Current State Of Rebreather Electronics

[stuartv]

You’re forgetting the analog secondary is powered by the sensors and not by a high current battery. Millivolts work in the 100% humidity of a loop, 9+ volts do not like humidity of any kind.

Where is there a 9V battery in the configuration I described? The only 1 ATA compartments would be the battery compartments inside the NERD or Petrel.

Giving a deco computer incorrect info? Why do you assume that all the MV readings given to your SP controller/deco computer are going to translate into the completely “correct” reading? Never mind the various glitches in components which can occur between reading millivolts and converting them into digital data inside a CCR. Do you analyze your O2 every day, even when you are diving with the same operation multiple days and you watch the person fill the tank each day? Are you using their O2 analyzer? How old is the cell in it? Are you or someone else calibrating it everyday? Add to that the fact that deco calculations are still just an educated guess, given the scant data available for constant PO2 diving. O2 exposure is even less well understood and tested. Even when you are confident about most variables, your displayed PO2 is still an educated guess because the computer only “knows” so many variables. The computer has no knowledge of whether your CCR was handled roughly by the boat crew the day before, or left on the boat deck in the hot sun for hours earlier that day. But if your sensors suddenly start behaving unpredictably, you might think of these things and wonder about the readings you’re getting. You might remember the sensor wiring harness looked a little frayed last time you prepped, even though you thought you put a new one in last trip. But did you?

I'm not assuming all the mV readings are correct. I'm assuming that there are enough correct readings given (i.e. at least one) for me to use the computer between my ears to know which ones are correct (if any) and decide whether to trust the resulting deco calculations.

And yes, I analyze my own tanks after every fill, using my own analyzer. Yes, I check the calibration each time I use my analyzer.

Yes, we know that none of this is perfectly accurate, nor does it perfectly model the physical process in our bodies.

But, when we do our best to give the electronics the most accurate info we can, the result has a very, very strong track record of getting us out of the water undamaged. Certainly, the modern electronics do a better job of keeping us undamaged when we we give them the most accurate info we can versus knowingly giving it inaccurate info (e.g. telling a computer we're using pO2 of 0.7, when we are really flying manually at 1.3)..

 

[RainPilot]

Some responses below:

not being able clear error codes

Error codes indicate an issue with the function of the unit (dead sensor, leaking loop etc). You clear them by rectifying the problem. Its like a low fuel light in your car. You clear it by adding fuel.

poor service

I have heard that, about the US specifically, and also from Poseidon in Sweden. Its getting better but it is a valid issue.

lockouts due to the internal calendar

The unit requires a service every 2 years, once you have gone 2 months over that date it will not pass the predive checks. That being said, the clear code is hardly secret knowledge, although I feel personally that a CCR diver should have their unit serviced regularly, 2 years is not crazy IMO. If you do the service course (like at DEMA for example) then you can service, clear and buy parts for your own unit no worries

several reports of rebooting underwater.

I shall research that, I don't know anyone who has had that problem but there's a lot of the units out there

With over 40 system checks before entering the water, there certainly is a lot that can go wrong.

Those checks are ALL things that the diver should be checking on ANY unit before a dive. Check battery voltages, sensor calibration, pos pressure loop test, BOV operation, gas contents confirmation etc

The fact that there is no way to know your PO2 in the event of an electronics failure or dying battery

On a rec setup, maybe, but that is predicated on "any error is a bailout" as advocated by all Rec CCR agencies. In a Tec config, I have a 3rd cell reader so 2 handsets reading 3 cells, same as most every unit out there.

You can’t even manually add O2

I have factory MAV on dil and O2 on my unit, both for my OTS CL and my BMCL. If you buy the full version of the unit, it comes with those included. Again, a rec version doesn't but you can just order them when you decide to go tec. They are standard drysuit inflation valves as used by many units so you don't need to stick with the factory ones.

 

[silent running]

Where is there a 9V battery in the configuration I described? The only 1 ATA compartments would be the battery compartments inside the NERD or Petrel.

If another handset computer of any make was hooked up to the loop there would be a 1ATA battery compartment hooked up to it and a cable with battery voltages flowing through it, another liability/failure point. And having a controller in the handset is completely unacceptable to me because it requires two way communication to the head and display at battery voltages, that’s just asking for trouble.

