Discrepancy between my computers during the dive

[pisauron]

It's page 9 on this thread. Let's do a TLDR.

Don't use computers from different manufacturers and expect the same results.

I got a different TLDR:

Computers with completely identical settings show massive differences in safety information and no one knows why.

 

[Jai Bar]

It's page 9 on this thread. Let's do a TLDR.

Don't use computers from different manufacturers and expect the same results.

Nope. Two computers using the same mathematical model with exactly the same parameters should yield similar results.

Lots of people think that is some complicated model involving lots of complex equations with multiple possible solutions tp the same inputs. But it is not. It is a simple model being used for more than 200 years, there is no voodoo involved.
 
You can expect some variance because of differences in depth sensors, rounding, noise filtering etc. I can understand a difference of 1-2 minutes. But 12 minute difference in NDL is very concerning,

 

[harvalen]

That's the plan. But which one though?

Shearwater or OSTC?

I found a similar issue with OSTC: Ceiling on 2nd dive - forum.heinrichsweikamp.com Just like here, with comments about computers being different, it was hard to get anyone to even understand that there was an issue.

Since it doesn't really affect my diving I didn't really put more effort into investigating after that. Looking back at those dives I do get Subsurface and OSTC to line up if I shorten the surface interval and works the same on a few other dives that I have. So now my guess is that the OSTC is counting time a bit too slow when it's in sleep mode on the surface. I'll see if I can do some more investigation next time I go diving. Maybe keeping the OSTC awake during the surface interval gives a different result.

To have a third term of comparison, we set up Subsurface with BUL settings and gradients 70/85. The result from Subsurface would have been even more conservative compared to both dive computers. (in the images, the deco suggested by Subsurface is shown in green, the one by RATIO in red).

For anyone interested i am attaching the logs here:

Garmin VS Ratio – Google Drive

drive.google.com

This is the ceiling I get when loading your dives in subsurface:

The only way I can get the subsurface ceiling to match what the Ratio support got is to set the surface interval to 3 min.

With the full 93 min surface interval the maximum Surface GF is 78% according to Subsurface, with a 3 min one it is 99%. It's hard to get the Subsurface and Ratio ceilings to match up. With 33 min surface interval they are similar length, but shifted a bit from each other, so it doesn't seem to just be a shortening of the surface interval. Maybe accumulation of shortened surface intervals for all the previous dives would do, but it starts getting messy.

(The reason for using 93, 33 and 3 min is just because the previous dive ended at 09:17 and it was nicer to use :20 :50 and so on for the minutes).

 

[pisauron]

I found a similar issue with OSTC: Ceiling on 2nd dive - forum.heinrichsweikamp.com Just like here, with comments about computers being different, it was hard to get anyone to even understand that there was an issue.

Since it doesn't really affect my diving I didn't really put more effort into investigating after that. Looking back at those dives I do get Subsurface and OSTC to line up if I shorten the surface interval and works the same on a few other dives that I have. So now my guess is that the OSTC is counting time a bit too slow when it's in sleep mode on the surface. I'll see if I can do some more investigation next time I go diving. Maybe keeping the OSTC awake during the surface interval gives a different result.





This is the ceiling I get when loading your dives in subsurface: View attachment 840699

The only way I can get the subsurface ceiling to match what the Ratio support got is to set the surface interval to 3 min.
View attachment 840701

With the full 93 min surface interval the maximum Surface GF is 78% according to Subsurface, with a 3 min one it is 99%. It's hard to get the Subsurface and Ratio ceilings to match up. With 33 min surface interval they are similar length, but shifted a bit from each other, so it doesn't seem to just be a shortening of the surface interval. Maybe accumulation of shortened surface intervals for all the previous dives would do, but it starts getting messy.

(The reason for using 93, 33 and 3 min is just because the previous dive ended at 09:17 and it was nicer to use :20 :50 and so on for the minutes).

So maybe the Ratio somehow messed up the surface intervals?

 

[scubadada]

Hi @harvalen

Seems odd, but very interesting. An error in the SI would affect both the NDL and the CNS O2, a unifying explanation. From the logs, I would imagine @pisauron could check the SI recorded for this dive by both computers, not that it would ensure that time was credited. I don't think I have ever heard of a dive computer error in not correctly crediting the SI, there's always a first time.

From the graphs of the implicated dive:
Ratio: 1 min NDL and 26% CNS O2
Garmin: 13 min NDL and 16% CNS O2

The difference in CNS O2 represents a significant portion of one 90 min half-life of O2 elimination

 

[Jai Bar]

So maybe the Ratio somehow messed up the surface intervals?

Uwatec had this bug in the past: when diving Nitrox it used the same percentage for the surface intervals (instead of Air):

Did Uwatec conceal a known computer flaw?

This post was moved from another computer thread, where the issue raised derailed that conversation. This interesting data perhaps warrants its own discussion. Twenty years - a long time. History creates memories that never entirely fade. Corporate actions have consequences, today and...

scubaboard.com

 

[pisauron]

Hi @harvalen

Seems odd, but very interesting. An error in the SI would affect both the NDL and the CNS O2, a unifying explanation. From the logs, I would imagine @pisauron could check the SI recorded for this dive by both computers, not that it would ensure that time was credited. I don't think I have ever heard of a dive computer error in not correctly crediting the SI, there's always a first time.

From the graphs of the implicated dive:
Ratio: 1 min NDL and 26% CNS O2
Garmin: 13 min NDL and 16% CNS O2

The difference in CNS O2 represents a significant portion of one 90 min half-life of O2 elimination

Thanks. How do i check the SI recorded?

Also i think Ratio support addressed that somehow

 

[scubadada]

Thanks. How do i check the SI recorded?

Hi @pisauron

I dive a Shearwater computer and an Oceanic computer. Both computers show the SI for the dive in the on-board log and in the downloaded log. I would imagine your Garmin and Ratio computer do too.

This is from my Shearwater computer download to the Shearwater cloud. It might also be interesting to compare the CNS O2 exposures for your two computers to see if the discrepancy persists. My Shearwater gives the CNS start and CNS end in both the on-board log and in the download. My Oceanic computer does not have this data. We already discussed that it would be interesting to compare end surface GFs if they were available. I don't use Subsurface, somebody else can help you with that.

 

[dmaziuk]

Thanks. How do i check the SI recorded?

Also i think Ratio support addressed that somehow

You can open subsurface XML file with an text editor (pref. a "smarter" one like notepad++). Each dive has its own block that starts like

<dive number='2' date='2014-04-04' time='14:59:00' duration='43:00 min'>

Start date and time will be there. SI is not recorded per se, add duration to start time to get the end time, then subtract from the next dive's start time.

 

[pisauron]

Hi @pisauron

I dive a Shearwater computer and an Oceanic computer. Both computers show the SI for the dive in the on-board log and in the downloaded log. I would imagine your Garmin and Ratio computer do too.

This is from my Shearwater computer download to the Shearwater cloud. It might also be interesting to compare the CNS O2 exposures for your two computers to see if the discrepancy persists. My Shearwater gives the CNS start and CNS end in both the on-board log and in the download. My Oceanic computer does not have this data. We already discussed that it would be interesting to compare end surface GFs if they were available. I don't use Subsurface, somebody else can help you with that.

View attachment 840757

If you look at the logs you will see the CNS changes. Dives start with CNS lower than the previous dive-end CNS.