Hi,
I have two things to respond to in this thread, so I'll start with the Saft battery warnings.
We monitor the voltage level to determine the remaining capacity of the battery. This isn't perfect, but it works fairly well for most battery types. With the Safts (or any Li-SOCl2 chemistry battery), they have a very flat operating voltage plateau for most of their life, then die relatively quickly towards the end. By relatively quickly, I mean an operating life of maybe 100 hours, then dying over a few hours time. So it should be possible to give a warning a few hours in advance.
However, with the Saft batteries, while the voltage plateau level is flat its actual level varies greatly with several factors. Temperature and current draw are two of the biggest factors affecting the voltage plateau. We need to detect a drop of about 0.15V to determine the battery is dying, but temperature and current draw can cause the voltage to fluctuate by over 0.3V! Fortunately, we can measure temperature and estimate current draw based on screen brightness and therefore compensate the warning voltages for these effects (this is why when someone asks at what voltage a Saft should be replaced we cannot give a simple number).
The problem you are seeing where the Saft battery gauge indicator is dropping from full to the battery completely dying is probably because the current draw compensation is not tuned correctly. Improvements to efficiency have been made so current draw is now lower. However, the tuning has not been updated. Most of us at Shearwater are using 1.5V alkaline or 1.5V Energizer Advanced Lithium batteries since they offer better value, so we haven't noticed this (these battery types aren't nearly so sensitive to current draw).
I do appreciate hearing about this. We are working on a firmware release cycle, so we will be sure to measure and adjust this tuning as needed.
Best regards,
Tyler Coen
Shearwater Research
---------- Post added June 4th, 2015 at 07:27 PM ----------
This response is regarding the NDL quickly dropping due to small depth variations during a slow ascent.
We received the log files (thank you), and while we have not yet performed complete analysis, it does not look unreasonable to me. We will still perform detailed re-simulations and verify the NDL numbers are correct.
The situation I believe is happening is this:
- A dive is performed to around 90ft. You stay there for a while and NDL goes down because tissue nitrogen loading has gone up.
- NDL gets down to a few minutes.
- The ascent then starts.
- As you get shallower, the NDL starts going up. However, it is important to remember for most of the tissue compartments, the nitrogen tissue loading is still pretty high. That is, you were almost in deco when at depth, so except for the fastest tissue compartments, most compartments still have significant nitrogen loading.
- The ascent is slow, maybe looking at a reef, just taking some time. So some tissue compartments are still on-gassing, or just balanced around on-gassing and off-gassing.
- You can get in situations where a bunch of factors are aligned, like the current tissue loading, on/off-gassing rates, ascent rate, M-values, etc. such that ascending or descending just a few feet can impact NDL significantly.
- Note that this only happens when ascending towards the end of a dive and tissue loading is high. You don't see this near the start of a dive at these depths, because tissue loading is then low.
Like I mentioned, we will of course perform a more detailed review to ensure there are no bugs or other problems here. This will take a few days. I just wanted to share my first analysis, which is that it seems reasonable and I have seen this before (during development simulations mostly).
Please let me know if you have any questions.
Best regards,
Tyler Coen
Shearwater Research