086 can arise from mismatches in these data frames, prompting me to investigate software aspects further.
Given that the front windshield was not replaced in the accident, I focused on the multi-function video controller. Using VDS2100, I checked its software version and found it did not match the vehicle’s specifications. Normally, the configuration should display “no fault,” but this unit had missing subsystems and incorrect coding. The software version incompatibility can be represented as a version mismatch equation:
$$ V_{actual} \neq V_{expected} $$
where \( V_{actual} \) is the installed version and \( V_{expected} \) is the required version for the BYD EV model. Attempts to reflash the software were blocked due to version conflicts, confirming the controller was faulty. I then replaced the multi-function video controller, performed a new calibration, and verified that all ADAS functions, including lane departure, were restored. The image below illustrates a typical ADAS setup in a BYD car, highlighting the integration of sensors and controllers:
To generalize the diagnostic approach for similar issues in BYD EV models, I developed a table outlining key steps and tools. This can assist technicians in systematically addressing ADAS faults:
| Step | Action | Tool Used | Expected Outcome |
|---|---|---|---|
| 1 | Initial fault code scan | VDS2100 | Identify stored codes like U014086 |
| 2 | Radar calibration | VDS2100 and road test | Restore basic cruise functions |
| 3 | Wiring inspection | Multimeter and visual check | Ensure no physical damage |
| 4 | Software version check | VDS2100 module readout | Detect mismatches in controllers |
| 5 | Component replacement | New multi-function video controller | Full ADAS functionality |
In conclusion, repairing collision-damaged BYD car models requires a holistic approach that considers hardware, software, and network communications. The fault in this BYD EV was multifaceted: initially, the uncalibrated radar caused partial failures, but the core issue lay in the multi-function video controller’s software encoding. This case underscores the importance of verifying software versions and performing comprehensive diagnostics in modern vehicles. For BYD EV owners and technicians, understanding these intricacies can prevent prolonged downtimes and ensure optimal performance of ADAS features. The resilience of BYD car systems depends on precise integration, and as technology evolves, continuous learning and adaptation are essential for effective troubleshooting.
Moreover, the mathematical models and formulas discussed, such as those for sensor fusion and communication protocols, provide a foundation for analyzing ADAS behavior in BYD EV models. For instance, the overall system reliability can be quantified using failure rate formulas:
$$ \lambda(t) = \lambda_0 e^{-\beta t} $$
where \( \lambda(t) \) is the failure rate at time \( t \), \( \lambda_0 \) is the initial rate, and \( \beta \) is a decay constant. By applying such principles, technicians can better predict and address issues in BYD car ADAS, enhancing safety and customer satisfaction. This experience with the BYD Song PLUS DM-i has reinforced the need for meticulous attention to detail in every aspect of automotive repair, from initial assessment to final validation.