computer redundancy

[gcbryan]

Does GUE training come with the obligation to post this sentence, or is it optional? If optional, has anybody actually opted not to? Just curious.

I can't answer that question but I do know that "you don't know what you don't know" and I do know that "air is for tires" and that "computers can rot your brain". I've further learned that "if the glove doesn't fit you must acquit".

And yes, now that I think about it I do recall that "my buddies brain is my backup". Not having had the training I'm confused about the posting requirements as well.

 

[DaleC]

thanks, daleC.

i think the main differences in algos can be seen in different scenarios in decompression diving (ndl diving might give a minute more or less but not have a significant impact IMHO). now, assuming you run buhlmann and vpm, i strongly believe that you find dive plans where one time buhlmann and another time vpm comes out ahead. if you have two computers running different algorithms, and you always go by the more conservative, wouldn't you add some safety margin?

just a thought.

It's a good question and I am more interested in generating discussion than just saying I'm right or wrong. I'm interested in the justifications.

To me it depends on why you are seeking redundancy. Some posters seem to run two different algorhythms and rely on the more conservative one but that leads me to ask, why run an algorhythm you don't trust? Pick a decompression model to trust and go with that.

I see redundancy as being able to detect whether my main computer is malfunctioning and giving me a false reading, not giving me two programs to choose from. Running the same model allows for fault detection; running different models does not.

Spot the failure:

a.) Two similar models: 1:1 1:1 1:1 1:1 1:2 1:1 ...

b.) Two different models: 1:1 2:1 1:2 2:2 1:2 2:1 ...

In a.) you cannot tell which value is false but you know there is a failure. In b.) you cannot tell if, when or where a failure occurs.

Here's a similar question though. When setting up doubles does one use two similar first stage regulators or two different models from different manufacturers with different performance outputs? Why/why not?

Someone mentioned using two computers to track varying functions which makes sense to me but I think you sacrifice one thing for the other. It all depends on what's more important.

Of course I am describing the perfect scenario. Knowing ahead of time which decompression model one prefers and being able to buy two computers that serve all long term applications. In reality most divers (like me) initially buy a base computer that the shop recommends and then down the road buy a second because the first is too limiting. Then you wear both.

 

[vladimir]

At least for me, the redundancy is not targeted at catching computer failures or algorithm flaws. My experience with failures is that they make themselves evident, though not necessarily intra-dive. For example, when my Air-Z died it was displaying a depth of 4 feet on the surface. I trust the algorithms on both my computers, though one is more conservative than the other. If the Uwatec dies I am happy to rely on the Suunto, and vice-versa. As mentioned, the reason I have two different brands is that together they offer an aggregate of features not offered on either one individually (logging software, size of display, displayed data, etc). The reason that I have a back-up at all is to rescue the dives that I have planned after one of the computers dies. On a ten-day liveaboard, if one dies on day two I will have all my nitrogen data on the other, and I can continue diving uninterrupted.

 

[DaleC]

the reason I have two different brands is that together they offer an aggregate of features not offered on either one individually.

That makes sense to me. It all depends on what one is seeking in redundancy.

 

[super7]

My contention is that if one wants to get the most benefit from redundant computers one should consider using the same algorhythm. If they are giving signifigantly different values then at least you know one is malfunctioning and can adjust your dive accordingly. If you use two computers that spit out different values because they are following different models how will you ever know if either one is malfunctioning. Different values would be expected.

If you dive one computer in gauge mode then it is just a very expensive Bottom timer/depth gauge.

I guess both computers could be affected at the same time, giving identical false readings but that is sort of like planning for multiple failures. It would be much more likely that they would malfunction seperately rather than simultaniously.

It's not a critism of people diving different computers because they have them or setting one into gauge mode, just a thought to stimulate discussion for those setting out to make a future purchase.
Should I buy two computers with the same algorhythm?
Different algorhythms?
One computer, one Bottom timer/depth gauge?

Well, I reasoned that if the computer died, I can still fall back on the gauge and start looking up the tables. In the unlikely event that the one in gauge mode dies, I still have the computer running calculations for me. Maybe I'm not exactly thinking redundency, but more towards a contingency plan.

 

[stargost]

Here's a similar question though. When setting up doubles does one use two similar first stage regulators or two different models from different manufacturers with different performance outputs? Why/why not?

.

I've been about to ask this question on separate post. I dont think people are using different reg, while indeed, this could be seen as further redundancy (different brand, different first stage like piston / diaphragm). Maybe it is not good looking.

