Manually voting out bad cells with Shearwater

[mr_v]

TL;DR: Shearwater should allow us to manually vote cells; why not?

Howdy folks,

If we track mV or if Shearwater does it for us, should we not be able to to identify bad cells and vote them out?

There are at least several ways of accomplishing that.

First, there is a pure manual way. You record mV in air, calculate the desired mV at your set points. Write it on a piece of tape and take the info w you on a dive. You can check mV at known depths if the cell is bad, vote it out manually.

Second, do this as a part of your "cells are having fun" routine. After performing a dil flush, note cells that are not showing the right PPO2, vote them out.

Third, use a bit more software power to do things semi-automatically. Just like option #2, but let Shearwater do calculations for you and make a recommendation.

Fourth, Shearwater tracks cell deviation overtime and alerts when cells are bad so you can vote them out or take them out. Will require a bit more coordination with cell changing procedures.

In other words, there are a few ways of thinking about it... Why not do it?


Cheers, --v

P.S. I got 10 hours on CCR, so if I missed something, please correct me and explain the reasoning. Anything helps at this point.

 

[tbone1004]

we have asked, it's up to @tcoen to implement. Divesoft has given us the ability for years and it is in my opinion a massive safety improvement over pure 3-cell voting logic.
Without a pressure testing head *Divesoft also has this and also tracks deviation and all sorts of other things* the unit can't really track anything with the cells other than their mV. Linearity from 0-1.0 is pretty irrelevant compared to linearity from 1.0-1.6 and without a third point for calibration above 1.0 it doesn't do a whole lot for you. No real benefit in the computer tracking this for you though so that's low on the list.

 

[mr_v]

we have asked, it's up to @tcoen to implement. Divesoft has given us the ability for years and it is in my opinion a massive safety improvement over pure 3-cell voting logic.
Without a pressure testing head *Divesoft also has this and also tracks deviation and all sorts of other things* the unit can't really track anything with the cells other than their mV. Linearity from 0-1.0 is pretty irrelevant compared to linearity from 1.0-1.6 and without a third point for calibration above 1.0 it doesn't do a whole lot for you. No real benefit in the computer tracking this for you though so that's low on the list.

Gotcha! You know, I am okay without a pressure head for now. I wonder if you can implement some support for a third point calibration. For example, we already do O2 flushes at 20ft, but the computer has no idea about it, it simply shows the result. What if we told the computer about the flush, as in "Hey, bud, I am flushing O2 at 20ft. Lemme know if mV is not what you expect based on the data you got so far!" Of course, there is some room for error there...

 

[JahJahwarrior]

I don't know if it's fair to say that tracking linearity from 0 to 1 is completely irrelevant. Is it fairer to state that a cell which is not linear from 0 to 1 is not going to be linear from 1 to 1.6, but that also, a cell which is linear from 0 to 1 is not gauranteed to be linear from 1 to 1.6?

That is, doesn't it make sense to check your cell in air, then in oxygen to confirm it is atleast linear on the surface, along with calculating your expect linearity at 1.6?

And, if a cell is 98% linear from .21 to 1, then shouldn't be expect that in a best case scenario, the cell is 98% linear from 1 to 1.6, rather than expecting it to be 100% linear from 1 to 1.6?

Additionally, is it better to check that 1.6 linearity at 20' on the way in, on deco, or both? IE, is it possible for a cell to perform linear for short time at high po2 and then fail shortly after? And should we target 1.6 or something very close to 1.6, like 1.5?

I do wish my computer would compare expected values and call out outliers, and give me the ability to manually vote out cells. I don't want to fight a rebreather that is injecting oxygen due to two bad cells; and I don't want to manually add oxygen while trying to only view one cell and ignore two others. I don't think that's an incredibly common scenario; proper gas planning should allow a bailout if it's really causing me that much hassle; and I don't know of any deaths that happened because people got confused due to bad cells and voting logic.

But I use an app to record cell voltages in air and o2 at surface, calculating linearity from .21 to 1, and predicting voltages at 1.6, so I can check that underwater. Also makes it easy to see that hey, last dive my cell started at 11.3 and was 99% linear, now my cell is suddenly at 9.5 and only 97% linear.