Trimix CF blending question

[omar]

Originally posted by oxyhacker
....Re the Graham's Law/ 3% error on helium mixes, that got kicked around a few years ago on one of the techdiving forums and no one seemed to be having any trouble with it.

As an aside, I do not put much stock into what is decided on techdiver or other lists regarding gas mixing. Very few people have the background to understand it. A good example are the claims about the stratification of gas(es) in tanks, helium leaking out etc. In fact a few years ago the consensus was that you could not analyze for helium.

The most recent fallacy is the 2 step gas analysis for helium. Where the mixer checks the oxygen content after the oxygen and helium are added and then again after topping with air. This sounds reasonable on first examination and it does provide a slightly better estimate of the helium content in the mix when compared to using ideal gas approximations without the two step gas analysis.

However, the physical properties of gases are not suspended or altered by doing a two step analysis and this method provides only a minor improvement in determining the real helium content of the mix. The two step analysis still suffers from the same problems when dealing with any gas mixture determined by using the Ideal gas laws. It is an approximation or estimate and it does not take into account real gas behavior, i.e. compressibility.

The assumption that the pressure and volume relationship is linear is still used and fill amounts or pressures are still determined using Boyle's law. Both of these assumptions do not work and are proven wrong every time someone fills tanks using fudge factors or correction factors.

What is interesting is that by using the 2-step method or the ideal gas equations the mix will always be light in helium (pun intended). The only one that I have seen acknowledge this is GI in an email to me a long time ago.

omar

 

[Capt Jim Wyatt]

Been busy doing the USS Wilkes-Barre every day this week. Using the new blender & mixing trimix. The learning curve was/is steep....

Biggest problem was flowing He into the blender stack & having it escape before the compressor inhaled it. Fixed that problem with a check-valve between the blender stack & the intake filter.

Next problem is the "LAG" time between changing flow rates of O2 and/or He and seeing the results on the Oxy & He analyzers. That lag time is around 2 minutes. Clearly I have to relocate the analyzers further upstream.

I am using 100% of the He in the bottles and getting average 18/45 mixes while shooting for 17/50. While shooting for nitrox 50 I am getting nitrox 43 on average.

This thing will definitely pay for itself soon.

I sure do appreciate all the VERY helpful information being provided here & look forward to more!!

 

[Uncle Pug]

Originally posted by Capt Jim Wyatt
Biggest problem was flowing He into the blender stack & having it escape before the compressor inhaled it. Fixed that problem with a check-valve between the blender stack & the intake filter.

Don't you have a filter on top of you blender? Seems that the air getting sucked into the mix would keep the helium from escaping.

BTW what compressor are you using and what oil? Any problems with overheating using such strong mixes?

For my EAN50 bottles I just add whatever O2 I need PP blending style and top with EAN32 from the bank... we analyze out at 50% +/- .5%

 

[oxyhacker]

This is actually a very interesting problem, and the reason we decided not to try and make too compact a mixer, or inject the gases too close to the top - a few extra inches of length are cheap insurance. Years ago when I worked tuning vehicles we found that many of them actually preformed worse with stubby "hotrod" filters - flat VWs and big Amal GPs especially needed a straight or conical stack between the carb mouth and the filter, since when the motor was running there was actually a cloud of gas-rich air floating above the carb mouth, being ejected and sucked in again with the pulsations of the pistons - without the stack it would be partially lost, leaning the mix. It's safe to assume a similar situation might exist with a continuous mixer.

With He the situation is worse since it may try to float out if the velocity in the mixer stack isn't enough to overcome its bouyancy. Airflow at very low pressure/suction can be very non-intuitive, with the main flow going one way and actually creating side and eddy currents going the other.

There's easy test you might want to try - just invert the mixer so the intake is on the bottom and the tube to the compressor is at the top, and see if the He in your final mix improves! Even better, temporarily add a foot or two of tubing to the intake end of your mixer too, so that any blowback is recovered, and see what happens.

