Accuracy of data presented in mV or PO2

Meng_Tze

Agree. That point was somewhat lost and is a core one

__________________
Homo Bonae Voluntatis

Sutty

Here's a possible failure mode of a computerised system: a factor must be stored in memory somewhere as a "calibration" figure. A value read from an A-to-D convertor is then multiplied by this factor to produce a PO2 value (there may be more than one step but this illustrates the point).

If the memory location is corrupted (eg memory location written to by faulty code elsewhere in the programming, ionising radiation, etc,etc) then you could multiply by an erroneous factor, and hence display an incorrect value of PO2. Of course there are ways to minimise this possibility, and its pretty unlikely anyway, And would only affect one cell readout unless such a stored factor was applied to all 3 readouts.

In the real world you would be really unlucky to get such an error with well written code. A random memory corruption is just as likely to hit anywhere else (eg in the memory holding the code).

However a completely separate 2nd system (or more), would allow you to cross-reference.

Neil

__________________
Never forget that life is a finite resource.

PhilSiswick

That, combined with the fact that (short of doing regular flushes) you can't really know whether the Vision has failed, as there is only one handset, is the reason I'm fitting a fourth cell. I'm surprised more people don't, just as a level of insurance/validation.

Cheers,

Phil

__________________
Phil (WSKD 0001)

I have always felt that the dive I am on is not nearly important as the dives I plan to be on the rest of my life. Tom Rose, 2007

The person who gets the farthest is generally the one who is willing to do and dare. The sure thing boat never gets far from shore. Charles A. Lindbergh

www.hugsac.org.uk

silent running

Hi Stuart, this is a good description of my own reservations about the length and complication of the signal chain from mv sensor to high current display, and which depends on a CPU and other hardware.

Speaking of those complications, isn't there also a pc of hardware required to average the fluctuating mv signal from the sensor in order to give a stable numeric value on the display? I wonder how this pc of gear separates the noise of an unstable cell from the normal up and down readings of a cell under continuous changing pressure as the diver swims in the water collum, while still managing to give a usable PO2 reading? -Andy

bgpartri

I think there are several issues here.

1) Adding a secondary display is a good idea. Whether you add a HUD, put one kiss display on to one of the sensors in addition to your computer, or put a fourth cell in through the offboard O2 port, adding a redundant display is a good idea.

2) The mV display and the PPO2 display are based on the same signal whether it is a KISS display or a computer - the output from the ADC. Is it more likely for the computer to fail? Some failures are more common and some are less common.

Can the mV display help trouble shoot? Yes it can. On the surface, it can allow you to spot broken wires and plugs on backwards. It can theoretically help you spot a current limited cell, but in reality it can't.

I don't find the mV display much use underwater.

3) But the real issue that I think is being missed here is that if you don't have a feel for a PPO2 display that is way off, you are in trouble anyway. Displays of any type that are showing a changing but wrong PPO2 due to an internal problem are very rare. Off the top of my head I can't think of a single instance. I have heard of, but never seen, a display hanging underwater, but I would think that the lack of changes to depth, PPO2, time, etc. would be pretty obvious.

However, wrong displays due to bad calibration, broken wires, flooded displays, current limited sensors, etc. are absolutely routine and certain to happen if you dive enough.

My point is that thinking that redundant displays on a system will give you reliable PPO2 readings makes me think that you are concentrating on finding unicorns when there are horses running all around you.

If you have done a bad calibration or you have three current limited sensors, all of your redundant displays will tell you the same lie. Both of these things have happened to me.

My current practise is to treat the sensors like they are a sneaky poker player. I'm never sure whether they're bluffing or not. So I do things to verify them. I calibrate in O2 then check them in air. I flush to O2 at 20 feet on every dive. I watch to see what they do on level swims. I listen to the solenoid firing pattern. I watch changes as I ascend or descend.

The more I do this stuff, the more I have come to understand that the O2 sensors and our treatment of them is one of the key weaknesses in rebreather diving.

YMMV

Bruce

__________________
http://www.shearwaterresearch.com

Good judgment comes from experience. Experience comes from bad judgment.

silent running

Thanks Bruce for this post and the earlier one, lots of good stuff to think about.

Your explanation of the way mv readings are derived and displayed is very helpful, but confirms what I previously thought-mv display is not of any use underwater. The mv display readings on the surface could take the place of a multi-meter, which is nice as you won't have to bring one an the trip. But would the surface mv values be affected by a bad cal and not quite as accurate as using a muti-meter directly on the sensors if you have reason to believe the sensor is faulty?-Andy

divetheworld

Strange that when you look back through this thread, you are all discussing some of the fundemental principles of compliance with EN61508. Yes, you are!

There is a difference between safety and reliability. Dr Mike put some excellent links up in another thread. If we are looking at potential failures from electronic components then we must consider all potential faults, whether hardware or software and calculate the failure probability and then mitigate that failure possibility by a combination of redundancy and diversity. Diversity is on of the best ways to achieve safe reliable systems as there will be no common mode fault between a system that uses diverse electronics. A good example is the Deep Pursuit as the Primary and secondary share absolutely no common components, firmware or software so a manufacturing fault, software problem or system error is unlikely to affect both critical systems. If this principle is applied to every critical part from the wiring through to the injection system then you will have a system that is fault tolerant with adequate redundancy and can be relied upon by the user.
At present, no system commercially exists that has sufficient inherrant safety and reliability to meet SIL3 or 4. But it can be done, without great cost.


Brent

__________________

lizardland

Like I said in the other thread, having a secondary is a good way of checking that the primary is working, not that the cells are working. By primary, I'm including all the supplementary parts, cabling, etc.

Another thing that I really do miss is Gordon's "dancing digit" i.e. the third decimal place, useless as a cell reading but seeing it move told you everything was still live.

Maybe having dived homebuilds for longer than I've dived a "real" rebreather has skewed my viewpoint, too long assuming that every system is inherently flawed.

__________________
Can you imagine drifting along in the sea with your mouth open and a load of f***ing plankton going in? You'd like it, would you?

www.westons-cider.co.uk

Azerbaijani Association of Technical Divers Publicity Officer and Goat Wrangler

PhilSiswick

It may take longer than with a third decimal place, but I always wait to see the second decimal place change on at least one cell before I end my handset check. To shorten the time I need to wait, I time it to coincide with the solenoid firing. It'll capture a frozen handset too.

Cheers,

__________________
Phil (WSKD 0001)

I have always felt that the dive I am on is not nearly important as the dives I plan to be on the rest of my life. Tom Rose, 2007

The person who gets the farthest is generally the one who is willing to do and dare. The sure thing boat never gets far from shore. Charles A. Lindbergh

www.hugsac.org.uk

David Pye

That's not particularly helpful, is it? Admittedly I was a bit unclear but I think we're using the word accuracy differently.

When I say I don't care about accuracy, I mean that I don't care about say +/- 0.1bar error in PPO2 for example. I do care if the data displayed on the screen is from 10 minutes ago, or bears absolutely no relation to the voltage output from the cells. That (to my mind) isn't accuracy in the terms of error margin.

If that makes sense.

David