AI computer or gauges for safety?

[fireflock]

The numbers you're looking for just don't exist. No one knows what the failure rate is for a AI wisdom expressed in any kind of useful term like failures per dive hour or something. I had one fail, but that doesn't tell anything useful. There are no supporting documents, or systematic studies.

At what point does the failure rate cease to matter? My guess is that AI computers, wireless computers, and analog SPG's all work well enough that the average diver won't see a problem more than once every few years at most.

Rich

[mitsuguy]

If you are dead set on this as your final conclusion, then none of us are going to be able to prove it otherwise...

Reasonably though, electronics fail, batteries fail even though they may be maintained and charged. The more electronics there are, the more likely of failure.

With mechanical gauges, there are still similar type failure points, but being worried about gauge creep and stuff is a non issue... I used to have a boost gauge in a car that was 5 psi off... at all times it was 5 psi off... I knew that and worked with it... even if your SPG is off by, say 10-15 percent, maybe even 25%, whats the worst that can happen? We plan our dives in thirds, a third to get there, a third to get back, a third just in case... so, if your gauge reads 25% high, your dive is done 25% faster...

Originally Posted by rawalker

As I stated regular maintenance was to be considered.
Accidental damage can happen to any piece of equipment and is not a consideration.
This is a question for state of the art regularly available products for the diving public.
If you feel Oceanic is the best AI wireless then we should be judging by that as state of the art. Or if you feel the Sherwood Wisdom hosed AI computer is a better representative of state of the art I'd like to see supporting documents on it's failure rate.
Failures are not to be considered battery gone dead as this would be a maintenance issue unless caused by a battery compartment flood.
This is to be a comparison of properly maintained devices without failures caused by user error!
Loosing sync for individual polling segments is not a failure and does not decrease dive safety!
Also when I say redundant wireless AI I mean 2 transmitters on the same bottle used for separate computers. (really redundant) I know this is not the way these computers are used today but they could be and it would increase the safety margin.
I do accept that sync can be lost but also that mechanical gauges (SPG) are more prone to inaccurate readings creating other safety concerns.

[ianr33]

Lets turn the question around. What do divers that are doing the most demanding dives in the world use? e.g WKPP with bottom times of around 5 HOURS at almost 300 feet,or maybe the wreck of the Britannic at around 400 feet.

I'm pretty certain they are using Brass spg's rather than redundant wireless computers.

[cummings66]

The OP would like to hear that computers are the king of the hill and nothing can knock them down, to that end you'll see he changed the question's parameters and now rules out loss of sync as a failure. By it's very warning it's a failure, if it wasn't it wouldn't tell you it could not do something.

It is critical to know how much air you have, without that data you can and for a new diver will run out of air, that means you have now increased the odds of you dying by a large margin, whether it be by drowning or embolism, OOA incidents account for a large percentage of diver deaths.

I maintain loss of link is a critical failure and not one you can endure if that computer is the only way you have to know the pressure remaining. You don't bet your life on a device that fails, and I had a loss of link last night. Once and for only a second or so. That's 3 losses in 2 years. Not many but one of them was for the entire dive and I used my B&G SPG so I could do the dive rather than abort.

What proof do you want that loss of pressure readings can cause serious dive consequences. Do you want DANS statistics on divers who die? Go to DAN and read for yourself the statistics. Do you want to see my logs where the computer lost sync? Will enough of us that say it happens convince you it's folly to believe the computer is infallible. You say don't count batteries, how about batteries that die shortly after being used? Look here, you'll see a post from a buddy of mine where his NEW VT3 had a battery croak in less than I think 10 dives with it. How about where the computer flooded after being properly serviced by a dive shop? Look on the board and you'll see those stories too. Several of them in fact. I had the battery in my transmitter fail without warning either, and the thing is supposed to warn you, except it didn't. It just quit.

Turn it around, how many people here have posted my B&G SPG failed and told me a bad number? Not many, I've never had a buddy have one fail that didn't suffer abuse, and it showed the signs of failure, fogging on the interior of the glass housing. The mechanics are more reliable. Make your own study, read the postings. Compare them, how many people have said my computer broke and I need a new one, I've lost lock, it flooded, battery died? More than you can count. How many have said my spg's flooded on me? Not one.

Use your eyes, you don't need a study, you have all the proof you need here in the forums. Branch out, look on TDS or scubaboard. Look for spg failures, now look for computer failures. You will discover computers die way more often.

Good luck finding data from a manufacturer on failures in a public forum, that's just not going to happen. I doubt you could call them up and have them tell you either. I repair electronics for a living and I can tell you without a doubt that new products fail. I can also tell you the manufacturers don't tell me the statistics, only the failure modes and common methods of repair.