Originally posted by Capt Jim Wyatt
Biggest problem was flowing He into the blender stack & having it escape before the compressor inhaled it. Fixed that problem with a check-valve between the blender stack & the intake filter.

 

[Uncle Pug]

Scott's mixer has the gases injected in the middle of the throat of a ventui... I also have about 6" of ball valve and tubing between the blender and the filter. The filter has its inlet holes at the bottom. I placed a large bag over the entire set up to see if that captured any escaping helium but noticed no effect on the post compression analysis. I though about inverting the intake and may give that a try to see if there is any change. From my mix logs it seems as though I am getting good use of the supply helium without losing too much if any. I need to get busy and do the math and see just how much helium is going into the cylinders and how much is going into the garage.

Jim... I have the same lag time of 2 minutes but it hasn't been a problem so far... it just took some fiddling at first to mark my calibration tape on the ball flow gauges for the various mixes.

 

[oxyhacker]

Two-step analysis may not be perfect, but it's hardly a "fallacy" - it's a useful tool, the only way for someone without a He analyzer to get a doublecheck on what is actually going into the tank. It's much better than nothing, and, as considerable experience in the field has shown, usually "good enough". I don't think it's any secret that two step analysis won't keep a mix from coming out light on the He if the mixer doesn't take into account compressibility, but that'st due to the physics of PP mixing not any inherent flaw in the method of analysis. And it might have prevented the recent death in Croatia, in a GUE trimix course, where the mixer added O2 twice rather than O2 and He.

But anyhow this thread is about continuous trimixing, and two-step O2 analysis works great there, since compressibility is squeezed right out of the picture as one is analyzing the gases at ambient.

As far as trusting info from net forums, while I am not as skeptical of their value as you seem to be (and as I recall both the stratification and selective leakage threads eventually reached the right answers), the key info I cited, that the 3% O2 sensor/helium error issue you bought up is not a problem with sensors using membrane diffusions barriers (which includes virtually all the sensors used in diving), comes from a paper from the Jefferson Accelerator Lab.

Now, has anyone heard anything about a "2/3% error due to cross sensitivity from
nitrogen when mixing trimix using a thermal He analzyer"?

Originally posted by omar

As an aside, I do not put much stock into what is decided on techdiver or other lists regarding gas mixing. Very few people have the background to understand it. A good example are the claims about the stratification of gas(es) in tanks, helium leaking out etc. In fact a few years ago the consensus was that you could not analyze for helium.

The most recent fallacy is the 2 step gas analysis for helium. SNIP
omar

 

[omar]

I guess I wasn’t specific enough. The fallacy that I was referring to was that the 2-step analysis eliminates the need for a helium analyzer. As I said the 2-step analysis is an improvement over only doing it once. But the only way to know for sure is analyzing for helium.

As far as “ a 2/3% error due to cross sensitivity from nitrogen when mixing trimix using a thermal He analyzer?” I am curious where this came from and what were the test conditions?

As you know all gases have a specific thermal conductivity and the term “cross sensitivity” is a misnomer because thermal analysis is non-specific. If this statement is referring to replacing one gas with another it would depend upon what gases are being used and how much of each. If it refers to oxygen and nitrogen, the thermal conductivity for both are virtually identical at 25 degC and replacing one with the other will not cause a 2/3% error.

With thermal conductivity analyzers the output will depend on what the calibration points are and the content of the reference gas is when it is compared to the test gas.

A temperature difference of 10 degC with a high helium content mix (>~65%) would cause a 2-3% error if there was no temperature compensation for the test gas to the reference point.

omar

 

[oxyhacker]

Gotcha, and I agree.

I think there's no question that 2 step analysis can produce a mix that is close enough for safe diving, especially with continuous mixing where compressibility isn't a factor, but it can never tell you exactly what is in the mix.

It's funny how that works - even if you can logically deduce what's in a mix, to within a high level of certainty, it's never as satisfying as being able to test for directly. Sort of like the difference between having your buddy, no matter how much you trust him/her, analyze your mix, and doing it yourself.

