What we now know about the airborne IMU

The six prepared recordings come from professional POS AV 510 systems and identify their IMU as type 6. There is now a direct published connection to the hardware family: NOAA researchers describe a POS AV 510 fitted with IMU-6 as a tactical-grade Litton-200 unit. This is strong support for the family identification, rather than an identification made from a different manufacturer's type-number table. Damiani and Mader, ION GNSS+ 2014 abstract

The inference for our files is consistent with the LN-200 family, rather than a proven exact part number. The manufacturer's LN-200 family description identifies three fiber-optic gyros, three silicon accelerometers and digital angle/velocity increments. Its hardware quality is appropriate to airborne inertial measurement. The exact variant, individual calibration and processing of our recordings still need confirmation. Northrop Grumman family description

Applanix's 2009 FAQ describes specially selected LN200ROM variants for its 510 systems, and notes that the same product can use different IMUs. This supports a possible variant; it does not tie LN200ROM to each of our six source hashes. Its navigation explanation also says GPS-estimated sensor errors correct the increments before navigation integration. That establishes a correction stage in the fused solution, but does not specify which side of it supplies legacy Group 4. Applanix FAQ, printed pages 2 and 6

The distinction that still matters

We need to know exactly what entered the recorded IMU stream. The same system can produce both sensor measurements and a GPS-aided navigation answer. Excellent navigation accuracy does not independently establish the accuracy or correction history of the recorded sensor increments.

The newer V6 manual describes separate raw-sensor and navigation logging, an unprocessed acquisition mode, and microsecond output time tags. These descriptions support the architectural distinction. They do not establish the legacy V5 Group-4 tap point, sensor integration period, physical latency or correction settings. Applanix V6 manual, printed pages 1-11 to 1-13

NovAtel documents LN-200 increments as 2⁻¹⁹ radians and 2⁻¹⁴ metres/second per count in its own RAWIMUSX output. Our empirical Applanix hypotheses are 0.4 arcsecond and 3.38 × 10⁻⁵ metres/second per count. The former is about 1.67% larger than that documented angular quantum; the latter is about 44.62% smaller than its velocity quantum. Different adapters can encode or rescale measurements differently. This comparison neither disproves the hardware family nor establishes which corrections Applanix applied. It prevents copying a family scale table into the legacy decoder without validating the output format. NovAtel RAWIMUSX scale table

Exact configurations to ask about

The configuration inventory associates the six existing source hashes with their recorded version strings and embedded dates. It was built from hash-verified completed evidence, without rescanning originals or inspecting gyro residuals. The S/N below is the logged POS system identifier; it is not an independently verified serial number for the installed IMU.

The technical question packet names those configurations and asks for the exact logging, scale, clock, correction and installation definitions. It is a prepared request, not a message that has been sent. A firmware-matched legacy ICD, acquisition description or calibration record could answer these questions; another generic datasheet cannot.

This review supersedes the earlier blanket statement that no public hardware-family association was available. The earlier evidence matrix remains intact as historical evidence. No quantitative calibration or scientific eligibility was promoted. All 232 originals remain, and empirical Earth fits/starts remain zero. The project can proceed with explicitly labeled processing hypotheses, but a measured Earth-model conclusion still needs the unresolved independent support.