Battery voltages vary from unit to unit, but I don’t think many are below 9 volts, some are more, hence the +...

I'm not assuming all the mV readings are correct. I'm assuming that there are enough correct readings given (i.e. at least one) for me to use the computer between my ears to know which ones are correct (if any) and decide whether to trust the resulting deco calculations.

Yes, and the less complicated the hardware between those sensor readings and a display, the better the chance you will get accurate info. I much prefer my needle gage secondary PO2 meter to 9 flickering digits. Because each time I scroll through the sensors, they are loaded up by the gage and I can see how fast they come up to PO2, which gives me a nice graphic indication of their output/behavior. If one of them is on its way out, it may rise more slowly the next time I scroll through. I can predict a cell getting voted out by the electronics after watching the needle gage. Simple and informative, with no A/D conversion or computer interpreting in between the cell and I.

And yes, I analyze my own tanks after every fill, using my own analyzer. Yes, I check the calibration each time I use my analyzer.

I tend to ask ahead what purity their O2 is and always check it in the beginning. The CCR calibration will tell if something is off from the previous fill. I find it interesting that the shop/dive op analyzer usually says something different from mine. Also repeated analysis of the same gas can bring slightly different results, even in a lab it’s sometimes hard to get the same results...

Yes, we know that none of this is perfectly accurate, nor does it perfectly model the physical process in our bodies.

But, when we do our best to give the electronics the most accurate info we can, the result has a very, very strong track record of getting us out of the water undamaged. Certainly, the modern electronics do a better job of keeping us undamaged when we we give them the most accurate info we can versus knowingly giving it inaccurate info (e.g. telling a computer we're using pO2 of 0.7, when we are really flying manually at 1.3)..

Most ECCRs have auto SP switch at 6 or so M, secondly, if your unit thinks you are at.7 and clears your deco while you’re really at 1.3, all the safer. And presumably while you are at your last stop, you would have plenty of time to check these settings and change them to be in sync.

There are accounts on the web of popular ECCRs rebooting and recalibrating underwater, buttons freezing/failing, handsets shorting out during a dive and taking the whole controller down, handsets failing to fire the solenoid because of a corroded contact after returning from factory service. The number of outright design flaws and mistakes by manufacturers made during service which have been found over the years is alarming to say the least. Some of us old timers have a hard time forgetting these things.

Are electronics getting better? Yes to some degree, but they are also getting more complicated and divers are getting more dependent on technology, and many of us are skeptical after seeing divers miss dives because of something silly like a solenoid error message. IMHO every CCR should be able to be run manually and every diver should be trained to do so, or it’s not a safe unit or a safe diver.

If I have to BO off a loop for anything short of a CO2 hit, it better be my own fault because I consider getting off the loop to be the absolute last resort. Once you do, you most likely now have less time to calm down, think and solve a problem, which is one of the great things about CCR. Real emergencies underwater are not a fun experience, no matter how experienced you are. Have you ever been caught in a 3+ knot tidal down current for 30 minutes? I have, and it would be almost impossible to bail off the loop in such a situation, should my fancy ECCR decide it didn’t like my sudden PO2 spike and a sensor being voted out after getting water on the face from tumbling end over end down to 60m in -15 seconds.

A modern ECCR computer collects lots of information and can make calculations very fast, but it doesn’t have a body full of sensory organs and a human brain with which to evaluate the total environment, the computer simply doesn’t have all the information. We do, and I would prefer to use all that information, human and electronics, together to make my own decisions...

 

[silent running]

Some responses below:



Error codes indicate an issue with the function of the unit (dead sensor, leaking loop etc). You clear them by rectifying the problem. Its like a low fuel light in your car. You clear it by adding fuel.



I have heard that, about the US specifically, and also from Poseidon in Sweden. Its getting better but it is a valid issue.