The cross sensitivity thing comes from Analox. I'm still trying to get some clarification from them as to exactly what they mean. But it raises an interesting question for you folks with He analzyers: when you mix up a heliox and analyzer it, how closely to the O2 and He analyzers typically agree?

Originally posted by omar
I guess I wasn’t specific enough. The fallacy that I was referring to was that the 2-step analysis eliminates the need for a helium analyzer. As I said the 2-step analysis is an improvement over only doing it once. But the only way to know for sure is analyzing for helium.

As far as “ a 2/3% error due to cross sensitivity from nitrogen when mixing trimix using a thermal He analyzer?” I am curious where this came from and what were the test conditions?

omar

 

[omar]

I have been giving some more thought to the "cross sensitivity" of a thermal conductivity analyzer to nitrogen and a resulting trimix error because of this "sensitivity".

If you were to run 100% oxygen across a thermal conductivity analyzer referenced to air the output would be about the equivalent of a 2.5%-3% helium mix referenced to air. This is to be expected because by nature thermal conductivity analysis is non-specific.

However, this does not mean that the analyzer will have this amount of error for a given helium content. The error would depend upon the calibration curve that the signal output is compared to to calculate a display value.

For a specific example: In the case of the Atomox helium analyzer the calibration curve is for the following points (you can send off an email to Atomox to verify this ;-)

(N2+ signifies that the balance of gas is primarily nitrogen but can include the common impurities found in the atmosphere)

O2%-He%-N2+%
----------------------
20/5/75
20/10/70
20/15/65
20/20/60
20/25/55
20/30/50
20/35/45
20/40/40
20/45/35
20/50/30
20/55/25
20/60/20
20/70/10
20/80/0
10/90/0

The specific value for each of these points is used for the calibration of the unit. Also the range of oxygen that is specified for an error of <1% is 10% to 40% O2.

The maximum amount of error from the replacement of oxygen with nitrogen would be from the difference in thermal conductivity between nitrogen and oxygen compared to the calibration point. The minimum amount of helium in the mix would also be a condition of the maximum error conditions because high helium contents will result in smaller signal changes as a result of changing the oxygen/nitrogen ratio. This mix would be 10/5/85 (not a very realistic mix) and would result in a -0.25% error of the output.

The difference the other way would be the replacement of nitrogen with oxygen with the minimum amount of helium in the mix. This would be 0.45% (a 40/5/55 mix - again not a very realistic mix).

So it comes down to what the calibration curve is.
I know that Atomox has a new Heliox analyzer in the pipeline that will do both O2 and He at the same time. It will be interesting to see the response of the sensors to a hot heliox.

omar

 

[roakey]

Originally posted by oxyhacker
And it might have prevented the recent death in Croatia, in a GUE trimix course, where the mixer added O2 twice rather than O2 and He.

I know it's fashionable to bash DIR and GUE, but check your facts first. No one died in a GUE training course...

Here's an accident analysis by Andrew via JJ:


Date: Tue, 18 Jun 2002 18:01:26 -0400
To: quest@gue.com
From: Jarrod Jablonski <jj@gue.com>
Subject: lessons learned

I am posting this message for Andrew Georgitsis who is out of touch training rebreathers in Japan. He sent me the following report but has since been out of email contact. I am posting it for him, knowing that he would prefer to see the information released.
Best wishes,
Jarrod

On this list, in our GUE classes, and within our projects we have consistently reiterated that divers must carefully identify the true risk of a given dive. With gas diving, oxygen toxicity remains the most consistent and yet often under appreciated risk. Irresponsible mixing, convoluted marking/procedures, and careless divers are almost exclusively the culprit in these cases. The following actual account depicts an all too common over confidence that nearly cost one diver their life. I encourage everyone to read this report and to appreciate the risk to which divers expose themselves and their team.

