Atomic Aquatics Cobalt Dive Computer

[RonR]

....
Now I do like the idea od OS independant fireware updates. What about uploading dive loads to a webserver over the USB with internet provided by the host? Then a user could login to said site and download a log file in different formats for different log programs. ...Could the usb port on the Cobalt be used to gain internet access from its host? Then firmware and logs could be transmitted via internet.
Nick

There are actually several who have done web based dive logs, from the late lamented Uemis, whose dive logging software was entirely web based, to some third party projects. A couple of these have contacted us with for information on incorporating the Cobalt, which we fully support. You do have the issue with data formats (dive data coming from the computer doesn't have a file structure, that needs to be created by the program). But the more you can take the 47 different variations of Windows and all the hardware configurations out of the picture, and stick within a standard communication method, the better off you are. What you are suggesting is, for dive logging, very sound.

Firmware updating is another question, in that the program in the computer needs to be replaced. For most dive computers, either this capability does not exist, or it is very rarely used. But we wanted the Cobalt to have frequent updates and enhancements, and not just be a static product. This process is vulnerable to failure if it is interrupted, either by a glitch in the host computer, a timing or driver problem, a bad USB port, a connection break, or (as happened to one of our users) a lighting strike. We realize we need to make this process simpler- particularly for Windows users, where almost all the problems have occurred. On the Mac side, where the firmware update is a standalone program, we have had very few problems, and most of those have traced back to hardware or cables (or the Lion update, and USB changes there). One approach we are considering is to just have the Cobalt appear as a USB mass storage device (like a thumb drive) that the user could drag a firmware file to. That will take some additional development on our end.

Just to be clear, Atomic is completely responsible for their dive logging and firmware updating programs, my comments about third party dive logging software were just to point out that there are some very good, well proven, and reliable programs out there that are feature rich and download directly from the Cobalt, with full support from Atomic. Atomic did contract with an outside developer to build their dive log- which I expect most other manufacturers have done as well. I'm more open about it. You would probably be surprised to know that some dive computer companies have little or no in-house engineering capability, but contract out (I know in part because we have been approached to do work for them). That's not a bad thing- the diving market is very, very tiny, and that means they can get good embedded engineers, designers, or programmers for specific projects. But doing all development in house and the company taking full responsibility are separate issues.

Ron

 

[RonR]

A customer beta program of the download tools, desktop software could do great things for the quality perception here. not of the firmware itself, but of the install process. msomething is obviously getting past the internal testing process.


Many of us have multiple home PC's (and in my case a Windows XP, Windows 7 - 64B, MAC, iphone, ipad, android phone). OK, i am a gadget guy and willing to help on occasion.

This is actually something I have started expanding, so we can "seed" new versions to a larger group. There are so many desktop computer variations that it's difficult to test enough. I will take you up on this offer- PM me with contact info.

Thanks, Ron

 

[nfarrar]

There are actually several who have done web based dive logs, from the late lamented Uemis, whose dive logging software was entirely web based, to some third party projects. A couple of these have contacted us with for information on incorporating the Cobalt, which we fully support. You do have the issue with data formats (dive data coming from the computer doesn't have a file structure, that needs to be created by the program). But the more you can take the 47 different variations of Windows and all the hardware configurations out of the picture, and stick within a standard communication method, the better off you are. What you are suggesting is, for dive logging, very sound.

Firmware updating is another question, in that the program in the computer needs to be replaced. For most dive computers, either this capability does not exist, or it is very rarely used. But we wanted the Cobalt to have frequent updates and enhancements, and not just be a static product. This process is vulnerable to failure if it is interrupted, either by a glitch in the host computer, a timing or driver problem, a bad USB port, a connection break, or (as happened to one of our users) a lighting strike. We realize we need to make this process simpler- particularly for Windows users, where almost all the problems have occurred. On the Mac side, where the firmware update is a standalone program, we have had very few problems, and most of those have traced back to hardware or cables (or the Lion update, and USB changes there). One approach we are considering is to just have the Cobalt appear as a USB mass storage device (like a thumb drive) that the user could drag a firmware file to. That will take some additional development on our end.

Just to be clear, Atomic is completely responsible for their dive logging and firmware updating programs, my comments about third party dive logging software were just to point out that there are some very good, well proven, and reliable programs out there that are feature rich and download directly from the Cobalt, with full support from Atomic. Atomic did contract with an outside developer to build their dive log- which I expect most other manufacturers have done as well. I'm more open about it. You would probably be surprised to know that some dive computer companies have little or no in-house engineering capability, but contract out (I know in part because we have been approached to do work for them). That's not a bad thing- the diving market is very, very tiny, and that means they can get good embedded engineers, designers, or programmers for specific projects. But doing all development in house and the company taking full responsibility are separate issues.

Ron

Is there space onboard to incorporate a small TCP stack and IP support and then a tftp or othere file transfer protocal? Can the environment support it? Then with host internet access the Colbalt could access an Atomic update server to download the update and upload logs its self. That would make frequent updates even easier. The Thumb drive idea is an easy one to implement and would be reliable.

In future hardware bluetooth could solve all of this. A mini PCIe card is $5-15 wholesale and the mini pcie socket is about 50 cents to a buck. Plus it reduces the amount of needed internal connections. If you used induction charging(could be a usb based cradle or collar) you could completely seal the electronics and have no external connections. Powermat from the cellular indus. comes to mind. Then the tank pressure readings are the only connection needed to come in the sealed housing.

