O2 sensor values on CCR dives

[Jef Driesen]

On 2017-11-27 17:27, Anton Lundin wrote:
> On 27 November, 2017 - Davide DB wrote:
>> On 17 November 2017 at 21:59, Davide DB <dbdavide at gmail.com> wrote:
>>> Dctool is a lucky process. 99% of the times I get errors like this:
>>> 
>>> [73.192176] ERROR: Failed to receive the response packet. [in
>>> ../../source/src/shearwater_common.c:319 
>>> (shearwater_common_transfer)]
>>> [73.192715] ERROR: Failed to download the dive. [in
>>> ../../source/src/shearwater_petrel.c:266
>>> (shearwater_petrel_device_foreach)]
>>> [73.193107] ERROR: Error downloading the dives. [in
>>> ../../source/examples/dctool_download.c:218 (download)]
>>> [73.193712] INFO: Write: size=9, data=FF0105002E902000C0

On windows (and linux too), I already have a build of dctool available 
which uses native bluetooth communication. This should be a bit more 
reliable then the serial emulation. (On the Petrel, I've noticed that a 
low battery also causes download issues, probably because the bluetooth 
chip draws more power.) The bluetooth build is available here:

http://libdivecomputer.org/builds/experimental/windows/dctool-bluetooth.exe

Instead of the serial port, you need to pass the bluetooth mac address 
as a hexadecimal number (e.g. prefix with 0x, and remove the colons).

> I haven't gotten around to it. Jef might be interested in looking at it
> to, so I'll ping him.

Do we have some ground truth to compare against? Last time I checked 
Shearwater Desktop didn't show the ppO2 values. And if they do, do they 
show the average ppO2 or the individual sensors?

The average ppO2 is stored directly in the data, so that shouldn't match 
perfectly. But the individual sensors is another story. In the data 
there is only the millivolt, and libdivecomputer converts this into a 
ppO2 value based on the calibration values. This conversion is reverse 
engineered, so a mistake there is not impossible. Note that 
libdivecomputer no longer reports the average ppO2, only the individual 
sensors.

Is the resulting ppO2 wrong for the entire dive, or only during the deco 
stop? The calibration does a simple multiplication with a constant 
value, so if there is something wrong there, I would expect to see the 
same kind of error throughout the entire dive.

Jef