Which CCR I think is the best...for me.

Deep Thought · 15th December 2005, 00:53

Quote: (Originally Posted by ScubaDadMiami)

During the demo that I attended, a diver trying out the unit accidentally left the DSV open. When the next diver put on the unit and climbed back into the spring, she stood there for about five minutes with it open while the instructor conducted the briefing.

When she put the mouthpiece in her mouth, "glurg, glurg, glurg." The instructor said, "Hey, everybody watching, this is the perfect chance for me to show you how the unit recovers from a flood." We watched as the the instructor instructed the diver, and the diver proceeded to clear the loop. After that, not only did this diver complete her demo but about seven more also did so after. I found this to be pretty impressive.

Yes impressive – but sorry not really - that example is not in itself impressive, but a common feature of OTS CL’s with T-pieces. With every Meg class I teach, students purposely pull the mouthpiece out and flood the hose in confined water and then proceed to drain water out of the exhale CL. I have yet to see any water getting into the canister in all the tests I have done. I have not had this same success rate with the inspiration, not sure but assume difference in design?

It is impressive in the sense that all rebreathers with OTS CL’s have the propensity to avoid fully flooding the loop if the mouthpiece pops out, etc…. I think it is a matter of terminology as clearing a hose flood and having the entire scrubber canister flooded is a different subject.

Water entering a DSV can still enter the scrubber can if the diver’s attitude in the water allows it to bypass the CL, or if the diver does not react correctly to the problem. If water gets to the scrubber can, the RPC is a definite advantage but not the only way to avoid caustic cocktails. The Meg, inspiration, and Classic KISS all have areas of stand-off where water can accumulate before interacting with the chemical.

I happen to like the concept of OTS CL’s for the possibility of floods, and I like the size and feel of the OTS CL’s on the Optima. Still an impressive unit, that much is clear.

Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out?

01RMB · 15th December 2005, 01:22

Quote: (Originally Posted by Gillenvy)

much of that is over my head but it remind me of a question which came to mind during the try out. Do all real time po2 computers average the two closest O2 sensors and ignore the 3rd as I think the optima/hammerhead electronics do? If so, does that mean you always need two reliable sensors to get an acurate read and thus only really have one extra sensor? If so, wouldn't it make more sense to have four sensors and dedicate a pair to each PO2 computer? If my line of thinking is correct, it seems like it would be nice to choose for yourself wich pair should be driving O2 injection rates...if one of the four O2 sensors was oviously erronious. Wouldn't it be advantagious to be able to isolate/overide the faulty pair and go off the pair that are giving the most like readings, no?

Gill
The new Vision electronics don't follow the old voting logic of picking the two closest sensors. It will vote out sensors that are not reacting to expectations ie when O2 is added.

Both controllers also look at all three sensors and the standby controller will take over from the first if there is no response or it is not acting according to expectations.

You can even loose the handset and the controllers will operate. The HUD will still monitor both controllers.

If you are having a real bad day - the Vision will still operate with one bad controller, two bad cells and no handset. Got to be happy with that.

***ScubaDadMiami*** · 15th December 2005, 01:24

Quote: (Originally Posted by Deep Thought)

Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out?

While I have no time on it yet, my understanding is that this was purposely done in testing for about 20 minutes, after which the unit was cleared and functioned without issue. I do not have the reference to show. However, I understand that this was done earlier.

***jradomski*** · 15th December 2005, 01:25

Quote: (Originally Posted by Gillenvy)

much of that is over my head but it remind me of a question which came to mind during the try out. Do all real time po2 computers average the two closest O2 sensors and ignore the 3rd as I think the optima/hammerhead electronics do? If so, does that mean you always need two reliable sensors to get an acurate read and thus only really have one extra sensor? If so, wouldn't it make more sense to have four sensors and dedicate a pair to each PO2 computer? If my line of thinking is correct, it seems like it would be nice to choose for yourself wich pair should be driving O2 injection rates...if one of the four O2 sensors was oviously erronious. Wouldn't it be advantagious to be able to isolate/overide the faulty pair and go off the pair that are giving the most like readings, no?

The only Rebreather I have ever seen with the ability of the USER to disable a specific sensor is the Infinito, which unfortunately has only been available to a select few.

On the HH electronics you can disable the solenoid firing and fly the unit manually by selecting an option in the primary handset.

Genesis · 15th December 2005, 04:23

The firmware I've got running on the bench will raise an alarm and refuse to fire the solenoid if any sensor is out of linearity with the other two.