The unit requires a service every 2 years, once you have gone 2 months over that date it will not pass the predive checks. That being said, the clear code is hardly secret knowledge, although I feel personally that a CCR diver should have their unit serviced regularly, 2 years is not crazy IMO. If you do the service course (like at DEMA for example) then you can service, clear and buy parts for your own unit no worries



I shall research that, I don't know anyone who has had that problem but there's a lot of the units out there



Those checks are ALL things that the diver should be checking on ANY unit before a dive. Check battery voltages, sensor calibration, pos pressure loop test, BOV operation, gas contents confirmation etc



On a rec setup, maybe, but that is predicated on "any error is a bailout" as advocated by all Rec CCR agencies. In a Tec config, I have a 3rd cell reader so 2 handsets reading 3 cells, same as most every unit out there.



I have factory MAV on dil and O2 on my unit, both for my OTS CL and my BMCL. If you buy the full version of the unit, it comes with those included. Again, a rec version doesn't but you can just order them when you decide to go tec. They are standard drysuit inflation valves as used by many units so you don't need to stick with the factory ones.

Thanks for all the info, glad to hear the tech version has manual adds, but it does make my point about over reliance on technology that the entry level Poseidon diver is not trained for manual operation.

While I don’t think CCRs are death traps, I do think they need to be taken very seriously and only very motivated people should be diving them. The trend towards automation goes against this by blurring the line between fault tolerance and idiot proof. I think CCR is a long way from mass appeal and may never achieve that, so what is the hurry to make them more appealing to people who may not really take to the activity in the long run, and might possibly become an injury or fatality in the short run?

As for predive checking, when you are diving multiple times a day, or on a multi day trip with a schedule, it’s often not feasible to do a complete system check between dives. And the simpler the unit is, the easier it will be to stay on top of vital functions. If I’m staring at a computer screen watching a 40 item checklist scroll through, I wonder if I’ll be as likely to notice the zip tie on my mouth piece is nicked, or that one of my BO regulator gages is broken...

 

[stuartv]

If another handset computer of any make was hooked up to the loop there would be a 1ATA battery compartment hooked up to it and a cable with battery voltages flowing through it, another liability/failure point. And having a controller in the handset is completely unacceptable to me because it requires two way communication to the head and display at battery voltages, that’s just asking for trouble.

Battery voltages vary from unit to unit, but I don’t think many are below 9 volts, some are more, hence the +...

That's exactly what I said - 1 ATA compartments for the internals of the NERD and Petrel.

But, where did I say anything about a controller? Where is the need for a 9V battery?

I said to have the NERD and replace the PO2 display with a Petrel. I didn't say anything about being a controller. So (as I said earlier), you'd have the same 2 cables and glands as if you had a NERD and your analog PO2 display.

And you'd have the same number of batteries and 1ATA compartments as your scenario of a NERD, an analog PO2 display, and a standalone backup deco computer.

And you'd actually have one less device to worry about (i.e. no analog PO2 display).

 

[stuartv]

Thanks for all the info, glad to hear the tech version has manual adds, but it does make my point about over reliance on technology that the entry level Poseidon diver is not trained for manual operation.

For the initial CCR training - where the diver is limited to recreational depths (or less) and no deco, I don't see the problem with training that teaches the very simple rule: If there is any problem, switch to BO and end the dive. In this case, the over reliance you are talking about seems to only have the downside that the diver might have some dives cut short or missed altogether.

 

[tbone1004]

@silent running uhh what? The analog secondary is not powered by the sensors, to your point, they don't deliver nearly enough current to do much of anything. The monitors are all drive by a real battery.

The 9v battery is to drive the solenoid because lower voltage solenoids that can run with something like an 18650 are usually physically larger due to higher amp draw. The 9v can be used with stepdown converters to run a board, but the only time any real voltage draw is being used is when the solenoid fires. That voltage does not have to go between the handset and the head.
So in a DiveCAN setup, you have a 9v that is driving the canbus board *VERY low draw*, and the solenoid *intermittent fairly low draw*. The analog mv connection is then isolated and very short which minimizes chance of electrical interference *a problem if you have analog wires in the same cable as the solenoid power cable, a la Predator controller, Hammerheads, etc, where that impulse can cause a jump in the mV reading*, and with the short distance effectively removes any risk of voltage drop due to the cable length. Shouldn't be anything relevant, but it is there.
The cables going out are now spitting out digital signals that can report when there is an error in the signal.

