CCR dive with AI has confused deco calcs

[jimbodude]

I'm importing CCR dives with AI.  After assigning cylinders to the imported gasses when using Air Dil and Air bailout, I see incorrect deco calculations, and the dive gas appears to have changed to Oxygen instead of using the reported PO2 and showing the Diluent as the dive gas.

Notice in this screenshot:

• In the "Information" box at the 48min point, the dive computer is reporting a deco ceiling and stop, where the calculated result is an NDL of ">2hrs"
• The dive gas in the graph section is "oxygen", not "air"
• We have correct PO2 data represented by the green line (but the yellow setpoint line is always 0.5 on the Nerd, because it isn't the setpoint controller - that's not a problem)
• Adding other cylinders with the correct CC gas doesn't change any of the calcs.

This does not appear to be a problem when using multiple BO options and trimix, though Subsurface assumed the transmitters were related to BO gasses, which I corrected.  The result is some extra cylinders defined. BO gasses are not included, which is fine because I didn't use them, but the imported state was messy, with BO gasses using the transmitter values.

Notice in this screenshot:

• The first 2 cylinders were imported as 2 of my BO gasses (Air and 50%), but they had the AI information.  I set the cylinder type and changed them to 15/45 and O2 to match the actual usage.
• This left the 2 copies of 15/45 gasses - I assume one for BO and one for CC, which matches what's defined in the computer.
• In this configuration, the calculations in the Information box match the computer-reported values pretty closely, the dive gas shows up correctly in the chart, and the tissue loading chart looks normal, but the cylinder list is wrong.

I have a Shearwater Nerd 2 with 2 AI transmitters connected, one for Diluent and one for O2.  I have defined other gasses on the Nerd for BO.  In the latest Nerd firmware, you can define names for the transmitters - mine are "D1" and "O2".

I know there isn't a direct relationship on the computer to connect the defined gases to the transmitter value, which is leading to this confusion.  Is there a more correct import workflow I should be using?  Is there a bug here?  Is there some improvement that could be made to the import process, now that there are gas names?  I think I still have a functional build environment if anyone has ideas to try or some guidance on where I could start to understand what's going on.

Imported on:

Subsurface-mobile, Android: 6.0.5112

Displayed on:

Subsurface, Mac: 6.0.5299-CICD-release

[Michael Keller]

Hi Jim.

On 16/02/25 18:36, jimbodude wrote:

After assigning cylinders to the imported gasses when using Air Dil and Air bailout,

By any chance, are you 'assigning cylinders to imported gases' by telling Subsurface that the gasmix of the diluent reported by your dive computer was incorrect, and that you actually did this dive using a 'diluent' of 100% oxygen? Because this is how your screenshot looks like.

If so, I wouldn't do that, unless you really did this dive using a 100% oxygen diluent - if what you want to do is to link the tank pressure sensor data to different gasmixes, do this by using the 'Sensors' column in the cylinder list:

Imported on:

Subsurface-mobile, Android: 6.0.5112

I'd probably update this to the latest version from https://subsurface-divelog.org/latest-release/ - there was a bug in the importer for CCR sensor values that resulted in the sensor values being stored incorrectly (https://github.com/subsurface/subsurface/pull/4419).

Ngā mihi

  Michael Keller

[jimbodude]

No, I definitely did not do this 120ft dive with Oxygen as my diluent, haha!  I just grabbed the latest Mac version of Subsurface, 5341 vs 5299.  I still don't see a "Sensors" column in my Equipment/cylinders section, and I don't see a setting that would enable that, but that definitely sounds like what I need.  How do I get that column?

[Michael Keller]

Hi Jim.

On 16/02/25 19:51, jimbodude wrote:
> No, I definitely did not do this 120ft dive with Oxygen as my diluent,
> haha!  I just grabbed the latest Mac version of Subsurface, 5341 vs
> 5299.  I still don't see a "Sensors" column in my Equipment/cylinders
> section, and I don't see a setting that would enable that, but that
> definitely sounds like what I need. How do I get that column?

Yes, unfortunately this is not very discoverable - right-clicking into
the headers row of the cylinders table should give you a drop-down that
lets you select the columns to display.


