Yokogawa AQ6370 optical spectrum analyzer repair

The original AQ6370 still measures well when it is working. 600 to 1700nm range, 0.02nm resolution, close-in dynamic range that held up against what was available at the time. Thirteen built-in analysis functions, Ethernet remote control, universal connector interface. Plenty of labs and production environments are still running these because the measurement capability covers what they need and replacing a functioning instrument is not a priority.
What has changed is the repair landscape. The AQ6370B, C, D, and E all have better parts availability and more active service infrastructure. The original AQ6370 is older than most of the technicians at instrument service depots. We have repaired enough of these to know exactly where they fail and what it takes to bring them back.
What Actually Goes Wrong on an Instrument This Old
The monochromator drive is the first thing to think about. The original AQ6370 uses a DC motor to drive the diffraction grating through a servo system with encoder feedback. After fifteen to twenty years of sweep cycles, the motor bearings wear, the encoder signal degrades, and the servo loop starts fighting itself. The symptom is usually intermittent sweep failures, inconsistent wavelength accuracy that varies between warm and cold starts, or the unit hanging during the internal auto-alignment routine. A fresh calibration does not fix a failing drive. The mechanical problem comes back.
Capacitor aging on the power supply and signal processing boards is a real issue at this age. Electrolytic capacitors in the power rails and analog circuits drift in capacitance and ESR over time. The symptoms are subtle at first: slightly elevated noise floor, level readings that are within spec but not as stable as they used to be, occasional unexplained measurement variance. By the time caps are visibly failed or causing hard faults, the board has usually been running degraded for a while. We check these during diagnostic on any AQ6370 that is this old.
The universal connector interface on the original AQ6370 is a different design from the later models. Users can replace the connector adapters themselves, but the underlying input optics and alignment mechanism behind that interface wear independently. An instrument that has been connecting fibers multiple times a day for fifteen years has a different input path condition than a lightly used unit from the same era.
The built-in reference source for wavelength calibration and auto-alignment weakens over time like any light source. On an instrument this age, if the auto-alignment routine is taking longer than it used to, or the instrument needs more warm-up time before wavelength accuracy stabilizes, the reference source is the first place to look.
Software and firmware on the original AQ6370 is fixed. There are no updates. If the internal software has corrupted storage or the system drive has degraded, recovery requires source media that is no longer distributed. We have what we need to handle this. Most service facilities do not.
How We Approach It
Older instruments need a longer diagnostic. We do not assume the only problem is what you described when you shipped it. DFB reference measurements at 1310 and 1550nm establish wavelength accuracy baseline. Sweep speed gets timed and verified. Dynamic range gets measured. The power supply rails get checked. Capacitor condition on critical boards gets assessed. Communication interfaces get tested. Input connector and optical path get inspected. We tell you what we found, all of it, before any repair work starts.
Repair is at the component level. On an instrument this old, swapping assemblies when you can fix the actual component is not good practice, it just moves the aging problem around. NIST-traceable calibration with a certificate comes with every completed repair. Test report showing pre and post repair measurements comes with it.