It will still give you all the information it has, but will not automatically act on it, and will instead alarm and force you to use the big computer between your ears.

If the sensors come back into agreement, it will resume normal operation.

You can also look at the millivolt levels at any time off the menu with a couple of button presses.

IMHO if you have a sensor "fight" during a dive the smart thing to do is thumb it then and there. Voting logic makes it too easy to ignore single cell faults and let the computer continue to "drive" with potentially disasterous consequences.

Garbage in, garbage out - no matter how many times you process it.

There are multiple potential failure modes depending on the specifics of the design of the Rebreather. As I have discovered during testing of various things, linear circuits do NOT always fail in expected ways (I have had a so-called linear op-amp decide to go DECIDEDLY non-linear on one channel only - unprovoked! That was cute.....)

You can only validate so much. If you see what is potentially bad data IMHO you have to take the safe choice and alarm rather than act on potentially bad information.

If you have two cells which agree and one which does not, how do you know whether two are good and one bad or vice-versa?

If you have a HUD and controller (or two controllers, whether only one is "smart" or both are) which both read the same cells there is always the possibility that a fault in one will contaminate the sensor outputs and thus screw the other. I have intentionally done some odd things with O2 cells, including feeding voltage into them to simulate "leakage" from a short in the wiring. That screws them HARD for an extended period of time! Bench testing has shown that such an event makes the affected sensor(s) likely to read singificantly high for quite a while after the fault is cleared.

This was a sobering discovery - that you can basically "charge" cells like a battery! While nobody would do that on purpose, it might happen due to an electrical fault that leaks current into the sensor wiring. What's worse is that the cell will still react to a PO2 spike in that condition.....

Wanna get bent? That'd be a great way to have it happen, and the bad news is that you might not catch it later on investigation, because after the unit is broken down and dries out the fault could disappear without a trace.... It takes very little stray current to do this. The raw facts are that you have a 50mv signal @ 1.0 PO2 (roughly); into the standard and expected 10kohms to terminate these sensors you have only .01 ma of current coming off those sensors! ANY leakage is going to result in absolute garbage for your "reading".

The discovery of the ability to "charge" these cells led me to change my design so that cable faults to a handset or battery cannister cannot feed back into the the O2 sensors. It is still possible for a circuit fault in the head to do it - that possibility cannot be eliminated as amplification, conditioning and conversion is necessary for a computer to process the information, but you can shorten those paths and protect against it as much as possible. It also led me to incorporate verification of the A/D chain during each polling sequence so that the conversion chain can be declared "healthy". This, however does not and cannot verify that the actual inputs are good - only that the ADC and communication to it are ok.

The idea of two controllers each reading two cells is an interesting one. Three cells actually increase the risk of one being bad over two. The idea that one can continue the dive with only two working appears to me to be quite dangerous. It might be interesting to use only two cells on the controller (if they disagree alarm and force the user to fly manually) and a third on a completely-indepedant backup (e.g. connected to a VR3 or KISS-style PO2 meter), thereby providing electrical isolation between the one and two...

***caveseeker7*** · 15th December 2005, 04:59

Quote: (Originally Posted by jradomski)

The only Rebreather I have ever seen with the ability of the USER to disable a specific sensor is the Infinito

I believe the Ouroboros also lets you en- and disable the sensors individually.

***jradomski*** · 15th December 2005, 05:07

Quote: (Originally Posted by Genesis)

The firmware I've got running on the bench will raise an alarm and refuse to fire the solenoid if any sensor is out of linearity with the other two.

It will still give you all the information it has, but will not automatically act on it, and will instead alarm and force you to use the big computer between your ears.

If the sensors come back into agreement, it will resume normal operation.

You can also look at the millivolt levels at any time off the menu with a couple of button presses.

IMHO if you have a sensor "fight" during a dive the smart thing to do is thumb it then and there. Voting logic makes it too easy to ignore single cell faults and let the computer continue to "drive" with potentially disasterous consequences.

Garbage in, garbage out - no matter how many times you process it.

There are multiple potential failure modes depending on the specifics of the design of the Rebreather. As I have discovered during testing of various things, linear circuits do NOT always fail in expected ways (I have had a so-called linear op-amp decide to go DECIDEDLY non-linear on one channel only - unprovoked! That was cute.....)