You should update the Android version of Subsurface as well - since we
are not distributing this through the Play store any more this has to be
done manually by downloading the .apk and then sideloading it.


Cheers

  Michael Keller

[James Cialdea]

Ah ha - perfect, thanks!  This makes much more sense, and things are now working correctly.  Thank you!

JCialdea

--
You received this message because you are subscribed to a topic in the Google Groups "Subsurface Divelog" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/subsurface-divelog/K9u2SBEyUlw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to subsurface-dive...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/subsurface-divelog/d93bc8da-833f-4a77-b91d-0f8fe55ffef0%40042.ch.

[jimbodude]

I've confused it again, but not the deco calcs this time.  Here's an example of a dive where I've reassigned the sensors to the diluent gas (the undeletable one) and an O2 bottle that I added.  There is a bunch of extra data now printed on the graph and the start/end pressures stayed on the original cylinders they were imported with.

How I got here:

• Imported the dive, resulting in cylinders 1 (BO Deco, 50%), 2 (BO Air), 3 (BO deep mix), and 4 (Diluent).  Cylinder 1 was assigned to sensor "0" ("D1" on my computer, which is the diluent sensor) and Cylinder 2 was assigned to sensor "1" ("O2" on my computer, the oxygen sensor).
• Since all the gasses imported were correct, I added cylinder 5, 100% O2.
• I assigned Cylinders 3 and 4 as 3L bottles
• I assigned the correct sensor to each of those

Any idea what is causing the overprinting on the graph?  Why don't the start/end pressures move with the sensor reassignment?  It's important that they get moved - if I were to assign the cylinder sizes to 1 and 2, I'd get gas consumption calculations there where no gas was actually consumed.

[Michael Keller]

Hi Jim.

On 16/02/25 20:21, jimbodude wrote:
> I've confused it again, but not the deco calcs this time.  Here's an
> example of a dive where I've reassigned the sensors to the diluent gas
> (the undeletable one) and an O2 bottle that I added. There is a bunch
> of extra data now printed on the graph and the start/end pressures
> stayed on the original cylinders they were imported with.

Have you tried saving the results, then closing the dive log and opening
it again?
I suspect that not everything that should be recalculated when
re-assigning tank sensors is at the moment.


I am actually working on an improvement of the functionality to
re-assign tank sensors, but I am short on example dives with multiple
tank sensors that I could use to test this.

So if you still have your logs in the form that you imported them, or if
you could re-import them and then share with me that would help me with
this work.


Ngā mihi

  Michael Keller

[jimbodude]

Save, close, and reopen looks like data loss to me.  No sensors assigned, most start/end pressures lost, and no option in the "sensors" dropdowns.

I'm happy to share this dive and any others - nothing secret here.  Let me know the most interesting format - I'll re-pull them and ship them to you when I get home.  (Which will be a few days - we're currently in Micronesia headed to US East Coast...)

[Michael Keller]

Hi Jim.

On 17/02/25 00:05, jimbodude wrote:
> I'm happy to share this dive and any others - nothing secret here. 
> Let me know the most interesting format - I'll re-pull them and ship
> them to you when I get home.  (Which will be a few days - we're
> currently in Micronesia headed to US East Coast...)

Just a .xml or .ssrf saved by Subsurface, after importing the dives with
multiple sensors into a new empty log file will give me all that I need.
Thanks.


Cheers

  Michael Keller

[jimbodude]

Perfect timing, I just got back online after the trip home.

Posting 3 files here (seems like they're too big for Groups to handle...):

- as_imported: just a moment after the import from the computer.

- notated: some trip grouping, gas types in the location field, and notes on the dives and trips to make things easier to navigate, sort, and understand.

- bottles_added: Did the workflow described in the 2-sensor trip notes for the first two 2-sensor dives.  I added notes about what looks wrong to each of those.  Added notes to the 2-sensor trip about the workflow and its pain points.

Let me know if I can help with anything - thanks for your attention on this.

[jimbodude]

Digging this back up.  I think the first issue to solve is that all CCR dives need an O2 cylinder.  I'm trying that here - https://github.com/subsurface/subsurface/pull/4758 - let me know what you think.  My computers aren't with me - I'll test it when I get them back.