RonR
Contributor
Well, I go away for a few days and look what happens… Seriously, not sure about weighing in here, but will add a few comments from my perspective.
First, I completely sympathize with free_electron's frustration, he's had two failures that should not have happened, plus the uncertainty generated by the recall. I wouldn't call the recall a failure, as it was a response to a very tiny number of computers in the field that appeared to have been improperly assembled. Atomic, out of a proper abundance of caution, decided to check all computers that had been manufactured prior to their instituting the new assembly process that made this problem impossible. I will add a couple of reassurances. The little ball/ glue seal in the battery compartment is actually a very good design- not high tech, but simple and rock solid reliable. And we have taken many of these batteries and subjected them to saltwater immersion, dead shorts, destructive abuse, etc. and have only come up with a weak fizzle- that's what the overpressure ball would release, if needed. These are not batteries likely to burn.
But I want to also make several points.
Diving is a tiny, specialized market. Trying to compare diving equipment manufacture to the standards set by manufacturers of electronics (or other) products where many millions are sold is just not going to be productive. The hard economic facts are that the resources available to those manufacturers- including pre production testing of large numbers of products to uncover rare failure modes- are not remotely possible for diving products. This is a market where sales are measured in the thousands.
Once a manufacturer has made essentially the same design (a small, round puck, for instance) for years, they will have largely worked out the bugs in manufacture, and reliability will probably be good. When new designs are developed, despite best efforts and good design, there are sometimes issues that turn up in the field that just did not appear in the sample numbers used for testing. That did happen with the Cobalt, and from nearly 30 years involvement in this business I can tell you it's happened to some extent with every new product I've been either part of or aware of. This isn't because manufacturers want to use consumers as testers, it's because of the small size of the diving market precludes the kind of testing that mass-consumer level products get. And building production is never the same as building prototypes. The large majority of early Cobalt adopters have not had any problems at all (and therefore don't post them here), but some have, sometimes repeatedly. Forums like this one, and Atomic's willingness to be open about any failures, tend to amplify those issues and make it seem as if they are much more common than they are.
As developers, we want to see two things when problems do happen:
First,we want to see Atomic being responsive to any consumer problems, and doing everything possible to make things right for the user ASAP. That isn't always 100% possible, but Atomic has, in my view, done an outstanding job of customer support. Compare Atomic's openness and responsiveness to other manufacturers- I'm pretty confident you would find they are in the top tier. I do agree that if they are out of new stock at the moment (and I know they are waiting on boards now, sales have exceeded anticipations), they should have contacted free_electron and sent a demo/ replacement before now.
BUT… the other thing we want to see, and something that is not visible to users, is that any problems occurring in production or assembly are aggressively identified and fixed. Even rare failures have to be treated as unacceptable. Atomic has done a tremendous job here- quality of their products is something they take very seriously. Problems have been researched, identified, and fixed. The Cobalts being produced now are not the same as our first production a couple of years ago. Atomic has aggressively acted to change their assembly process and designs to prevent even minor or rare failures. When I look back at logs of problems from early production, I see overwhelmingly issues that could no longer occur with these changed processes.
Nothing made by humans is perfect. The LP sensor failure free_electron experienced is in a component used by many dive computer manufacturers. They fail very rarely, but it does happen to all of us. Some percentage of Cobalts will leak or fail- though I don't believe it will be any more than other computers, and now quite probably less. Windows compatibility will still give us trouble. We will keep working on ways to make upgrading firmware easier and less dependent on the desktop OS. We will keep working to make this product better and more reliable- and being very open when we fall short is a big part of this effort.
Ron
First, I completely sympathize with free_electron's frustration, he's had two failures that should not have happened, plus the uncertainty generated by the recall. I wouldn't call the recall a failure, as it was a response to a very tiny number of computers in the field that appeared to have been improperly assembled. Atomic, out of a proper abundance of caution, decided to check all computers that had been manufactured prior to their instituting the new assembly process that made this problem impossible. I will add a couple of reassurances. The little ball/ glue seal in the battery compartment is actually a very good design- not high tech, but simple and rock solid reliable. And we have taken many of these batteries and subjected them to saltwater immersion, dead shorts, destructive abuse, etc. and have only come up with a weak fizzle- that's what the overpressure ball would release, if needed. These are not batteries likely to burn.
But I want to also make several points.
Diving is a tiny, specialized market. Trying to compare diving equipment manufacture to the standards set by manufacturers of electronics (or other) products where many millions are sold is just not going to be productive. The hard economic facts are that the resources available to those manufacturers- including pre production testing of large numbers of products to uncover rare failure modes- are not remotely possible for diving products. This is a market where sales are measured in the thousands.
Once a manufacturer has made essentially the same design (a small, round puck, for instance) for years, they will have largely worked out the bugs in manufacture, and reliability will probably be good. When new designs are developed, despite best efforts and good design, there are sometimes issues that turn up in the field that just did not appear in the sample numbers used for testing. That did happen with the Cobalt, and from nearly 30 years involvement in this business I can tell you it's happened to some extent with every new product I've been either part of or aware of. This isn't because manufacturers want to use consumers as testers, it's because of the small size of the diving market precludes the kind of testing that mass-consumer level products get. And building production is never the same as building prototypes. The large majority of early Cobalt adopters have not had any problems at all (and therefore don't post them here), but some have, sometimes repeatedly. Forums like this one, and Atomic's willingness to be open about any failures, tend to amplify those issues and make it seem as if they are much more common than they are.
As developers, we want to see two things when problems do happen:
First,we want to see Atomic being responsive to any consumer problems, and doing everything possible to make things right for the user ASAP. That isn't always 100% possible, but Atomic has, in my view, done an outstanding job of customer support. Compare Atomic's openness and responsiveness to other manufacturers- I'm pretty confident you would find they are in the top tier. I do agree that if they are out of new stock at the moment (and I know they are waiting on boards now, sales have exceeded anticipations), they should have contacted free_electron and sent a demo/ replacement before now.
BUT… the other thing we want to see, and something that is not visible to users, is that any problems occurring in production or assembly are aggressively identified and fixed. Even rare failures have to be treated as unacceptable. Atomic has done a tremendous job here- quality of their products is something they take very seriously. Problems have been researched, identified, and fixed. The Cobalts being produced now are not the same as our first production a couple of years ago. Atomic has aggressively acted to change their assembly process and designs to prevent even minor or rare failures. When I look back at logs of problems from early production, I see overwhelmingly issues that could no longer occur with these changed processes.
Nothing made by humans is perfect. The LP sensor failure free_electron experienced is in a component used by many dive computer manufacturers. They fail very rarely, but it does happen to all of us. Some percentage of Cobalts will leak or fail- though I don't believe it will be any more than other computers, and now quite probably less. Windows compatibility will still give us trouble. We will keep working on ways to make upgrading firmware easier and less dependent on the desktop OS. We will keep working to make this product better and more reliable- and being very open when we fall short is a big part of this effort.
Ron