Is it more complicated? Obviously, with very simple systems like this digital is always going to be more complicated than analog, but it's also more reliable and contrary to your point, minimizes the risk of a faulty reading because the the cells can plug directly into the board that is processing their signals. They don't have to plug into a cable, that then plugs into a computer, than then analyzes the signal.
The computers are already converting the analog mV value to a digital signal, processing it, and sending it out, but with the canbus systems, you are just allowing it to be more reliable. Two-way communication does have to happen, but it's all digital with no power transmission which is fine.

On your analyzer comment, are you calibrating with air? If so, does your analyzer factor in barometric pressure, temp, and humidity? Even if it does, you are relying on a purely linear output from the sensor at 5x the calibration value which is exceptionally rare. Having good O2 to calibrate to at ppO2=1.0 gives you infinitely more repeatable analysis and any change in the calibration is likely your sensors themselves shifting. Over-calibrating is not a good thing.... Most dive shops will calibrate at air, usually not adjust for temp/pressure/humidity, and when you go to analyze O2, even a few percent in linear deviation will make a surprisingly large difference in the gas analysis.

Things like the Poseidon were designed to be idiot-proof, and have a lot of things build into them to make them idiot-proof, to varying degrees of success. I don't think that's the answer. I think the DiveCAN concept of going to digital communication is vastly superior to analog communication, I run my Meg at 0.6 on the solenoid and manually run it at 1.0-1.1. I much prefer that to the concept of CMF rebreathers *which I also own* for a myriad of reasons, but at the same time, I'm a cave diver, and running off of a solenoid in most of the caves in Florida is horrifically inefficient because of the ups and downs. In training I had to maintain flat ppO2's as an emphasis on control and attention to the unit, but in reality my ppO2 in certain passages will bounce between 0.7 and 1.2 especially if I'm on a scooter. If I wasn't running it manually, the solenoid would be freaking out and my buoyancy would go to hell in a handbasket every time I was ascending.
I don't want my unit to lock me out, ever, and they don't.
One thing I do like with the Liberty is that you can choose to vote out a sensor yourself. I think this is a MASSIVE safety improvement over the standard voting logic process.

 

[silent running]

That's exactly what I said - 1 ATA compartments for the internals of the NERD and Petrel.

But, where did I say anything about a controller? Where is the need for a 9V battery?

I said to have the NERD and replace the PO2 display with a Petrel. I didn't say anything about being a controller. So (as I said earlier), you'd have the same 2 cables and glands as if you had a NERD and your analog PO2 display.

And you'd have the same number of batteries and 1ATA compartments as your scenario of a NERD, an analog PO2 display, and a standalone backup deco computer.

And you'd actually have one less device to worry about (i.e. no analog PO2 display).

Sorry for the misunderstanding, I see what you are suggesting, and it’s fine if you want full computer redundancy, but with that comes complications with batteries of whatever higher voltage, which are connected to sensors through cables, which have connections which can corrode and fail.

To be clear I do not view my analog secondary to be anything to worry about, quite the opposite. The ability to isolate the low current sensor output side from the high current battery side (Solenoid, HUD, SP controller) is a huge benefit from a trouble shooting standpoint and gives me the ability to turn off the electronics entirely if there is a short, dying battery or a malfunctioning solenoid. I view any battery voltage flowing underwater through cables, some of which are not even designed for marine applications, to be a liability. And to reiterate, there is only low current/MV passing through the secondary cable, which means no chance of a short. The only failures of them I have seen is the cable glans being kinked enough to cause 1 of the sensors not to read, no affect on any other sensor or part of the system.

I was using the example of controllers in handsets to further illustrate the point that two way communication of information powered by battery voltage is even more vulnerable to failure from shorts/water ingress/battery failure and bad connections. Sorry for muddying the waters with that, but there was a recent near death situation from just that issue and I thought it a relevant, if not totally applicable thing to mention in this general discussion of CCR electronics.

A standalone computer does not have real time sensor information, but it does have the added benefit of not being able to screw up/short out anything else it might be connected to because of water ingress and stray voltage. Simplicity in life support makes me all warm and fuzzy..

 

[silent running]

@silent running uhh what? The analog secondary is not powered by the sensors, to your point, they don't deliver nearly enough current to do much of anything. The monitors are all drive by a real battery.