Just a Thought. Thanks for sparking my engineering juices. Back to the web site idea. My thought is that is an easier environment to develop in and web developers can be found much easier. Also deployment is faster. You can closely control versions. One customer wont have an outdated version of the software. Eveyone has the same version on the web server. You could also have and more closely monitor special development servers for beta testers. If it is not possible to feasible to implement tcp/ip into the firmware its a moot point.

 

[TnT]

I'm not convinced having the device upload your dives directly to some online service is a good idea. I have two good reasons for that. The first one is that it requires internet access, which is not always available on many diving locations. I usually travel with my laptop, and can easily download dives with a desktop application regardless of whether there is internet available or not. The second reason is privacy concerns. Not everyone wants to submit their personal data to some online service controlled by a company (even if they have nothing but honest intents). Even if it's not for privacy reasons, having to rely on a web service as the only means to get at your data is bad. Just look at what happened to uemis. It went bankrupt and as a result their users can download their dives anymore.

Note that I'm not saying online service are bad. I just don't think it's a good idea if it's the only option. Maybe you have already heard about diveboard? It's an online service which uses a browser plugin (for windows, linux and mac) to download dives directly from within your browser. Pretty cool, and it does support the Cobalt too. That's exactly what I meant when saying supporting third-party developers can lead to very innovative ideas.

 

[mbirnbaum]

I just ran the update to 1.17 and it went through without a hitch, however the gas used statistic is messed up. Shows stuff like 90L for something that should have around 1990 L.
On v1.13 that statistic was OK but the RMV statistic got messed up - showed me 7L/min for what should have been between 16 and 18L/min.
V1.12 was fine with all of that, but had a randomly occurring bug with gas switching that I notified atomic of through my vendor but never got an answer.

I know it's hard writing software but please run unit tests on it to make sure simple things like those bugs don't make it into the publicly deployed firmware.
The statistic bugs are easy to spot but something that occurs more dynamically like the gas switching bug isn't - and that's a far more important one, so it's all the more important to have thorough automated tests to go through as many scenarios as anyone can think of and then more as bug reports come in and get addressed.

 

[RonR]

I just ran the update to 1.17 and it went through without a hitch, however the gas used statistic is messed up. Shows stuff like 90L for something that should have around 1990 L.
On v1.13 that statistic was OK but the RMV statistic got messed up - showed me 7L/min for what should have been between 16 and 18L/min.
V1.12 was fine with all of that, but had a randomly occurring bug with gas switching that I notified atomic of through my vendor but never got an answer......

Can you give me specifics (PM is fine) on what you are seeing? Are you saying these numbers in the log changed between versions, for the same dives? If the tank size were not recorded correctly in the dive data, that could give erroneous reading for consumption, but they would remain the same once they were there. What tank size is entered in the Gas/ Tank Settings screen? The Cobalt's dive log doesn't display the tank size used, but that data is stored with the dive and can be accessed.
And please tell me what you saw with gas switching, so I can make sure it was addressed.

Ron

 

[staggsSevin]

So I got to page 36 on this thread and decided to say heck with it and just post my question. Did Atomic ever come out with a case for the Cobalt? I'm fixing to retire my Icon HD (with wireless) and move back to a console so I'll be picking two of these guys up and was trying to figure out if there was a case made just for them or see what other people are using. In the past I would just put them in my reg bag (not including the Icon, but I did so with my Luna) but with this guy I won't be doing that.

So any cases out there made just for the Cobalt?

 

[Downing]

Still no case. I've been told that Atomic's "looking into it" for what that's worth.

I use a Blue Reef neoprene generic computer cover. It's almost twice as long as it needs to be but is a reasonable fit otherwise and provides plenty of padding. Blue Reef, if you don't know, is Leisure Pro's housebrand. The price, I think it was $8, was certainly right but it's like parking a Ferrari inside a Greyhound Bus Station.

I'll gladly toss them in the trash if Atomic ever produces a case or even just a custom cover.

You might want to consider using a Zagg screen protector, too. The supplied Atomic screen protector isn't bad, but nothing beats a Zagg and you can get one specifically for the Cobalt.

 

[staggsSevin]

Still no case. I've been told that Atomic's "looking into it" for what that's worth.

I use a Blue Reef neoprene generic computer cover. It's almost twice as long as it needs to be but is a reasonable fit otherwise and provides plenty of padding. Blue Reef, if you don't know, is Leisure Pro's housebrand. The price, I think it was $8, was certainly right but it's like parking a Ferrari inside a Greyhound Bus Station.

I'll gladly toss them in the trash if Atomic ever produces a case or even just a custom cover.

You might want to consider using a Zagg screen protector, too. The supplied Atomic screen protector isn't bad, but nothing beats a Zagg and you can get one specifically for the Cobalt.

Sounds good thanks Downing.

 

[Straegen]

One approach we are considering is to just have the Cobalt appear as a USB mass storage device (like a thumb drive) that the user could drag a firmware file to. That will take some additional development on our end.

This... do this. Far simpler system and the firmware can vet the new uploaded firmware before an overwrite. The Cobalt may have this already but a lot of BIOS MBs now have a stored factory flash that can be reloaded at any time in case a flashing goes bad. Basically the store the factory flash on a completely separate chip.

This is actually something I have started expanding, so we can "seed" new versions to a larger group. There are so many desktop computer variations that it's difficult to test enough. I will take you up on this offer- PM me with contact info.

I suggest setting up a series of VMs for internal testing before sending out a beta. Our test bench has 21 OS variations including Mac and Linux as well as most Windows flavors at a variety of Service Pack levels. It is time consuming to setup, but once done pretty easy to run a regression test. Using an MSDN subscription and Oracles VirtualBox the costs are relatively cheap.