Forgive the tone of my posting, I happen to believe that in diving we should stack the deck to make the odds of surviving a failure greater and not less. There is no proof of what you want, but there is proof divers die when they run out of air. With an AI computer that is what we're talking about. Running out of air if that fails which they do. You have ruled that out too fast.

How do you intend to know how much air you have if the only means you can read it with lost sync?

[cummings66]

Actually, several years back Zenith told me in a service school they had a 95% failure rate of the products coming from Mexico, so I can't say they won't tell you, but it's rare for failure statistics to make it out of headquarters unless it gets to be HUGE.

[rawalker]

If your saying that not knowing your pressure for 15-30 seconds is critical to your dive then I guess you have your SPG glued to your mask! This is the type or loss of sync that computers experience normally and it may only happen once or twice in any 15 minute dive period. This is not a failure!!! I do happen to use 2 computers (an Oceanic Atom and a VT3) and haven't experience more than 1 missed poll on any 15 minute interval and then only on 1 of the computers.
As for electronic failure anyone that has worked as an electronic tech within a manufacturing environment can tell you that 90% of all failures of solid state electronic happen within the first 24 hours of use.
of the remaining 10% of failures over 9% happen within 72 hours that is why manufacturers put their devices through burn in periods. Rules out bad batches do to botched preceedures. It leaves a very small percentage of failures in the field and most of those failures are not the electronics but interface components such as switches. Life support devices need to show an even lower margin of failure to avoid litigation over product safety.
I myself have worked over 25 years in electronic technologies from DOD electronic at component level to low voltage controls, sound systems, large scale environmental controls (HVAC and refrigeration), fire alarm systems, preaction and suppression system.
Knowing what I do I find it hard to believe that any manufacturer would still be in business marketing dive computers that presented a significantly higher risk when used and maintained properly.

[mitsuguy]

Originally Posted by rawalker

As for electronic failure anyone that has worked as an electronic tech within a manufacturing environment can tell you that 90% of all failures of solid state electronic happen within the first 24 hours of use.
of the remaining 10% of failures over 9% happen within 72 hours that is why manufacturers put their devices through burn in periods. Rules out bad batches do to botched preceedures. It leaves a very small percentage of failures in the field and most of those failures are not the electronics but interface components such as switches. Life support devices need to show an even lower margin of failure to avoid litigation over product safety.
I myself have worked over 25 years in electronic technologies from DOD electronic at component level to low voltage controls, sound systems, large scale environmental controls (HVAC and refrigeration), fire alarm systems, preaction and suppression system.
Knowing what I do I find it hard to believe that any manufacturer would still be in business marketing dive computers that presented a significantly higher risk when used and maintained properly.

you are out of your mind if you believe those statistics...

why does RAM go bad in computers, sometimes randomly, after a few months of use, or a few years of use? Why does a processor go bad when it is kept at proper temperatures? Why does a network card go bad randomly... I worked in the computer repair industry for a good number of years, and have seen hundreds or even thousands of solid state devices fail for no apparent reason at all...

how about an crank position sensor in a car? Any reason they go bad? - they don't even having moving parts, and no electronics - just a magnetic sensor that has a reluctor wheel that "flies" close to it. Yet they still fail on vehicles, a year later, 2 years later, 10 years later... thats just one sensor... PCM's in cars go bad as well, and they are protected inside the car in most applications, but, they do fail, at a high rate, and usually 5-10 years after the car was first driven.

You basically just said that only 1% of electronics fail after the first three days of use. Anyone else here believe that?

[CompuDude]

Ask an engineer at a corporation a question about failure rates, and he'll quote the party line every time.

As a mechanic at the local repair shop, and he'll tell you what ACTUALLY dies in the field, vs. what the manufacturer would have your believe.

You can believe all you want about the bullet-proof reliability of electronics. And since there is no scuba police, you're welcome to dive however the heck you want.

But it seems pretty clear you're not going to get the answer you want here, because you have stacked the deck against answers with the only sorts of statistics anyone other than the actual manufacturer has access to... and I guarantee you they're not going to give those numbers up. If they were as good as you claim, they'd be publishing them on billboards.

Every bit of anecdotal evidence I have seen has said B&G is more reliable than wireless AI. Turns out, my personal experience... and first hand observations of many others... agrees with that evidence.

What you choose to do with this is entirely up to you.

[rawalker]

Originally Posted by mitsuguy

you are out of your mind if you believe those statistics...

why does RAM go bad in computers, sometimes randomly, after a few months of use, or a few years of use? Why does a processor go bad when it is kept at proper temperatures? Why does a network card go bad randomly... I worked in the computer repair industry for a good number of years, and have seen hundreds or even thousands of solid state devices fail for no apparent reason at all...

how about an crank position sensor in a car? Any reason they go bad? - they don't even having moving parts, and no electronics - just a magnetic sensor that has a reluctor wheel that "flies" close to it. Yet they still fail on vehicles, a year later, 2 years later, 10 years later... thats just one sensor... PCM's in cars go bad as well, and they are protected inside the car in most applications, but, they do fail, at a high rate, and usually 5-10 years after the car was first driven.

