There is a limited amount of information released about the onset of the LN-OJF accident, but I have noted some similarities between it and the G-REDL accident, and I want to share my observations. First the source material:
G-REDL Accident Report:
Directly after the loss of MGB oil pressure, the helicopter deviated from its cruise conditions. With only four seconds of flight data (before CVFDR recording seized), analysis of the helicopter response was limited. During these four seconds, the helicopter initially rolled left to -11.3° before rolling right to 9.1° within two seconds. Magnetic heading increased by 30 degrees.
The recorded flight data was provided to the manufacturer who ran a simulation to analyze the consistency between flight control inputs and the helicopter response during the final seconds of operation. The results indicated that, two seconds after the MGB oil low pressure warning, the helicopter was no longer responding normally to flight control inputs.
Recorded fault messages downloaded from the SMDs revealed that, at approximately the same time as the final radio transmission (20s after MGB oil low pressure warning), data from both the Heading Sensor Units (HSU) and the radio altimeter was lost to all four SMDs concurrently. The radio altimeter is located in the tail section of the helicopter on the shelf above the CVFDR, and the HSUs further aft.
Damage to the tail boom of the helicopter confirmed that it had separated from the fuselage prior to the impact with the sea. It had suffered from multiple main rotor blade strikes on the boom and at the base of the fin. Damage to the tail rotor drive shaft was consistent with the shaft rotating when it was struck by the main rotor blades.
LN-OJF Preliminary report:
The recordings on the CVFDR showed that everything appeared to be normal until a sudden catastrophic failure developed in 1-2 seconds. The CVFDR recordings ended abruptly at the same time. A spectral analysis of the CVR data has been carried out. There was no obvious indication of an abnormality before the sudden detachment of the rotor head.
LN-OJF Eyewitness statement:
I was looking at the helicopter as it was flying normally. Then after a few seconds it started veering back and forth quite substantially followed by 3-4 quite loud bangs where we then clearly see parts coming off of the helicopter. A few seconds after that there is one additional bang after which the helicopter goes straight down and the rotor has come off.
In summary, it seems that two things happened shortly after commencement of the respective catastrophic events:
1. Both aircraft apparently veered back and forth initially.
I will not attempt to try and explain the exact reason for the "veering" in either accident, but I found it interesting enough to remark the similarity.
2. The CVFDR on both aircraft seized recording after 2-4 seconds.
As for the CVFDR stopping so prematurely in both cases, consider the following:
- In the G-REDL case it has already been documented that the MR blades struck the tail-boom multiple times, and that it eventually seperated prior to impact with the sea.
See photoshop illustration of G-REDL.
- In the LN-OJF case I believe that the MR blades struck the tail-boom as well. I base this belief on my observations of the wreckage images and videos released publicly. From watching the chilling video of the
falling rotor hub
frame by
frame, I'd argue that one can see several blades damaged (some severely)
before hitting the ground. The approximate pattern of damage can then be confirmed by
looking at the
images of it on the ground. Damage to the tail-boom itself can be
seen in
several different
images, as well as in
this video where it is being hoisted out of the water.
So what does this have to do with the CVFDR stopping abruptly? Well, while I cannot say for sure which CVFDR was installed in G-REDL, the one installed in LN-OJF have been
pictured in press releases. It is a combined CVR and FDR from Honeywell, see
manual here. The location where it is normally installed in the helicopter (L1, L2 & 225) is the forward port side of the tail/cargo-hold wall as far as I understand it (see final image). To avoid overwriting the latest data after a crash, the CVFDR will stop recording if:
1. It receives no power
The CVFDRs in the puma fleet should
AFAIK be connected "direct battery", meaning it is among the very last of the essential electrical consumers that should lose power after failure of the generators.
2. Its immersion switch is triggered
It was not immersed in water at the time it seized recording.
3. Its inertia switch is triggered (6g)
I propose that one or more MR blades hit the tail-boom and the resulting G-load, however short, triggered the inertia switch. This, in my opinion, is the reason it stopped.
Photoshop illustration of LN-OJF.
TL;DR: I note some similarities in initial aircraft behaviour as well as hypothesize that the reason for the CVFDR seizing recordings only seconds into the catastrophic events for both G-REDL and LN-OJF, is due to MR blade strikes on the tail-boom and thereby triggering the CVFDR's inertia switch.