You can only validate so much. If you see what is potentially bad data IMHO you have to take the safe choice and alarm rather than act on potentially bad information.

If you have two cells which agree and one which does not, how do you know whether two are good and one bad or vice-versa?

If you have a HUD and controller (or two controllers, whether only one is "smart" or both are) which both read the same cells there is always the possibility that a fault in one will contaminate the sensor outputs and thus screw the other. I have intentionally done some odd things with O2 cells, including feeding voltage into them to simulate "leakage" from a short in the wiring. That screws them HARD for an extended period of time! Bench testing has shown that such an event makes the affected sensor(s) likely to read singificantly high for quite a while after the fault is cleared.

This was a sobering discovery - that you can basically "charge" cells like a battery! While nobody would do that on purpose, it might happen due to an electrical fault that leaks current into the sensor wiring. What's worse is that the cell will still react to a PO2 spike in that condition.....

Wanna get bent? That'd be a great way to have it happen, and the bad news is that you might not catch it later on investigation, because after the unit is broken down and dries out the fault could disappear without a trace.... It takes very little stray current to do this. The raw facts are that you have a 50mv signal @ 1.0 PO2 (roughly); into the standard and expected 10kohms to terminate these sensors you have only .01 ma of current coming off those sensors! ANY leakage is going to result in absolute garbage for your "reading".

The discovery of the ability to "charge" these cells led me to change my design so that cable faults to a handset or battery cannister cannot feed back into the the O2 sensors. It is still possible for a circuit fault in the head to do it - that possibility cannot be eliminated as amplification, conditioning and conversion is necessary for a computer to process the information, but you can shorten those paths and protect against it as much as possible. It also led me to incorporate verification of the A/D chain during each polling sequence so that the conversion chain can be declared "healthy". This, however does not and cannot verify that the actual inputs are good - only that the ADC and communication to it are ok.

The idea of two controllers each reading two cells is an interesting one. Three cells actually increase the risk of one being bad over two. The idea that one can continue the dive with only two working appears to me to be quite dangerous. It might be interesting to use only two cells on the controller (if they disagree alarm and force the user to fly manually) and a third on a completely-indepedant backup (e.g. connected to a VR3 or KISS-style PO2 meter), thereby providing electrical isolation between the one and two...

I don't think its a good idea to automatically stop the addition of oxygen becuase 1 cell drifts.. This happens more frequently that you'd expect, but usually comes back into line.. just a small droplet of water can cause a cell error, the injecting of oxygen especially in designs that inject near the cells introduces so dry gas which can help the issue..

Also I have seen on more than 1 occasion where you get po2 spike on a specific cell becuase it doesnt equalize the pressure as fast as the others (especially true if oxygen is injected near the cells),

Voting logic works in most cases, there really isnt a good reason to abandon it, but it would be a good idea to be able to override it, if the big computer determines there is a problem..

Genesis · 15th December 2005, 05:28

Thanks Joe.... I'll see how it all works out underwater. One of the nice things about "rolling your own" is that changes are just a source-code edit and recompile/download - a few minutes.....

DepthCharge · 15th December 2005, 06:28

Quote: (Originally Posted by Deep Thought)

Yes impressive – but sorry not really - that example is not in itself impressive, but a common feature of OTS CL’s with T-pieces. With every Meg class I teach, students purposely pull the mouthpiece out and flood the hose in confined water and then proceed to drain water out of the exhale CL. I have yet to see any water getting into the canister in all the tests I have done. I have not had this same success rate with the inspiration, not sure but assume difference in design?

It is impressive in the sense that all rebreathers with OTS CL’s have the propensity to avoid fully flooding the loop if the mouthpiece pops out, etc…. I think it is a matter of terminology as clearing a hose flood and having the entire scrubber canister flooded is a different subject.

Water entering a DSV can still enter the scrubber can if the diver’s attitude in the water allows it to bypass the CL, or if the diver does not react correctly to the problem. If water gets to the scrubber can, the RPC is a definite advantage but not the only way to avoid caustic cocktails. The Meg, inspiration, and Classic KISS all have areas of stand-off where water can accumulate before interacting with the chemical.

I happen to like the concept of OTS CL’s for the possibility of floods, and I like the size and feel of the OTS CL’s on the Optima. Still an impressive unit, that much is clear.

Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out?

