Matching Perdix and MultiDeco

[RonMurray]

I have been using MultiDeco using Buhlman ZHLC (GF 30/70) for some time on my phone and like it. I recently purchased a Shearwater Perdix which seems to also uses the ZHLC algorithm - but try as I might, I can't get Multideco and the Searwater the agree on dive plans. Not big difference mind you - a 135' dive on just air yields about a 2 minute total dive time difference. (From memory so I may be inexact on these datails).

Not a big deal, but it bugs me. I have checked and rechecked to make sure all of the settings that i can find match (gradient factors, ascent rates, salinity, even gas consumption).

Anyone have any ideas?

(My apologies if this an improper forum)

 

[rongoodman]

It doesn't surprise me that two different implementations of the same algorithm might be a little off. How does the Perdix compare to ZHL B on MultiDeco?

 

[scubadada]

Hi @RonMurray

Is this a no deco or a deco dive? For NDLs, I believe Shearwater, and my Nitek Q, calculate NDL with the longest time for which you can still surface at, or under, the GF hi, using the dictated ascent rate. GF lo does not kick in until you cannot meet that requirement and go into deco. MuliDeco does not calculate NDL, per se, and the GF lo appears to always be in play.

Once you go into deco and the GF lo kicks in to dictate the initial stop, I would generally expect the computer and MultiDeco to calculate the same run time.

 

[victorzamora]

My understanding is that ZHL-C is the dive computer and real-time variant of the algorithm. ZHL-B is the variant meant for deco planners. My recollection is that it's less processor-intensive but also less flexible. The differences in planning are as OP mentioned: a minute or two here or there. Close enough that, for me, there's no point in worrying as I'll never actually do a bottom time exactly as planned.

 

[RonMurray]

Right on Victor, I'm not concerned about the minute or two variance - the imponderable vagaries of human physiology alone obviate the need to anguish over them. But my desire for order in the universe makes me wish they matched
I assumed the same algorithm, given the same inputs, would generate he same result.

Live and learn...

 

[victorzamora]

Right on Victor, I'm not concerned about the minute or two variance - the imponderable vagaries of human physiology alone obviate the need to anguish over them. But my desire for order in the universe makes me wish they matched
I assumed the same algorithm, given the same inputs, would generate he same result.

Live and learn...

I definitely understand. My wife and I joke that my "crazy" doesn't like things like that slide. However, my crazy and I have sat down and had long discussions regarding tiny differences. Things like that still bug me, but now I can sleep at night.

As an engineer, it bugs me as well. If you do the math, it should give you the same answer. However, in engineering you also learn the phrase "close enough" early on and use it frequently.

 

[joe10540]

They may both be using the same algorithm but one might have a small amount of a built in safety factor perhaps only an extra 1-2% built into the run time. That wouldn't to show up until you got either very deep or long run times. That might explain the minor differences you are seeing

 

[KenGordon]

My understanding is that ZHL-C is the dive computer and real-time variant of the algorithm. ZHL-B is the variant meant for deco planners. My recollection is that it's less processor-intensive but also less flexible. The differences in planning are as OP mentioned: a minute or two here or there. Close enough that, for me, there's no point in worrying as I'll never actually do a bottom time exactly as planned.

C and B differ in the coefficients only. The CPU requirements are identical. To switch between one and the other the implementation only has to swap a pointer. C's coefficients make it slightly more conservative in the middle compartments. It is said this is to make up for the lack of rounding of the profile.

OP. Try selecting C in the planner.

One likely difference is that computers use the lowest ceiling rather than the first stop to start interpolating from GF low. See what happens if you use something like 70/70 for the plans.

It is a shame Shearwater do not provide planning in their desktop software.

 

[Shearwater]

The difference between the desktop plans and a real-time plan is in the way they handle stops of less that a minute. In desktop algorithms, if there is a 10 second stop, the algorithm will show it as a 1 minute stop in the plan. The real-time computer is calculating the profile in real-time, so the 10 second stop clears in 10 seconds.

Bruce

 

[Shearwater]

It is unlikely that Shearwater will offer a desktop planner. There are several good planners that are free or inexpensive so there is very little opportunity to add much value.

Bruce