Common-mode voltage (\(V_{cm}\)) represents the average voltage of two signal lines with respect to a common reference point, typically ground. It is essential in differential signal systems, power electronics, and EMC testing.

The common-mode voltage is mathematically defined as:
\(V_{cm} = \frac{V_+ + V_-}{2}\)

Where:
– \(V_+\) is the voltage on the positive signal line relative to ground
– \(V_-\) is the voltage on the negative signal line relative to ground

To measure Vcm:

• Oscilloscope: Connect probes to \(V_+\) and \(V_-\), measure each with respect to ground, and compute the average.
• Differential Probe: Use to extract both differential and common-mode signals.
• Multimeter: For DC or low-frequency, measure \(V_+\) and \(V_-\) relative to ground and average them.

Measuring Vcm helps identify noise, verify balance, and improve performance.

Common-mode impedance (\(Z_{cm}\)) is the impedance that common-mode currents encounter between signal lines and ground.

To measure \(Z_{cm}\) accurately, use an impedance analyzer (e.g., Keysight E4990A). Connect the two signal lines together to form a common-mode path, then measure the impedance between this node and ground.

Alternatively, use a Vector Network Analyzer (VNA) with a mode-conversion setup to measure \(Z_{cm}\) using mixed-mode S-parameters.

For a basic estimation, inject a common-mode signal into the system and measure voltage and current. Then calculate:
\(Z_{cm} = \frac{V_{cm}}{I_{cm}}\)

Proper grounding and shielding are essential to reduce noise and ensure accurate readings.

Insertion Loss (IL) is a fundamental parameter used to describe how much signal power is lost when a component, such as a filter, connector, or cable, is inserted into a transmission path. It is typically expressed in decibels (dB) and is crucial in high-frequency, RF, and communication systems, where maintaining signal strength and quality is important.

The insertion loss quantifies the difference in signal amplitude or power before and after the device under test (DUT). The voltage-based formula for insertion loss is:
\(\text{Insertion Loss (dB)} = 20 \log_{10} \left( \frac{V_{\text{in}}}{V_{\text{out}}} \right)\)

Where:
\(V_{\text{in}}\) is the voltage before the DUT
\(V_{\text{out}}\) is the voltage after the DUT

For power-based measurements, the equation becomes:
\(\text{Insertion Loss (dB)} = 10 \log_{10} \left( \frac{P_{\text{in}}}{P_{\text{out}}} \right)\)

A higher insertion loss indicates greater signal attenuation. For example, an insertion loss of 3 dB means approximately 50% of the power is lost. Insertion loss is often measured using a Vector Network Analyzer (VNA), which directly measures transmission parameters such as S21.

A Large Language Model (LLM) is a type of artificial intelligence designed to understand and generate human language. LLMs are built using deep learning techniques, especially the transformer architecture, which enables them to analyze context and generate coherent text.

Below is a simplified diagram showing the core of an LLM — the Transformer Block:

Diagram: The transformer architecture behind large language models.

🔹 1. Input Embedding
Text is first converted into a format the model can understand — numerical vectors — using a process called embedding. These vectors represent the meaning and position of each word.

🔹 2. Multi-Head Attention
This is how the model determines which words in the input are most relevant to each other. It can focus on multiple relationships at once, which helps it understand context.

🔹 3. Feed Forward Network
After attention is applied, the data is processed through a small neural network to refine the result.

🔹 4. Output Embedding
The final output is converted into predictions — such as the next word in a sentence — or passed on to deeper layers for further processing.

LLMs like ChatGPT use stacks of these transformer blocks to power natural conversations, summarize information, write code, and more — all by predicting the most likely next token in a sequence.

Leakage current refers to the current that flows from live conductors to the ground line through insulation materials or capacitors under normal operating conditions. In electronic devices, this current primarily flows through Y capacitors (safety capacitors) from the live or neutral line to the protective ground line. The leakage current is directly proportional to the Y capacitor value, following the formula:

Practical Examples (230V/50Hz grid):

As demonstrated, when the Y capacitor value doubles, the leakage current also doubles proportionally.

Risks of Increasing Y Capacitor Values

While increasing Y capacitor values can improve EMI performance, it introduces several concerns:

Safety Standard Considerations

Different application scenarios require compliance with different safety standards, each with varying leakage current requirements, thus affecting Y capacitor selection:

Design Recommendations