I was lucky enough to be on a trip when they were testing the Optima. Rudy flooded to the cannister, cleared and completed 10-15 min of deco. Back on the boat, he opened the cannister and dumped out at least 8 oz. of water. Impressive I'd say. Later,
John

silent running · 15th December 2005, 07:04

Quote: (Originally Posted by DepthCharge)

I was lucky enough to be on a trip when they were testing the Optima. Rudy flooded to the cannister, cleared and completed 10-15 min of deco. Back on the boat, he opened the cannister and dumped out at least 8 oz. of water. Impressive I'd say. Later,
John

Hello John, how exactly did he clear it underwater, did he roll it forward into the lungs? -Andy

15th December 2005, 00:53	#41 (permalink)
Deep Thought Custom Title Allowed! Current Rebreather/s: Inspiration Vision Classic Kiss Other Rebreather/s: Inspiration Classic Inspiration Vision Evolution Megalodon Sport Kiss Classic Kiss Optima Join Date: Apr 2005 Location: Great Lakes Posts: 190	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by ScubaDadMiami) During the demo that I attended, a diver trying out the unit accidentally left the DSV open. When the next diver put on the unit and climbed back into the spring, she stood there for about five minutes with it open while the instructor conducted the briefing. When she put the mouthpiece in her mouth, "glurg, glurg, glurg." The instructor said, "Hey, everybody watching, this is the perfect chance for me to show you how the unit recovers from a flood." We watched as the the instructor instructed the diver, and the diver proceeded to clear the loop. After that, not only did this diver complete her demo but about seven more also did so after. I found this to be pretty impressive. Yes impressive – but sorry not really - that example is not in itself impressive, but a common feature of OTS CL’s with T-pieces. With every Meg class I teach, students purposely pull the mouthpiece out and flood the hose in confined water and then proceed to drain water out of the exhale CL. I have yet to see any water getting into the canister in all the tests I have done. I have not had this same success rate with the inspiration, not sure but assume difference in design? It is impressive in the sense that all rebreathers with OTS CL’s have the propensity to avoid fully flooding the loop if the mouthpiece pops out, etc…. I think it is a matter of terminology as clearing a hose flood and having the entire scrubber canister flooded is a different subject. Water entering a DSV can still enter the scrubber can if the diver’s attitude in the water allows it to bypass the CL, or if the diver does not react correctly to the problem. If water gets to the scrubber can, the RPC is a definite advantage but not the only way to avoid caustic cocktails. The Meg, inspiration, and Classic KISS all have areas of stand-off where water can accumulate before interacting with the chemical. I happen to like the concept of OTS CL’s for the possibility of floods, and I like the size and feel of the OTS CL’s on the Optima. Still an impressive unit, that much is clear. Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out? __________________ Creature Of Habits; Mostly Bad www.gudcscuba.com
(Offline)

15th December 2005, 01:22	#42 (permalink)
01RMB Evolving Current Rebreather/s: Evolution Other Rebreather/s: Join Date: May 2005 Location: Brisbane, Australia Posts: 192	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by Gillenvy) much of that is over my head but it remind me of a question which came to mind during the try out. Do all real time po2 computers average the two closest O2 sensors and ignore the 3rd as I think the optima/hammerhead electronics do? If so, does that mean you always need two reliable sensors to get an acurate read and thus only really have one extra sensor? If so, wouldn't it make more sense to have four sensors and dedicate a pair to each PO2 computer? If my line of thinking is correct, it seems like it would be nice to choose for yourself wich pair should be driving O2 injection rates...if one of the four O2 sensors was oviously erronious. Wouldn't it be advantagious to be able to isolate/overide the faulty pair and go off the pair that are giving the most like readings, no? Gill The new Vision electronics don't follow the old voting logic of picking the two closest sensors. It will vote out sensors that are not reacting to expectations ie when O2 is added. Both controllers also look at all three sensors and the standby controller will take over from the first if there is no response or it is not acting according to expectations. You can even loose the handset and the controllers will operate. The HUD will still monitor both controllers. If you are having a real bad day - the Vision will still operate with one bad controller, two bad cells and no handset. Got to be happy with that. __________________ Cheers Rod
(Offline)

15th December 2005, 01:24	#43 (permalink)
*ScubaDadMiami* Moderator Current Rebreather/s: Optima Other Rebreather/s: Join Date: Sep 2005 Location: Florida, USA Posts: 2,075	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by Deep Thought) Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out? While I have no time on it yet, my understanding is that this was purposely done in testing for about 20 minutes, after which the unit was cleared and functioned without issue. I do not have the reference to show. However, I understand that this was done earlier. __________________ Howard Packer IANTD CCR Instructor Miami Beach, Florida CCRDiveTraining.com
(Offline)

