Feature Article in《Image Laboratory》Magazine — The Role and Quality Importance of High-Speed Cables in Machine Vision Systems

(Figure 1) Machine Vision System

In machine vision, the cable is not merely a connection medium; it is a vital channel for data transmission. It directly influences the integrity of image signals, the transmission rate, and the stability of the entire system. This article will explore the factors that affect the high-frequency characteristics of cables and propose several design guidelines.

The ease of judging signal correctness in a digital communication system can be measured by the Signal-to-Noise Ratio (SNR). When the signal energy significantly exceeds the noise (SNR >>60dB), the signal judgment accuracy approaches 100%. Conversely, as the noise energy increases, the signal judgment accuracy decreases.

     

The magnitude of the signal energy is determined by the power of the signal generated by the source and the loss incurred in the transmission channel. Cable loss is the main source of channel loss and can be classified into two types: Return Loss (RL) and Insertion Loss (IL).

Return Loss (RL) and Impedance Matching

Return Loss is the amount of energy reflected to the source due to impedance mismatch in the transmission channel or interface. If the impedance of the cable, connector, or PCB trace (e.g., 50Ω、75Ω、100Ω) is not matched, a portion of the signal energy will be reflected. This reflected wave superimposes onto the original wave, distorting the original signal. Furthermore, because the reflected signal superimposes with a time delay, it introduces jitter and latency into the waveform. These phenomena degrade signal integrity, cause the eye diagram to close (Figure 2), and affect judgment accuracy at the receiving end.

RL(dB) = -20 log10(|Γ|) where Γ is Reflection Coefficient

Γ = (ZL – Z0) / (ZL + Z0), where ZL is the load impedance, and Z0 is the system's characteristic impedance.

Where ZL is the load impedance, and Z0 is the system's characteristic impedance. The greater the difference between ZL and Z0, the larger the reflection coefficient and the greater the Return Loss. Conversely, when the difference between ZL and Z0 is minimized, or when ZL matches Z0, the reflection coefficient is minimized, and the Return Loss is also minimized.
 

 (Figure 2) Eye Pattern with Impedance Match (Left) vs. Eye Pattern with Impedance Mismatch (Right)

Controlling and matching the characteristic impedance is the primary method for reducing Return Loss. Figure 3 shows the impedance distribution curve of a cable assembly. Manufacturers of high-speed connectors and cables continuously enhance manufacturing quality and stability to control and temper impedance variations. The geometric structure of a shielded twisted-pair cable, including the conductor spacing, lay length (twist pitch), equivalent dielectric constant, and shielding layer, determines the cable's characteristic impedance.
 

(Figure 3) Impedance Distribution Curve Diagram

In the "last mile," which is the cable assembly manufacturing process, the stripping, wire dressing, and soldering processes (the "Cable Management" area in Figure 3) damage the cable's geometric structure. For instance, stripping back the shielding layer, as shown in Figure 4, can drastically change the impedance. Traditional processing often relies on manual labor and simple tools, resulting in inadequate tolerance control and severe impedance variations.

Based on years of experience, COMOSS has developed an automated and human-machine collaborative work model, complemented by precision tooling. This approach minimizes the stripped length of the shielding layer (Figure 4) and ensures that tolerances are maintained within a stable control range. In addition to dimensional control, covering the processed area with materials of a higher Dielectric Constant (Dk) helps to mitigate the high impedance curve (Figure 5), thereby keeping the impedance stable within a specific range and reducing Return Loss.
 

(Figure 4) Shielding Layer Stripped Length




(Figure 5) Impedance Distribution Curve vs. Different Dielectric Constant (Dk) Materials

Insertion Loss is the phenomenon where signal energy is attenuated due to factors such as conductor resistance, dielectric loss, radiation loss, and discontinuities. The majority of this loss originates from the cable itself.
Regarding loss caused by conductor resistance, the Skin Effect dominates in the high-frequency domain, where most current concentrates on the conductor surface. Therefore, the conductivity and surface roughness of the conductor significantly affect the degree of loss. Consequently, the plating material and the plating process of the conductor are critically important. Currently, silver-plated copper wire is widely used to reduce conductor attenuation.