I remember designing a electronic system for a military application, and the specification required that the 2 redundant systems were very different:
- 2 different processors (DSP) from different brand/manufacturers
- the (simple) software running on them was to be written by different programmers. Software was doing the exact same thing (not 2 different algo for detection)

The test protocol was of course the same, but the idea was indeed you dont want the failure to be 'model' dependent.

However they were a few problems as beyond the digital, there were a couple of analog blocks which also were supposed to be different and ending having too wide variations over extreme conditions, so eventually, the analog redundancy was kept the same (components/supplier) - even military can be flexible ;-)

I guess the point here is whatever redundancy approach you choose - same computer or not, same algo or not, there is going to be a few compromises !

 

[drbill]

I dive with a Tec 2G and a Nitek Duo because they are the computers I have (well, if you don't count my Orca Skinny Dipper). I log the Nitek Duo info on my website and download the Tec 2G data to my computer.

 

[Rhone Man]

Here's a similar question though. When setting up doubles does one use two similar first stage regulators or two different models from different manufacturers with different performance outputs? Why/why not?

I've been about to ask this question on separate post. I dont think people are using different reg, while indeed, this could be seen as further redundancy (different brand, different first stage like piston / diaphragm).

Good idea. I don't think it is entirely the same question (different considerations), but it sounds like an interesting discussion topic nonetheless.

 

[DA Aquamaster]

I have a Shearwater Predator and an Uwatec Luna with the PMG upgrade.

Both bring something different to the dive. The Shearwater is a very simple, easy to program in the field and easy to read during the dive computer that offers a great deal of flexibility in terms of profile due to the adjustable gradient factors. It is the "primary" computer.

I have absolutely no use for AI an I saw no need to buy a transmitter for it, but the Luna brings several other features to the dive such as a very good digital compass and some book mark/log/download software features that I find are very useful in terms of reviewing and debriefing a dive. Like the Tec2 G it has more of a bend and mend model with a tendency toward shallower stops than the Shearwater using default 30/85 GF's. But they seem to play well together in terms of total run time.

The display on the Shearwater Predator is great on the arm holding the light where it sits in the dark, but it gets a bit spastic when the display tries to adjust for changing light levels making it less happy on the non light holding arm where it is intermittently lit by the dive light. In contrast, the Luna has a great display that is easy to read with the light and is well suited to being carried on the non light holding arm.

Two different deco models does not have to be a major issue. As long as the shallower stop computer does not mind deeper stops and ends the total deco within a few minutes of the deeper stop computer, there is no real downside. The normal offset and differences in stops and depths also serves as a sanity check to ensure both computers are functioning normally. If they are not disgreeing by the usual amount, experience with similar dives in the past and/or with the pre-dive plan and expected deco should provide more than enough clues regarding which computer is faulty.

That last line is however going to be true even if you use 2 identical computers. You should be doing enough pre-dive planning to have a good idea of what to expect for the deco obligation and consequently enough to spot faulty deco instructions in a computer. If you are not doing that, you are playing fast and loose with the gas planning and that will eventually bite you in the ass real hard.

-----

Different brand/model regs does not in my opinion offer any advantage in redundancy. Reg failures are rare unless your personal equipment maintence is really poor and using regs of different brands complicates the annual service logistics even more than using regs of different models. In that regard, it is arguably an "equipment solution for a skills deficiency" example.

I use piston first stages in applications where they make sense (stage/deco regs) and I use diapragm regs where they make sense (cold/dirty water). I use balanced second stages in all applications as they are in the big picture no more failure prone and not much more complicated than unbalanced regs, but perform a lot better.

 

[DaleC]

I understand the real world practicality of buying different computers for different reasons and winding up with different algorhythms that one is able to dive with regardless but, I still struggle with seeing the upside to intentionally choosing different decompression models.

I also don't claim that either myself or others are right or wrong, I'm more interested in the thinking behind peoples choices.

I brought up the first stage reg question because I was wondering how many people who would choose different models because they fear inborn structural failures also consider that when using identical regulators.

One can also choose computers with different functions that are also based on the same decompression models. The Suunto Gekko and Vyper are two examples I am familiar with.

but let me set out a scenario:

Computers A and B use the same decompression model and (combined) have all the features one is looking for...

Computers C and D use different decompression models and (combined) have all the features one is looking for.

All things being equal, which ones would you choose?