O2 Injection Point

There is a lot of talk about where to inject, but there is a little fact being ignored: if you move the inject point away from the sensors, that is the same as increasing the response time of the sensor. At the lowest tidal volumes, the response time could be 10 seconds if the inject point is on the exhale breathing bag, and sensors in the normal place on the inhale side of the scrubber. This will increase the PPO2 error.

The amount of the error can be calculated quite easily. Put the extra response time into the Rebreather model we posted, it comes up with 10% PPO2 error. That is the difference between PPO2 set point and actual PPO2. So a 1.0 set point will go up to 1.1 and down to 0.9, even if the injection is perfect. Even the average point can be 10% off under these conditions.

This may seem quite surprising, but is quite logical: if your tidal volume is low, and you put in a huge lag between the injector and the sensor, then you are breathing gas that has been through your lungs once, filling up the exhale bag and scrubber before you inject gas. When you inject gas, you are injecting too far downstream to deal with the gas in the scrubber and most of it in the exhale bag.

It is compounded with assymetric breathing patterns: breathing in fast and out slow.

I will post some graphs and figures from this when I have time over the next few days.

It means there is only one correct place to inject gas: right up close to the O2 sensors, which need to be on the scrubber inhale side (they cannot be in the inhale bag otherwise condensation problems would wreck/cause error in the sensors).
Alex

Hello Alex, thanks for the data/info. I didn't know the lag time from injection could be so long. But this is in a BM CL unit you're talking about, right? Originally, I think Jonny posted this ? about the OTS CL Meg. If we're talking about an OTS CL unit, I don't think there's a more convenient/ergonomic place to put the manual add button, than in the CLs. And putting it in the exhale lung would get it closer to the sensors. I suppose you could rig up something like the MKs have, with a button on the back side with a whip that goes into the head. But I think I would still prefer having the button/fittings where I can see them and keep from getting banged around if they were out of my sight. In my experience, the reaction time after pushing the button on the Prism exhale lung is negligable-3 or 4 seconds-and after I've taken a breath...
Obviously placement of the solenoid in the head is the way to go for the reasons you mentioned. In the Prism, the solenoid is just before the scrubber inlet, to give it the most time to mix with the loop before hitting the sensors. -Andy

There is a lot of talk about where to inject, but there is a little fact being ignored: if you move the inject point away from the sensors, that is the same as increasing the response time of the sensor. At the lowest tidal volumes, the response time could be 10 seconds if the inject point is on the exhale breathing bag, and sensors in the normal place on the inhale side of the scrubber. This will increase the PPO2 error.

(snip)

Cheers

Alex

ANDY,
Its a combination of placement, the divers tidal volume, the divers breathing rate and the volume of gas required between the inlet point and sensors...

\Each position has its advantage.. a reasinable trade of is just to have an injection point as close to the start of teh scrubber as possible, so the lag isnt that great..

Now that is a good point.
A fine balance between delay and too close then. I expect too close would be more tolerant to maintaining a stable SP than too far away.Brent

But I wasn't sure whether Alex understood how this thread started...
Both the Prism and the Inspo have man add in the exhale lung, which gets the inject point closer to the sensors and has no possibility for interference with mixing and PPO2 reading from a partial inhale of fresh O2 inject from an inhale CL inject point. So what is the advantage of having an O2 man add in the inhale lung? I'm aware that no design is perfect for everything, but this is one design feature I can't see any big advantage in... -Andy

If the sensors are a long way from the injection point, there is a lag, and this results in a PPO2 error. This is quite a subtle problem, and the magnitude of it only becomes apparent when one has a full rebreather model, or is testing on a breathing machine where inhaled PPO2 is being measured.

Cheers

Alex

The delay should be as short as possible, even if it causes the sensors to read high momentarily as O2 is injected. The result will be much tighter O2 control.

Wouldn't it would be much easier to control the PPO2 if rebreather manufacturers would use a continuous injection, e.g. with a needle valve and stepper motor or via a proportional valve or via a 'leaky valve' combined with a small solenoid?

As for the mixing: in order to prevent a 'slug' of O2 moving around the loop, couldn't this also be improved by injecting in different places (e.g. exhale lung and scrubber head or just before and after the scrubber)?

Of course better mixing would make it more difficult to verify response of sensors electronically...

Steven

Hi Steven,
Reading your post and this "slug of O2" moving around the loop... I do not believe that density differential of the gases in a Rebreather can lead to the formation of a "slug".Yann.

Hi Yann, don't look at it as a slug of "pure" O2, just a slug where the mixture is richer than in the rest of the loop.

On the SK* that I owned this was clearly visible in the output of the cells:
inject -> ppo2 RISES -> then drops again -> then rises again but not as high -> then drops again but not as low etc... it takes about 3 breaths to get it completely mixed. Don't expect a 'spectacular' rise, I'm talking about a small peak (maybe 0.1 bar PPO2) here.


steven

*On an eCCR you will probably not see this as clearly as the controller would probably display a "running average" of measured values. (e.g. take 16 samples every second, sort them, throw away the 4 lowest and 4 highest, average the remaining 8, and then insert in circular buffer of 8 values of which you show the running average.)