You basically just said that only 1% of electronics fail after the first three days of use. Anyone else here believe that?

To answer your question simply. Devices are designed to work within a given environment for a given length of time based on costs and safety requirements. Yes devices fail after time but that is usually considered in its design and manufacturing costs. For your example since not critical to safety the magnetic sensor is designed so that a percentage of them will fail after a year or 2 but many more survive many years. This is under the heating and cooling cycles under the hood. To electronics heat is the enemy. As for memory and CPU they also undergo heating and cooling cycles but there is often another problem (installation without proper ESD precautions (electro static discharge)) Would you like to see my 10 year old Pentium III computer that still works perfectly the only non original parts are the hard drives. (drives are mechanical they break)
It was built very carefully using proper methods to insure the components were not damaged during assembly. I know I built it.
I presently have another PC running 24/7 for the last 18 months. It had a cooling fan replaced about 6 months ago but otherwise is doing what it is supposed to.
A dive computer is a much simpler device overall it's design considers a underwater working environment and life support but it uses much less power and produces little measurable heat. It's really just a calculator with timer and pressure transducers. The transmitter uses a bit more power but because of it's limited needed range still uses little power and produces little heat. Good designs should be rewarded with trust and poor ones should be avoided.
Do you use a smoke detector in your home? They are designed to last 5 years and then should be replaced. During that time you should replace the battery yearly and check the detector monthly but you sleep every night and trust your life with it.
Granted the underwater environment is more dangerous to people but that is why we use redundant systems and most of us dive with a buddy. (a buddy is another redundancy)

[mitsuguy]

Originally Posted by rawalker

To answer your question simply. Devices are designed to work within a given environment for a given length of time based on costs and safety requirements. Yes devices fail after time but that is usually considered in its design and manufacturing costs. For your example since not critical to safety the magnetic sensor is designed so that a percentage of them will fail after a year or 2 but many more survive many years. This is under the heating and cooling cycles under the hood. To electronics heat is the enemy. As for memory and CPU they also undergo heating and cooling cycles but there is often another problem (installation without proper ESD precautions (electro static discharge)) Would you like to see my 10 year old Pentium III computer that still works perfectly the only non original parts are the hard drives. (drives are mechanical they break)
It was built very carefully using proper methods to insure the components were not damaged during assembly. I know I built it.
I presently have another PC running 24/7 for the last 18 months. It had a cooling fan replaced about 6 months ago but otherwise is doing what it is supposed to.
A dive computer is a much simpler device overall it's design considers a underwater working environment and life support but it uses much less power and produces little measurable heat. It's really just a calculator with timer and pressure transducers. The transmitter uses a bit more power but because of it's limited needed range still uses little power and produces little heat. Good designs should be rewarded with trust and poor ones should be avoided.
Do you use a smoke detector in your home? They are designed to last 5 years and then should be replaced. During that time you should replace the battery yearly and check the detector monthly but you sleep every night and trust your life with it.
Granted the underwater environment is more dangerous to people but that is why we use redundant systems and most of us dive with a buddy. (a buddy is another redundancy)

I, too, still have a very old computer that runs like new (slow)... Now you are bringing in limits as to how long things are designed to last. Are you saying that a new Dive Computer should be more reliable than an old one, because the last statement you made in the post above was that 99% of all solid state electronics fail within the first 3 days, and that if it made it past that, there was only a 1% failure rate...

as for the car sensor - you are crazy if you think they design those parts to only last a year or so... they design them to last forever (or at minimum, the warranty period, which is at least 5 years on new vehicles), and some do, but some fail.

As for computer parts dying, how are you going to tell me that a piece of RAM installed in a computer for two years goes through more abuse than a two part dive computer and transmitter that are always in motion (potentially running into things, moved around from tank to tank), actually have pressure sensors, wireless transmitters, batteries, can be in hot temperatures or very cold temperatures... I have built hundreds of computers - nothing more than a tile floor and a ESD work bench... I had a dual cpu P3 server run for 4 years without ever being turned off or even restarted... again, not saying that there aren't some electronics out there that can and will work for seemingly ever, but that is the exception, not the norm...

again, you are looking for someone to back up your own theories and thoughts, and I am here to tell you, its just not going to happen... as everyone else has said already, the brass/glass spg is and will always be "ole reliable" in comparison to a wireless AI computer...