When is ZH-L16C GF not ZH-L16C GF? When it is a G2 Tek computer

[Belzelbub]

If you want to tilt at ScubaPro windmills, have at it.

Those must be some expensive windmills.

 

[Sevenrider860]

** Corrected Original G2 Runtime value. **

Using MultiDeco as the offline dive planner and the dive planning modes on the Shearwater, Garmin, and G2 Tek computers.

Plans were created using ZH-L16C with gradient factors of 50/75.
90 Feet / 60 Minutes / 21%

MultiDeco
Total Runtime: 129 Minutes
First stop 40 feet for 20 seconds

Shearwater
Total Runtime: 132 minutes
First stop 40 feet for 1 minute

Garmin
Total Runtime: 131 minutes
First stop 40 feet for 1 minute

G2 Tek
Total Runtime: 153 minutes
First stop 40 feet for 5 minutes


The G2 Tek Dive Planner does not allow you to see anything but the first stop and the Total Ascent Time in their dive planner. Total Ascent Time shown on the Scubapro is 93 minutes, so that would make the total runtime 153 minutes. MultiDeco and the non-Scubapro computers all came up with a 1 minute (or less) stop at 40 feet. The G2 Tek has a 5 minute stop.

MultiDeco, Shearwater, and Garmin all generate dive planes that vary no more than 3 minutes between them and they all generate a 1 minute (or less) stop at 40 feet. The G2 Tek dive planner is not close to these other "industry standard" ZH-L16C implementations.

 

[lexvil]

Using MultiDeco as the offline dive planner and the dive planning modes on the Shearwater, Garmin, and G2 Tek computers.

Plans were created using ZH-L16C with gradient factors of 50/75.
90 Feet / 60 Minutes / 21%

MultiDeco
Total Runtime: 129 Minutes
First stop 40 feet for 20 seconds

Shearwater
Total Runtime: 132 minutes
First stop 40 feet for 1 minute

Garmin
Total Runtime: 131 minutes
First stop 40 feet for 1 minute

G2 Tek
Total Runtime: 153 minutes
First stop 40 feet for 5 minutes


Total Ascent Time shown on the Scubapro is 93 minutes, so that would make the total runtime 153 minutes. MultiDeco and the non-Scubapro computers all came up with a 1 minute (or less) stop at 40 feet. The G2 Tek has a 5 minute stop.

As I mentioned in my original post...MultiDeco, Shearwater, and Garmin all generate dive planes that vary no more than 3 minutes between them and they all generate a 1 minute (or less) stop at 40 feet. The G2 Tek dive planner is not close to these other "industry standard" ZH-L16C implementations.

Can you use the pressure pot to put the computer through real simulated conditions to see if it tracks with the planner? I would hope that the planner software is screwed up not the actual algorithm.

 

[Sevenrider860]

Can you use the pressure pot to put the computer through real simulated conditions to see if it tracks with the planner? I would hope that the planner software is screwed up not the actual algorithm.

I updated the post. I screwed up the run time. The G2 now shows itself to be more conservative using identical values.

I do plan on diving these again in the next month.

 

[Sevenrider860]

I pulled out the HUD. The planner on the HUD is actually a simulator.

The Scubapro HUD results in:

Total Runtime: 138 minutes
First stop 40 feet for 2 minutes

This confirms my impression from some previous dives that the HUD tracks more closely to the Shearwater that the G2 Tek does.

 

[Sevenrider860]

Here are some NDL times from the G2 (ZH-L16 ADT MB PMG) at various MB levels compared to the NDL times of computers that implement ZH-L16C + GF. The G2 Tek wont go below a GF High of 50, so that is the bottom limit I used for the computers.

Diving the Shearwater on Medium Conservatism (40/85) gives you 44 minutes of NDL. Which is equivalent to 45/90 on the G2 Tek and a little less than MB L0 on the G2. From my testing, the G2 Tek NDL times are not affected by any changes to GF Low.

The Scubapro HUD does track closer to the Shearwater than the G2 Tek which seems to support the rumor I have heard that the ZH-L16C + GF algorithm on the HUD came from the SeaBear software developers and not from the current G2 Tek/Luna 2.0 software developers. SeaBear was purchased by Johnson Outdoors (Scubapro) in 2015. Dr. Arne Sieber was the founder of SeaBear and he stayed with Scubapro as the R&D Manager for Scubapro Dive Computers until 2017. I think the Scubapro HUD was released in 2018, so it makes sense that Dr. Sieber would have had a lot of input into the ZH-L16C + GF algorithm that was used in this computer. I am pretty sure the G2 Tek was released in 2022, so it is doubtful that Dr. Sieber had much, if any, input into the software development of the ZH-L16C + GF algorithm used in the G2 Tek.

