This video details the repair of an Agilent MXG Vector Signal Generator (N5182A) that exhibited a "Fractional-N PLL unlocked" error, causing signal loss at specific output frequencies, particularly below 5 GHz. The diagnostic process involved a systematic teardown and analysis of the instrument's complex internal block diagram, focusing on the Phase-Locked Loop (PLL) responsible for frequency synthesis. Through extensive probing with an active RF probe and a spectrum analyzer, the video verified the functionality of the Voltage Controlled Oscillator (VCO), frequency doubler, and filter bank switches across their operating ranges. The custom fractional-N divider and phase detector were also meticulously tested in isolation by overriding the PLL's tuning voltage. The root cause was eventually narrowed down to a degraded gain in the Gallium Arsenide phase detector (Agilent 1GC14218), leading to loop instability. After replacing the faulty component with one from a donor board, the instrument successfully locked across its full 250 kHz to 6 GHz range, passed all internal self-tests, and exhibited phase noise performance largely consistent with the manufacturer's datasheet.
Repairing the Agilent MXG Vector Signal Generator: A Deep Dive into PLL Diagnostics
0:09 Instrument Introduction: The video focuses on repairing an Agilent MXG Vector Signal Generator, model N5182A, a 6 GHz synthesizer capable of arbitrary waveform generation for RF carriers.
0:50 Problem Identification: The instrument initially works at higher frequencies (5-6 GHz) but shows a "Fractional-N PLL unlocked" error and loses its output signal when tuned below approximately 3.1 GHz (and other specific bands).
1:47 Initial Diagnostics with Spectrum Analyzer: A continuous sweep with max hold on a Siglent spectrum analyzer (SSA5085A) clearly maps the working and non-working frequency bands, indicating intermittent PLL lock.
3:18 Block Diagram Analysis: The repair focuses on the instrument's complex internal block diagram, particularly the Phase-Locked Loop (PLL) which includes a custom VCO (750 MHz - 1.5 GHz), frequency doublers, filters, a fractional-N divider, and a phase detector. Most components reside on a single, large board.
10:15 Initial Component Verification: Using an active RF probe and spectrum analyzer, components like the VCO, doubler, and filter bank were individually tested within the instrument's working frequency ranges (e.g., 2 GHz output, 1 GHz VCO frequency). All signals appeared correct and clean.
20:11 Frequency Sweep Test: When tuning the instrument downwards, the PLL consistently unlocked around 1.9 GHz. This initially led to suspicions about the filter bank.
24:07 Filter Bank Verification: The filter bank and its switches were isolated from the loop and measured with a network analyzer (Siglent SNA5014A). Results confirmed they were functioning correctly and switching at 1.8 GHz, indicating the 1.9 GHz unlock was a coincidence, not a direct filter issue.
26:01 VCO Functionality Test: The PLL loop was broken, and the VCO's tuning voltage was overridden by an external source. This test confirmed that the VCO could sustain oscillation and tune across its entire desired frequency range (750 MHz - 1.5 GHz), ruling out the VCO as the problem.
30:31 Fractional-N Divider Test: The VCO frequency was manually set to an "unlocked" frequency (e.g., 1.6 GHz), and the synthesizer was programmed to request that frequency. The fractional-N divider was observed to correctly produce the 50 MHz comparison frequency for the phase detector, suggesting it was functional.
32:40 Problem Narrows to Phase Detector: With all other major PLL components verified, the phase detector (Agilent 1GC14218, Gallium Arsenide) became the primary suspect. The hypothesis was that its gain had degraded, causing loop instability.
34:16 Solution: Phase Detector Replacement: A replacement phase detector was carefully desoldered from a donor board and installed on the main board.
35:55 Repair Confirmation: After replacement, the instrument successfully locked and produced a stable output across its entire frequency range (250 kHz to 6 GHz), resolving the "Fractional-N PLL unlocked" error.
37:12 Self-Test and Performance Verification: The instrument passed all internal self-tests. Phase noise measurements at 1.9 GHz showed performance generally consistent with the datasheet, though slightly degraded at some offsets and improved at others, which could be attributed to the unit's age or slight model variations.
39:56 Key Takeaway: The repair highlighted a subtle failure mode where a degraded component gain within the PLL feedback loop led to instability and an "unlocked" condition, an unusual but fixable problem.