Hi WetLens,
I responded to your email, but am posting here for others that may be interested in the NDL question.
We have reviewed the logs and did not find anything out of the ordinary.
Your logs were used to run simulations on desktop computers. We obtained the same results as displayed by your Petrel. This is not all that surprising since the simulation program has the same code base as the Petrel firmware. However, it does prove that there were no glitches or temporary errors on your unit.
It terms of validating whether the software is performing correctly as per the decompression model is a bit trickier. I have not taken the time to check this dive in great depth, looking precisely at which compartment is controlling the NDL time and how it interacts with depth, inert gas loading, on/off gassing rates, and M-Value limits.
However, the observed behaviour is reasonable for the situation. The important thing to realize in this situation is that tissue loading is quite high (as deco stops were almost needed when at 80ft). When descending to the lower depths, the tissue loading is still high for most tissue compartments. Initially the NDL is likely controlled by the fastest compartments and thus rises quickly as depth decreases. But after staying at moderate depths for a while, slower compartments become involved and the NDL becomes more constrained. I have attempted to show this in the attached diagram. The smooth lines are the tissue tensions for the fastest 7 compartments. The dashed lines show inspired PPN2 at the depths of 38ft and 43ft (about how much depth was varying when NDL was unstable). You can see that where the NDL is unstable, the tissue tensions are quite close to the inspired PPN2 values. For at least one compartment, it moves from on-gassing to off-gassing with this small change in depth. This closeness of inspired PPN2 to the tissue loading is what causes the NDL time sensitivity to small changes in depth.
Best regards,
Tyler Coen
Shearwater Research