It is my opinion that had Scubapro implemented the same ZH-L16C + GF algorithm in the G2 Tek that they have in the HUD, I don't think there would be any issues over the algorithm. There would still be complaints about SOS mode if you ascend above 3 feet for 3 minutes after missing a required GF 100/100 stop and I am sure some other complaints, but the algorithm itself would not be questioned.

 

[stuartv]

** Corrected Original G2 Runtime value. **

Using MultiDeco as the offline dive planner and the dive planning modes on the Shearwater, Garmin, and G2 Tek computers.

Plans were created using ZH-L16C with gradient factors of 50/75.
90 Feet / 60 Minutes / 21%

MultiDeco
Total Runtime: 129 Minutes
First stop 40 feet for 20 seconds

Shearwater
Total Runtime: 132 minutes
First stop 40 feet for 1 minute

Garmin
Total Runtime: 131 minutes
First stop 40 feet for 1 minute

G2 Tek
Total Runtime: 153 minutes
First stop 40 feet for 5 minutes


The G2 Tek Dive Planner does not allow you to see anything but the first stop and the Total Ascent Time in their dive planner. Total Ascent Time shown on the Scubapro is 93 minutes, so that would make the total runtime 153 minutes. MultiDeco and the non-Scubapro computers all came up with a 1 minute (or less) stop at 40 feet. The G2 Tek has a 5 minute stop.

MultiDeco, Shearwater, and Garmin all generate dive planes that vary no more than 3 minutes between them and they all generate a 1 minute (or less) stop at 40 feet. The G2 Tek dive planner is not close to these other "industry standard" ZH-L16C implementations.

Thank you for doing that.

I ran the same dive in the Subsurface dive planner. I got 131 minutes runtime.


Subsurface (6.0.5054-CICD-release) dive plan created on 2/29/24
Runtime: 131min

	depth​	duration​	runtime​	gas​
➘	90ft	3min	3min	air
➙	90ft	57min	60min
➚	40ft	2min	62min
-	40ft	2min	64min
➚	30ft	0min	64min
-	30ft	10min	74min
➚	20ft	0min	74min
-	20ft	21min	95min
➚	10ft	0min	95min
-	10ft	36min	131min
➚	0ft	0min	131min

CNS: 13%
OTU: 37

Deco model: Bühlmann ZHL-16C with GFLow = 50% and GFHigh = 75%

But, when I first ran it, my total runtime was 166 minutes. I had my last stop depth set to 20'. Changing it to 10' changed the plan to the above.

Is it possible that your G2 Tek is set to assume a last stop depth of something other than 10'?

Alternatively, I wonder if the discrepancy is coming from the G2 Tek assuming slower ascent rates, per its Ideal Ascent Rate chart.

 

[scubadada]

...From my testing, the G2 Tek NDL times are not affected by any changes to GF Low...

Thanks for the interesting table. From what has been posted in this thread, it does appear likely that the G2 Tek runs a revised, more conservative version of Buhlmann ZH-L16C with GF. This strategy has been in place for a long time. Pelagic Pressure Systems/Oceanic PZ+ is said to be "Buhlmann ZH-L16C based", straightforward and honest. I look forward to hearing what @BoltSnap learns from discussing this topic with his contacts at Scubapro.

For Buhlmann ZH-L16C with GF, the NDLs are entirely dictated by the GF high. The GF low does not kick in until decompression, when it dictates the depth of the first stop.

 

[Sevenrider860]

For Buhlmann ZH-L16C with GF, the NDLs are entirely dictated by the GF high. The GF low does not kick in until decompression, when it dictates the depth of the first stop.

That is my expectation and what I would consider correct operation. But I have seen posts in the past that there are some computer implementations that use GF Low values in the NDL calculation.

 

[inquis]

The GF low does not kick in until decompression, when it dictates the depth of the first stop.

Perhaps "should not" is more accurate. I can see a poor implementation screwing this up and appreciate the verification by @Sevenrider860.

Edit: need more coffee... so... slowww.