I've received this by PM (the last line explains why), which takes us back to the original Pierson report. It is clearly relevant, but it's not my composition and will therefore not respond to any questions about it.
"Sorry for the length, but to understand Ed Pierson's report one needs the parts he left out.
He mixes two separate issues.
The first issue: A resolver is a variable transformer that measures the alignment of the magnetic field generated by a reference/excitation winding coil using the sense winding coils. If either the reference coil or the sense coils are interrupted, for example by a break, then the resolver can no longer function correctly.
https://en.wikipedia.org/wiki/Resolver_(electrical) There is a fourth winding that causes the current in the reference winding; it can be treated the same as part of the reference winding for discussion purposes.
To emphasize, there are two sense coils. One provides the sine of the rotation angle multiplied by the magnetic field and other constants, the other is at a right angle to the first and provides the cosine of the rotation angle. It effectively "resolves" the mechanical angular motion into these two electrical components. By comparing the magnitude of the two signals one can get the ratio and therefore the tangent and then develop the arc-tangent to get the angle.
The reason for looking at the ratio is that losses in the system will be proportional to the signal strength. However, this means if one of the sense coils develops a high-resistance, such as from a broken wire barely making contact or a corroded connector, the angle reported will not represent the actual angle. In the Lion Air report I believe the maintainers had previously sprayed some cleaner on some connectors, which did not help, before replacing it with the miscalibrated unit.
If a wire breaks completely, that can be detected during test. If the reference winding is broken no current will flow and both sine and cosine components will be zero. If the sine or cosine winding is broken then turning the vane will show one output changing and the other remaining at zero output.
Where that leads to in the Lion Air case is that a broken wire in one of the sense windings was apparently being pulled apart by the surrounding epoxy when the epoxy was cold, causing a total failure that was reported by the system.
When the AoA assembly was cold electricity could not be conducted across the gap. As the assembly heated the epoxy expanded and brought the broken ends together. This would initially allow microscopic contact, resulting in an unstable conductivity making the ratio with the other winding appear to vary even if/when the vane was not being moved. Finally, at a sufficiently increased temperature, the two ends of the wire would be forced together to conduct sufficiently to give a stable and correct reading.
The heat that was being applied to the assembly during the Lion Air bench test appeared to be from the anti-ice heater in the AoA vane.
The second issue: The AoA vane has a heater. This is a resistive element that converts electrical power into thermal energy. The system can detect if there is current flowing through that heater and if that current is zero when voltage is applied the system does the Power = Current * Voltage-drop calculation and determines that with zero current there is zero electrical power and therefore the unit cannot produce heat. There is apparently an integral regulator to limit the maximum temperature and prevent melting the AoA vane. This failure is part of the anti-icing failure system to report.
Where Pierson fails is when he claims the heater that affected the failed winding at Lion Air has anything to do with the heater error that was reported after the AoA vane was apparently torn from the aircraft in Ethiopia.
What is certain is that the Ethiopian plane did not have a history of gradually increasingly unreliable AoA readings; in fact the AoA reading was accurate until after take-off, at least to the extent it was reading as expected and agreed with the second AoA sensor. Had there been a wire break in the resolver it would not have become the exact value that dropping against the high AoA stop, which happens when the counterweight no longer had a weight balance from the AoA vane. However that is exactly what is expected were the AoA vane torn off by a bird strike. In addition, losing that vane would also lose electrical continuity through the heater, causing the anti-icing failure warning, which is what the Ethiopian data showed.
When Ed Pierson writes: "Why did the sensor signal go unstable when heater power was removed during testing of the “removed AOA Sensor” from the Lion Air airplane?" he is mistakenly linking the effect of heat from a heater that is not part of the resolver and is trying to build on that false assumption. Had the resolver been put into a toaster-oven and no power applied to the anti-ice heater the resolver would have acted the same. I doubt that he would then blame toaster ovens as being part of a conspiracy. It went unstable because the epoxy slowly cooling gradually reduced the force with which the wires were held together, gradually increasing the resistance until a gap formed.
Note that the 60C temperature was a moving target. At some point it worked at 20C, then at 30C, 40C, and 50C. The damage done by forcing the broken wires into end-to-end contact will accumulate to where it will eventually never function. I have personally seen a mechanical system fail the same way - differential thermal expansion causing cumulative damage and a sliding window on the ability to function until it no longer could.
Where he really runs off the rails is "a factory environment under duress with a shortage of electricians" because the factory electricians do not manufacture resolvers, electrical technicians at the supplier do. There is nothing at all having to do with the factory environment that affected either crash involved AoA sensor. Honestly if it takes a Scanning Electron Microscope to see a defect and seeing that only after destructive disassembly, that's beyond what a typical Boeing assembler can detect on the factory floor.
I expect that the Blacksburg information will be used to guide remove and inspect/replace with new resolvers based on which technician or which process was being followed per which serial numbers, alongside a design review to ensure it cannot recur.
I can agree with his righteous indignation over how a factory is run and how procedures are developed, but they have no bearing on the two crashes.
I'm on the banned-from-posting list, so it may be that direct messages to me get bounced. That's a choice by the moderators and I respect it. It's their platform and up to them what to make public."
MechEngr is offline