While conducting a Tech 1 in Croatia this last week we were faced with the following incident, one that should prove educational to all. It is for this reason that we wanted to bring it public attention. The course was conducted in Croatia, on an Island called Pag. The initial part of the training, DIR fundamentals and critical skills went well, with students undergoing training with myself and Richard Lundgren. After two days of fundamentals and 4 days of critical skills training, we were ready to move forward to the experience portion of the class. During this portion, the students plan and execute two dives to a max depth of 36 m, on a 30/30 triox mix and decompress on Nitrox 50.

The logistics of the course were coordinated by the local dive facility. The owner of the facility was involved with the class. Upon returning from the sixth day of training, Diver X who also is the facility owner, began the nightly filling process for the next day of diving. Oddly enough in Croatia, the same valve fitting (threads) is used for all gases, facilitating confusion for those that are not properly vigilant. In fact, diver X confused the supply bottles, accidentally filling oxygen instead of Helium into his own back tanks. Failing to heed the directions of his instructors to properly analyze his gasses, diver X marked all his cylinders as analyzed while, in fact, his back tanks were never analyzed. Clearly such a mistake placed the diver and his team at tremendous risk; in this instance it almost cost him his life. Individuals must bear in mind that these actions do not occur in a vacuum, and that rescuers and other team members are compromised when they must take extreme actions to safeguard the safety of other members.

Following the execution of all pre-dive drills (conducted in seven minutes) both groups proceeded with their dive. Nineteen minutes into the dive, Diver X gently flashed me with his light to get my attention. He pulled his regulator out of his mouth (as if he was giving me an OOA) but instead began to convulse at 36m. I immediately donated my regulator, grabbed his harness (with my left hand) and tried to put a regulator in his mouth. His convulsions were very strong and I could not initially get the regulator in his mouth. I proceeded to swim him closer to the wall to avoid being swept away by the current. Diver X convulsed for a solid two minutes. We then started up, my right arm under his right arm, holding him firmly while keeping the regulator in his mouth. During this process I tried to remain conscious of maintaining an open airway.

We then proceed up, with the help of my assistant, controlling all buoyancy with my left hand. It took about one minute to reach 17m; there Diver X began to convulse again. Following this convulsion I slowly led him to the surface. Upon surfacing, I called for the surface support boat, which initiated an emergency response. I removed my mask and his and prepared to start mouth-to-mouth breathing. Fortunately he was breathing, making this unnecessary. We then removed his gear and put him on the Zodiac.

Upon establishing that all divers were safely at the surface with no need for additional decompression, we decided to start back to the dock, where we could meet the ambulance. During the ride to shore, Diver X began to come around, his color improved and his breathing became more rhythmic. We administered oxygen during the return trip to the dock and by the time we had arrived, he was feeling and looking much better. Upon reaching the dock the ambulance took over, taking him to the hospital for further testing. Further testing proved that there was no lung damage or bends and only as a precaution did they treat him for near drowning.

Subsequent analysis proved that Diver X's cylinders had been improperly marked. At this point we re-analyzed all diving cylinders, demonstrating that the only improper marking had been on Diver X's cylinders. It turns out that although his tanks were marked for 30/30 he was, in fact, breathing 50.1%.
Although this event ended well it was clearly filled with dangerous potential. All divers should use this as yet another example of the importance in following careful procedures for gas analyzation and tank marking. This must be the case whether they fill themselves or whether others fill the tanks for them. In this particular case Diver X allowed his personal filling of the tanks and the accuracy of the other mixes to induce a false sense of confidence. The diver later told me that he is keenly aware of his error, embarrassed by his arrogance and grateful to be alive. It is my hope that this account will help others realize that short cuts often fail and seemingly impossible things can and do occur. Ignoring the proper procedures can cause an accident very quickly.

Andrew Georgitsis
GUE Training Director

---------------------------------------------------------------------
To unsubscribe, e-mail: quest-unsubscribe@gue.com
For additional commands, e-mail: quest-help@gue.com