OzTek 2017

Dirk Hohndel dirk at hohndel.org
Sun Jan 22 08:10:17 PST 2017


> On Jan 22, 2017, at 2:32 AM, Anton Lundin <glance at acc.umu.se> wrote:
>> Shearwater we have great relationships with (ok, they still owe me a
>> response on more details of the new BTLE communication, but hey, they
>> already sent me a Perdix AI). We used to have a contact at Mares, but he
>> left. Do they still produce new dive computers?
> 
> I would love to get some feedback from them about their calibration
> values of their O2 sensors. I've reverse-engineered a patch which
> produces what looks to be the right values, but it would be great if
> someone with internal knowledge about how its thought could comment on my
> analysis.
> 
> ( The short and curly: Shearwater stores their voted/computed/averaged
> ppO2, and the mV from the actual sensors in each sample. They also store
> a "calibration value" in the dive header of unknown unit. If I take that
> * about 1000 (1024?) the average of the sensors will produce the same
> ppO2 as they have stored in my sample data. Is that bogus math or is it
> the right<tm> way? )

Can you write this up with a bit more context and maybe the pertinent code
sample and send it to me? I'll be happy to poke them about this.

>> No idea about Divesoft.
> 
> We already support the Divesoft Freedom-series computers. We talked to
> them over email, and they sent us a header-file with the actual structs
> they write to their log-files, so that way I could correct all the
> mistakes I/we did when reverse-engineering the format.
> ( Things like Temperature in 1/10 C, 10 bit signed wasn't that easy... )
> 
> Divesoft makes a really good helium analyser to. Way better than Analox
> variant.
> 
> A itch to scratch is to make all those file-based importers behave more
> like a regular dc-download, IE point them to a directory, and they will
> import all the dives that you're missing. This goes for the Divesoft,
> Seabear and probably a couple more.

That would indeed be easier for the user. We could trigger this through
the selection of dive computer vendor.

>> I know that people keep hoping for support for Liquivision support. Not
>> sure if there's any hope to move forward on that...
> 
> Direct download would be cool, but we atleast got a direct importer
> nowadays in core/liquivision.c. A quick glance says it looks reverse
> enginnered, but hey, we got one. 

I keep forgetting about that - maybe in part because we get a request
for direct download support maybe once a quarter.

/D


More information about the subsurface mailing list