15th December 2005, 01:25	#44 (permalink)
*jradomski* Moderator Current Rebreather/s: Inspiration Classic Sport Kiss Optima rEvo Other CCR Home Build Other Rebreather/s: Inspiration Vision Evolution Megalodon Classic Kiss rEvo Other CCR Home Build Join Date: Mar 2005 Location: "Da" Bronx Posts: 2,899	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by Gillenvy) much of that is over my head but it remind me of a question which came to mind during the try out. Do all real time po2 computers average the two closest O2 sensors and ignore the 3rd as I think the optima/hammerhead electronics do? If so, does that mean you always need two reliable sensors to get an acurate read and thus only really have one extra sensor? If so, wouldn't it make more sense to have four sensors and dedicate a pair to each PO2 computer? If my line of thinking is correct, it seems like it would be nice to choose for yourself wich pair should be driving O2 injection rates...if one of the four O2 sensors was oviously erronious. Wouldn't it be advantagious to be able to isolate/overide the faulty pair and go off the pair that are giving the most like readings, no? The only Rebreather I have ever seen with the ability of the USER to disable a specific sensor is the Infinito, which unfortunately has only been available to a select few. On the HH electronics you can disable the solenoid firing and fly the unit manually by selecting an option in the primary handset. __________________ Joe Radomski CCR Trimix Instructor Trainer ANDI Instructor Trainer Director #10 All posts are personal opinions and DO NOT reflect any affiliated agency unless specifically stated. Last edited by jradomski : 15th December 2005 at 01:27.
(Online)

15th December 2005, 04:23	#45 (permalink)
Genesis Bubbless Box of Death Current Rebreather/s: Home Build Other Rebreather/s: Home Build Join Date: Oct 2005 Location: Sunny Florida Posts: 1,394	Re: Which CCR I think is the best...for me. The firmware I've got running on the bench will raise an alarm and refuse to fire the solenoid if any sensor is out of linearity with the other two. It will still give you all the information it has, but will not automatically act on it, and will instead alarm and force you to use the big computer between your ears. If the sensors come back into agreement, it will resume normal operation. You can also look at the millivolt levels at any time off the menu with a couple of button presses. IMHO if you have a sensor "fight" during a dive the smart thing to do is thumb it then and there. Voting logic makes it too easy to ignore single cell faults and let the computer continue to "drive" with potentially disasterous consequences. Garbage in, garbage out - no matter how many times you process it. There are multiple potential failure modes depending on the specifics of the design of the Rebreather. As I have discovered during testing of various things, linear circuits do NOT always fail in expected ways (I have had a so-called linear op-amp decide to go DECIDEDLY non-linear on one channel only - unprovoked! That was cute.....) You can only validate so much. If you see what is potentially bad data IMHO you have to take the safe choice and alarm rather than act on potentially bad information. If you have two cells which agree and one which does not, how do you know whether two are good and one bad or vice-versa? If you have a HUD and controller (or two controllers, whether only one is "smart" or both are) which both read the same cells there is always the possibility that a fault in one will contaminate the sensor outputs and thus screw the other. I have intentionally done some odd things with O2 cells, including feeding voltage into them to simulate "leakage" from a short in the wiring. That screws them HARD for an extended period of time! Bench testing has shown that such an event makes the affected sensor(s) likely to read singificantly high for quite a while after the fault is cleared. This was a sobering discovery - that you can basically "charge" cells like a battery! While nobody would do that on purpose, it might happen due to an electrical fault that leaks current into the sensor wiring. What's worse is that the cell will still react to a PO2 spike in that condition..... Wanna get bent? That'd be a great way to have it happen, and the bad news is that you might not catch it later on investigation, because after the unit is broken down and dries out the fault could disappear without a trace.... It takes very little stray current to do this. The raw facts are that you have a 50mv signal @ 1.0 PO2 (roughly); into the standard and expected 10kohms to terminate these sensors you have only .01 ma of current coming off those sensors! ANY leakage is going to result in absolute garbage for your "reading". The discovery of the ability to "charge" these cells led me to change my design so that cable faults to a handset or battery cannister cannot feed back into the the O2 sensors. It is still possible for a circuit fault in the head to do it - that possibility cannot be eliminated as amplification, conditioning and conversion is necessary for a computer to process the information, but you can shorten those paths and protect against it as much as possible. It also led me to incorporate verification of the A/D chain during each polling sequence so that the conversion chain can be declared "healthy". This, however does not and cannot verify that the actual inputs are good - only that the ADC and communication to it are ok. The idea of two controllers each reading two cells is an interesting one. Three cells actually increase the risk of one being bad over two. The idea that one can continue the dive with only two working appears to me to be quite dangerous. It might be interesting to use only two cells on the controller (if they disagree alarm and force the user to fly manually) and a third on a completely-indepedant backup (e.g. connected to a VR3 or KISS-style PO2 meter), thereby providing electrical isolation between the one and two... Last edited by Genesis : 15th December 2005 at 04:30.
(Offline)