Incorrect, I’m speaking of the Prism 1 which like the MK15’s, uses higher output sensors to drive needle gages directly and requires no battery to display PO2.

The 9v battery is to drive the solenoid because lower voltage solenoids that can run with something like an 18650 are usually physically larger due to higher amp draw. The 9v can be used with stepdown converters to run a board, but the only time any real voltage draw is being used is when the solenoid fires. That voltage does not have to go between the handset and the head.
So in a DiveCAN setup, you have a 9v that is driving the canbus board *VERY low draw*, and the solenoid *intermittent fairly low draw*. The analog mv connection is then isolated and very short which minimizes chance of electrical interference *a problem if you have analog wires in the same cable as the solenoid power cable, a la Predator controller, Hammerheads, etc, where that impulse can cause a jump in the mV reading*, and with the short distance effectively removes any risk of voltage drop due to the cable length. Shouldn't be anything relevant, but it is there.
The cables going out are now spitting out digital signals that can report when there is an error in the signal.

Is it more complicated? Obviously, with very simple systems like this digital is always going to be more complicated than analog, but it's also more reliable and contrary to your point, minimizes the risk of a faulty reading because the the cells can plug directly into the board that is processing their signals. They don't have to plug into a cable, that then plugs into a computer, than then analyzes the signal.
The computers are already converting the analog mV value to a digital signal, processing it, and sending it out, but with the canbus systems, you are just allowing it to be more reliable. Two-way communication does have to happen, but it's all digital with no power transmission which is fine.

I have no problem with Canbus, sounds like a good system if you need to move around lots of digital info. The Prism MK15 architecture allows the isolation of the low current analog sensor side from the high current computer side, no need to pick a side you like better, they either play nice or you separate them..:-)

[QUOTE="tbone1004, post: 8760076, member: 136424"][USER=24468]
On your analyzer comment, are you calibrating with air? If so, does your analyzer factor in barometric pressure, temp, and humidity? Even if it does, you are relying on a purely linear output from the sensor at 5x the calibration value which is exceptionally rare. Having good O2 to calibrate to at ppO2=1.0 gives you infinitely more repeatable analysis and any change in the calibration is likely your sensors themselves shifting. Over-calibrating is not a good thing.... Most dive shops will calibrate at air, usually not adjust for temp/pressure/humidity, and when you go to analyze O2, even a few percent in linear deviation will make a surprisingly large difference in the gas analysis.[/QUOTE]

I don’t currently own an analyzer, had borrowed one for my last trip. And yes, of course I calibrated with O2. But good O2 is not always available in all parts of the world.

[QUOTE="tbone1004, post: 8760076, member: 136424"][USER=24468]
Things like the Poseidon were designed to be idiot-proof, and have a lot of things build into them to make them idiot-proof, to varying degrees of success. I don't think that's the answer. I think the DiveCAN concept of going to digital communication is vastly superior to analog communication, I run my Meg at 0.6 on the solenoid and manually run it at 1.0-1.1. I much prefer that to the concept of CMF rebreathers *which I also own* for a myriad of reasons, but at the same time, I'm a cave diver, and running off of a solenoid in most of the caves in Florida is horrifically inefficient because of the ups and downs. In training I had to maintain flat ppO2's as an emphasis on control and attention to the unit, but in reality my ppO2 in certain passages will bounce between 0.7 and 1.2 especially if I'm on a scooter. If I wasn't running it manually, the solenoid would be freaking out and my buoyancy would go to hell in a handbasket every time I was ascending.
I don't want my unit to lock me out, ever, and they don't.
One thing I do like with the Liberty is that you can choose to vote out a sensor yourself. I think this is a MASSIVE safety improvement over the standard voting logic process.[/QUOTE]

Yes, sounds like step in the right direction.

I can vote out the sensor myself too, by turning off the electronics on the Prism and have done so in order to further diagnose a problem. If I need to, I can isolate the low voltage sensor side of the Prism and can lock out the electronics if they misbehave, not the other way around..:-)[/user][/user]

 

[JohnnyC]

I've come to a conclusion. We should all just dive RB80's with the P-port plugged.