Gas needs 999 bar

[DennisA]

During a few dives last week, the gas needs for my second bailout gas switched between 999 bar, and the expected value several times.

I have attached a few profiles.

CCR, Auto SP, 2 BO gasses, FW 2.97


Any idea what caused this?

Best regards

Dennis

[Ralph]

Hi Dennis,

what file type are the *.dive / how to open them?

BR
Ralph

[DennisA]

It is the profile dump from ostc tools
http://www.angermayr.eu/index.php/en/so ... ctools-win

[Ralph]

Ok, seems so i need some more information on your OSTC settings: how did you set the OC (bailout) gases: Settings "first" / "deco" and switching depth for the two gases. Further more: have you activated delayed ascent?

The gas needs are computed from the stops table and the ascents, plus the need during the delayed ascent phase at bottom depth if configured. 999 bar is the highest value that can be displayed, so internal calculation may have computed an even higher number. This usually happens in case the stops table "explodes" due to none-availability of appropriate deco gases, which in turn may happen if the configured switch depths do not allow the OSTC to use a deco gas although it "is there". Opposite to the normal stops table, which can be viewed through one of the custom views, there is no way to view the internal bailout deco stops table as such. But an exploding bailout stops table shows up by an fTTS (in fact: B/O) ascent time that goes through the ceiling.

BR
Ralph

[DennisA]

Gas1: 21/20 0m
Gas2: Nx75 11m

fB/O: 3
Calc.Gas B/O: yes
Bottom gas: 25 l/min
Decogas 20 l/min

Tank size:
Gas1: 7l @ 180 bar
Gas2: 7l @ 180 bar

I don't recal if the fB/O showed unrealistic values.
TTS was as expected.

[Ralph]

Hi Dennis,

i just ran your dive through my OSTC in deco calculator mode with 40m for 30 mins, 21/20 @ 1,3 in the loop, B/O gases 21/20 as FIRST and Nx75 as DECO for 11 meters and up. The gas prediction yields 788 liters (113 bar) on the 21/20 and 644 liters (92 bar) on the Nx75. So with all these details configured correctly there should not have been any exploding of the internal B/O stops table and pressure needs displayed should stay within above max numbers.
Honestly, i have no direct idea on the cause of the intermittent "999" right now. I will run the dive through the simulator mode and have a sharp look on the displayes, anyhow i would like to ask you to try to gather some context information in case you observe this behaviour again.

UPDATE: did manage to reproduce the error in simulator mode. Now the search for its cause can start...

thanks & best regards,
Ralph

[DennisA]

That's excellent! Good luck hunting!

[Ralph]

First hunting results: The faulty code section has been identified, the bug affects the 2nd plan only.

--> Normal deco information is not affected. <--

Under certain circumstances, the switching of the bailout gas during the initial ascent, i.e. before reaching the first required stop, is calculated in a wrong way.

Next step: will find out why the code works in most, but not all cases...

[DennisA]

Thanks for the update!


Happy hunting!

[Ralph]

Hi,

here's the final result: the OSTC calculated that you can ascent to 11 meters to switch to Nx75. Then deco builds up progressively, with stops first at 3 meters or whatever depth configured, then 6 and 3 meters, then 9, 6 and 3 meters and so on. But actually when the OSTC prompts for the first stop, the stop depth shown is not the exact depth where you reach your GF-low, but the next deeper multiple of 3 meters. So it happens that your real first stop is e.g. at 10 meters, the OSTC calculates ascent to 11 meters, gas switch here, then 1 more meter up, reaching first stop there at now 10 meters, resulting in a stop put at 12 meters. But that's actually deeper again and this makes the gas-needs algorithem producing erratic results. Once the true first stop depth became deeper than 11 meters, all was fine again.
I have sent the fix to Matthias, there will soon be a patch Version released.

BR
Ralph

[heinrichsweikamp]

Hi,

Thanks a lot for looking into this so quickly (tu)

Matthias

[DennisA]

Ralph, your dedication to the projet is admirable!

[Ralph]

Thank you!


Well, the gas-needs code and the whole rebreather bail-out functionalities were written by me, so i had a personal interest in understanding the bug and fixing it...

[heinrichsweikamp]

If someones want's to try: Attached is the 2.97SP2 with this 999bar issue fixed.

regards,
Matthias