The stranding configuration of the conductor (Figure 6) also affects loss. For the same AWG (American Wire Gauge), a greater number of strands increases the conductor's Flexibility but also increases high-frequency signal loss. Therefore, conductor flexibility and signal loss present a trade-off, requiring optimal design based on the specific application.
 

(Figure 6) Relationship between Stranding Count and Flexibility

The solution to dielectric loss is relatively simple: the basic principle is to select insulating materials with low loss and a low Dissipation Factor (Df).

In cable applications, interference acts as noise and is a primary factor contributing to signal integrity degradation, particularly in high-speed transmission, industrial environments, and machine vision systems. Common interference sources include: Electromagnetic Interference (EMI) from external equipment such as motors, inverters, power modules, or other high-frequency devices; and pulse interference from the external environment, such as lightning surges or Electrostatic Discharge (ESD).

In the structure of high-speed cables, in addition to adopting a Shielded Twisted Pair (STP) structure for the differential signal pairs, the cable assembly as a whole also needs shielding to resist signal interference from external devices and the environment. Shielding typically involves a combination of aluminum foil shielding and metal copper wire shielding. The metal copper shielding can be either the highly flexible Spiral type or the tightly structured, less prone-to-loosen Braid type (Figure 7).

Using both aluminum foil shielding and metal copper wire shielding concurrently enables full-band interference suppression and ensures signal quality.
 

(Figure 7) Shielding Structures

Crosstalk is another common internal noise source in cables. Since multiple signals are transmitted simultaneously within the cable and numerous conductors are in proximity, different electromagnetic fields interfere and couple with each other. The shielding for differential signal pairs, similar to the overall cable shielding, can be enhanced by adding a metal braid shield alongside the aluminum foil shield to significantly reduce crosstalk between different signal pairs.

The cable shielding structure is the first step in noise suppression. The next critical steps are grounding and the Integrity of the Shield during assembly. Shielding primarily confines noise to the shielding layer, preventing it from coupling onto the signal wires and compromising signal integrity. Grounding provides a discharge path for noise energy, eliminating the noise. If the shield grounding is inadequate or absent (known as shield floating), noise energy can escape as radiation, potentially affecting its own signal or other equipment. Furthermore, shielding integrity (Figure 8) is crucial. Higher frequency signals are more likely to radiate out through even tiny gaps. Generally, gaps must be smaller than 1/10 of the wavelength to prevent signal radiation. Therefore, ensuring complete, gap-free shielding during cable processing is extremely important for enhancing noise immunity.

(Figure 8) Shielding Integrity

In addition to signal energy loss and internal/external interference, the inconsistency in signal propagation delay, or Skew, is another common issue affecting signal integrity. Skew refers to the delay difference within the same differential pair (Intra-Pair Skew) or the delay difference between different pairs (Inter-Pair Skew).

For example, in a differential pair signal, if the transmission times of the positive and negative signals are not synchronized, the receiver cannot completely cancel the common-mode noise, leading to bit errors.

The primary cause of Skew is mismatched conductor lengths, which can arise from varying twist pitches in twisted pairs or unequal tension applied to the two wires during stranding. These manufacturing processes can lead to length discrepancies within a pair. Consequently, in higher-speed transmission applications (where signal integrity requirements are more stringent), the cable structure is commonly switched from Twisted Pair to Coaxial or Twinax (Figure 9) to mitigate Skew.
 

(Figure 9) High-Speed Cable Structures

Conclusion
Today's machine vision cameras are no longer low-resolution, low-rate devices. As resolution and frame rates increase, the data volume grows exponentially. This massive amount of data must be transmitted through the cable to the processing unit in an extremely short timeframe. If the cable fails to support high-speed transmission, it will cause image delay, frame dropping, or even data errors, thereby compromising inspection accuracy and production efficiency. This is precisely why the quality of high-speed cables is critical. This article has explained the common factors affecting cable performance in the high-frequency domain and the corresponding solutions, hoping to be beneficial for your future applications.

References: VESA DisplayPort (DP) Standard Version 1.4

This Article was Published in the Japanese Magazine《画像ラボ》NOV.