15th December 2005, 04:59	#46 (permalink)
*caveseeker7* Who loves ya, baby Current Rebreather/s: Sport Kiss Other Rebreather/s: Prism Topaz Join Date: Jan 2005 Location: Too far from Neverland Posts: 5,435	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by jradomski) The only Rebreather I have ever seen with the ability of the USER to disable a specific sensor is the Infinito I believe the Ouroboros also lets you en- and disable the sensors individually. __________________ Cheers Stefan "Political Correctness is a doctrine, fostered by a delusional, illogical minority, and rabidly promoted by an unscrupulous mainstream media, which holds forth the proposition that it is entirely possible to pick up a turd by the clean end.'!"
(Offline)

15th December 2005, 05:07	#47 (permalink)
*jradomski* Moderator Current Rebreather/s: Inspiration Classic Sport Kiss Optima rEvo Other CCR Home Build Other Rebreather/s: Inspiration Vision Evolution Megalodon Classic Kiss rEvo Other CCR Home Build Join Date: Mar 2005 Location: "Da" Bronx Posts: 2,899	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by Genesis) The firmware I've got running on the bench will raise an alarm and refuse to fire the solenoid if any sensor is out of linearity with the other two. It will still give you all the information it has, but will not automatically act on it, and will instead alarm and force you to use the big computer between your ears. If the sensors come back into agreement, it will resume normal operation. You can also look at the millivolt levels at any time off the menu with a couple of button presses. IMHO if you have a sensor "fight" during a dive the smart thing to do is thumb it then and there. Voting logic makes it too easy to ignore single cell faults and let the computer continue to "drive" with potentially disasterous consequences. Garbage in, garbage out - no matter how many times you process it. There are multiple potential failure modes depending on the specifics of the design of the Rebreather. As I have discovered during testing of various things, linear circuits do NOT always fail in expected ways (I have had a so-called linear op-amp decide to go DECIDEDLY non-linear on one channel only - unprovoked! That was cute.....) You can only validate so much. If you see what is potentially bad data IMHO you have to take the safe choice and alarm rather than act on potentially bad information. If you have two cells which agree and one which does not, how do you know whether two are good and one bad or vice-versa? If you have a HUD and controller (or two controllers, whether only one is "smart" or both are) which both read the same cells there is always the possibility that a fault in one will contaminate the sensor outputs and thus screw the other. I have intentionally done some odd things with O2 cells, including feeding voltage into them to simulate "leakage" from a short in the wiring. That screws them HARD for an extended period of time! Bench testing has shown that such an event makes the affected sensor(s) likely to read singificantly high for quite a while after the fault is cleared. This was a sobering discovery - that you can basically "charge" cells like a battery! While nobody would do that on purpose, it might happen due to an electrical fault that leaks current into the sensor wiring. What's worse is that the cell will still react to a PO2 spike in that condition..... Wanna get bent? That'd be a great way to have it happen, and the bad news is that you might not catch it later on investigation, because after the unit is broken down and dries out the fault could disappear without a trace.... It takes very little stray current to do this. The raw facts are that you have a 50mv signal @ 1.0 PO2 (roughly); into the standard and expected 10kohms to terminate these sensors you have only .01 ma of current coming off those sensors! ANY leakage is going to result in absolute garbage for your "reading". The discovery of the ability to "charge" these cells led me to change my design so that cable faults to a handset or battery cannister cannot feed back into the the O2 sensors. It is still possible for a circuit fault in the head to do it - that possibility cannot be eliminated as amplification, conditioning and conversion is necessary for a computer to process the information, but you can shorten those paths and protect against it as much as possible. It also led me to incorporate verification of the A/D chain during each polling sequence so that the conversion chain can be declared "healthy". This, however does not and cannot verify that the actual inputs are good - only that the ADC and communication to it are ok. The idea of two controllers each reading two cells is an interesting one. Three cells actually increase the risk of one being bad over two. The idea that one can continue the dive with only two working appears to me to be quite dangerous. It might be interesting to use only two cells on the controller (if they disagree alarm and force the user to fly manually) and a third on a completely-indepedant backup (e.g. connected to a VR3 or KISS-style PO2 meter), thereby providing electrical isolation between the one and two... I don't think its a good idea to automatically stop the addition of oxygen becuase 1 cell drifts.. This happens more frequently that you'd expect, but usually comes back into line.. just a small droplet of water can cause a cell error, the injecting of oxygen especially in designs that inject near the cells introduces so dry gas which can help the issue.. Also I have seen on more than 1 occasion where you get po2 spike on a specific cell becuase it doesnt equalize the pressure as fast as the others (especially true if oxygen is injected near the cells), Voting logic works in most cases, there really isnt a good reason to abandon it, but it would be a good idea to be able to override it, if the big computer determines there is a problem.. __________________ Joe Radomski CCR Trimix Instructor Trainer ANDI Instructor Trainer Director #10 All posts are personal opinions and DO NOT reflect any affiliated agency unless specifically stated.
(Online)

15th December 2005, 05:28	#48 (permalink)
Genesis Bubbless Box of Death Current Rebreather/s: Home Build Other Rebreather/s: Home Build Join Date: Oct 2005 Location: Sunny Florida Posts: 1,394	Re: Which CCR I think is the best...for me. Thanks Joe.... I'll see how it all works out underwater. One of the nice things about "rolling your own" is that changes are just a source-code edit and recompile/download - a few minutes.....
(Offline)

15th December 2005, 06:28	#49 (permalink)
DepthCharge New Member Current Rebreather/s: Inspiration Classic Home Build Other Rebreather/s: Home Build Join Date: Mar 2005 Location: upstate NY Posts: 230	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by Deep Thought) Yes impressive – but sorry not really - that example is not in itself impressive, but a common feature of OTS CL’s with T-pieces. With every Meg class I teach, students purposely pull the mouthpiece out and flood the hose in confined water and then proceed to drain water out of the exhale CL. I have yet to see any water getting into the canister in all the tests I have done. I have not had this same success rate with the inspiration, not sure but assume difference in design? It is impressive in the sense that all rebreathers with OTS CL’s have the propensity to avoid fully flooding the loop if the mouthpiece pops out, etc…. I think it is a matter of terminology as clearing a hose flood and having the entire scrubber canister flooded is a different subject. Water entering a DSV can still enter the scrubber can if the diver’s attitude in the water allows it to bypass the CL, or if the diver does not react correctly to the problem. If water gets to the scrubber can, the RPC is a definite advantage but not the only way to avoid caustic cocktails. The Meg, inspiration, and Classic KISS all have areas of stand-off where water can accumulate before interacting with the chemical. I happen to like the concept of OTS CL’s for the possibility of floods, and I like the size and feel of the OTS CL’s on the Optima. Still an impressive unit, that much is clear. Question to any of the Optima divers who have time on these units; anyone have experience with a full flood that entered the canister? In the event of a flood that wetted the RPC, could the dive be aborted on the rebreathers, were you able to clear the canister, or bail out? I was lucky enough to be on a trip when they were testing the Optima. Rudy flooded to the cannister, cleared and completed 10-15 min of deco. Back on the boat, he opened the cannister and dumped out at least 8 oz. of water. Impressive I'd say. Later, John
(Offline)

15th December 2005, 07:04	#50 (permalink)
silent running Mature mouth breather Current Rebreather/s: Prism Topaz Other Rebreather/s: Join Date: Jun 2005 Location: U.S.A. Brooklyn, New York Posts: 1,814	Re: Which CCR I think is the best...for me. Quote: (Originally Posted by DepthCharge) I was lucky enough to be on a trip when they were testing the Optima. Rudy flooded to the cannister, cleared and completed 10-15 min of deco. Back on the boat, he opened the cannister and dumped out at least 8 oz. of water. Impressive I'd say. Later, John Hello John, how exactly did he clear it underwater, did he roll it forward into the lungs? -Andy Last edited by silent running : 15th December 2005 at 07:07.
(Online)

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode