Atomic Aquatics Cobalt 2: unsafe at any depth?

[damdiver89074]

Atomic Aquatics Cobalt 2: unsafe at any depth?

I purchased TWO Cobalt 2 computers on Feb 25th.
My wife was learning SCUBA diving so our first two dives with the computers were in a thirteen foot deep pool the following week.
At the end of the two dives neither unit had logged the first dive. Out of the water, Unit 1 was reading a current depth of 8 feet with a max depth of 24 feet. Unit two was reading a current depth of 4 feet with a max depth of 18 feet. The two units were sent to Atomic Aquatics for replacement.
I received two new computers but was unable to dive with either new unit. One unit was reading that it was 4 feet underwater and the second unit had an inoperable compass.


On April 30th I emailed Atomic Aquatics and requested to speak to someone from the company. I was hoping for an apology for getting four out of four failed units, but would have been happy if they had simply acknowledged my concerns. As of today, June 13th, I have never heard from a company representative.
The designer of the computer (not an Atomic Aquatics employee) did contact me and acknowledged that they were aware that a problem exists with the pressure sensor but that the magnitude, extent, scope, and predictability of the problem was unknown. The problem was know, but Atomic Aquatics was selling their computers anyway.


That second set of failed units was returned May 1st and I was promised a new set of hand tested computers as soon as possible. Up until today, six weeks later, I was still waiting. Today I have decided to try and get my money back from my original February purchase. Yesterday my dealer spoke with someone at Atomic who said they should be able to send me out working units in a week or so but that they still looking into the problem and they still don't know the magnitude, extent, scope, and predictability of the problem. They are still selling the computers though.

 

[Dr. Lecter]

RonR, as you may recall I was a pretty staunch defender of the Cobalt 1 a couple years back, when people were losing their minds over annoyances like rusty keypads, complicated or unavailable firmware/updaters/logging programs, and a number of user-caused battery issues. Then the HP connection recall happened, and I still though Atomic handled it well. It eventually seemed like AA had a handle on the Cobalt's reputation as a great computer with constant issues related to materials selection and/or assembly. That said, for a $1,200 computer, there did seem to be a lot of hiccups.

Personally, I sent back the same Cobalt 1 three times:

• First for a saltwater destroyed battery compartment due to a tail of glue carelessly left across an o-ring sealing surface, which very slowly let water into the battery compartment;
• Next, two of the buttons finally exhibited signs of magnet rusting, when they became non-responsive and it needed the HP recall check/cover replacement;
• Finally, shortly after coming back all ready to dive, it went back for a complete warranty replacement due to its implosion at 190' in the middle of a deco dive. Thankfully all this resulted in was using my new Petrel, and me having a brand new Cobalt to sell to someone else so I could buy a second Petrel.

Each time I received flawless easy service by calling AA directly, getting someone from the factory floor on the phone right away, and very speedy (2-4 days on the first two, a week or two on the last one) FedEx return service, completely gratis. RonR was also extremely helpful here -- without him, I think SB would have turned the Cobalt 1 into something resembling the xDeep BLACK's disaster of a rollout.

But now this (I think this is about the 6th separate post I've seen about this sensor issue) with the Cobalt 2 rollout (plus a new crop of firmware issues)? It may be time for AA to sell/license the tech to someone else and get back to doing what they apparently do best: regs and masks.

Is there any evidence other DCs have gotten these bad sensors, since they're so widely used? Or did AA just totally luck out and get the whole bad batch?

 

[Scared Silly]

FWIW, It seems that Suunto had a series of computers with bad pressure sensors. I do not know the similarities and differences in the sensors.

 

[Dr. Lecter]

Yeah, it's not unheard of for this kind of issue to crop up: which has been true of each and every issue that's cropped up with the Cobalts. But after awhile, I think 'totally normal consumer electronics SNAFUs' fatigue sets in, and given the relative charlie-foxtrot that the Cobalt 1 was, it looks like there's a pretty thin veneer of goodwill for the Cobalt 2's growing pains.

But there's no doubt the Cobalt is one of the best recreational computers on the market, and AA still has very good (I've gotten the sense things may have become slightly more corporate since the Huish buyout) customer service/support. Hopefully that will be enough to staunch this mess.

 

[damdiver89074]

I agree that the Cobalt is a great computer on paper. But in my case 4 failures in 4 computers is still a 100% fail rate even if it is a small sample size. All of these computers passed their in house QA and testing then failed immediately after. The bigger concern is that if the sensors are to blame, then when exactly will they fail? I don't know of any mechanical failures that improve over time. If someone was lucky enough to get that 1 in 5 computer that actually worked out of the box, would it fail on the first dive, the third dive or the tenth? They look great at the dealers, but I'd rather have something I can depend on.

As far as AA customer service goes: I'm still waiting since April for my email, phone call or whatever. My dealer has spoken with them on the phone several times, at times right in front of me. On the last call we were told they'd try and get us a working pair in a week or two even though I'd purchased them at the end of February! That's extremely poor customer service.

 

[RonR]

OK, deep breath, and in keeping with our policy of openness…

First, let me correct one thing. As soon as Atomic became aware that there was a problem with depth sensors, one that they were not catching in testing, they stopped shipping Cobalt 2's completely. I at one point thought they were still shipping units that passed multiple tests, since it was clear that most Cobalt 2's were working fine, but they stopped cold until they knew what the problem was and could ensure that it wouldn't worsen or suddenly appear in use. That turned out to take a while.

At the time the OP and I had a discussion, we had seen fewer than 20 instances of the issue- then he got four in a row. It turned out to be a detective nightmare involving the Swiss manufacturer of the sensor and a number of engineers to track down what was responsible. We now know, but just as of the last couple of days. It has taken us this long to be sure we understand the issue and confirm through production testing that it is no longer present- and as you can imagine it's not great to be unable to ship your flagship product in the height of the season.

The identical depth sensor in question is used on many- perhaps most- dive computers, and has been for many years, to the tune of many tens (hundreds?) of thousands. Suunto, Shearwater, Uwatec, the Cobalt 1, and others all use or have used this same sensor- so it is not a common failure point. And as it turns out it wasn't exactly a bad batch of sensors, though it really looked like that at the beginning.

The way the problem manifested was that the Cobalt would pass all testing at Atomic, including the wet test dive, then- only after exposure to water- start reading inaccurately. Not by much, but potentially enough to make it not come out of a dive on surfacing, which means being off by three feet or more. Atomic didn't catch this initially because the wet test is almost the last thing they do. The difficult part was that the sensors were perfectly accurate before a wet test, and if left to sit for some time- days- would appear to fix themselves and work properly on the next dive. It was both intermittent and self repairing. That also kept the sensor manufacturer from diagnosing the issue, as by the time they got to Switzerland they worked fine.

Without getting too deep into how we figured this out: The center of these sensors is filled with a very delicate gel material that keeps water off the electronics while transmitting pressure freely. We traced the problem to a protective cap the manufacturer recently began supplying to help prevent damage to the sensor during assembly. This cap could, in some situations, release material that would combine with the gel- but that only reacted when exposed to water. This reaction apparently slightly pressurized the gel on water exposure and caused the sensor to read inaccurately. Most sensors were slightly affected or not at all, but some, like the ones the OP received, could go off by several feet. We now believe this variation was a function of how long these caps were in place during assembly- so naturally they tend to manifest in clusters that were manufactured and then shipped together.

Needless to say, we are not using these caps any longer, and I'm notifying other computer manufacturers who I know use this sensor so they don't have the same issue. On computers that don't provide a way to check atmospheric pressure reading it could be difficult to detect.

The part of all this is that the teething problems in the Cobalt one- buttons, firmware updating, compass calibration, etc. really were well addressed in the Cobalt 2. This issue hit out of the blue, and is completely unrelated to the design or anything Atomic did.

I don't have any explanation for the lack of a reply from Atomic, except to say that it is not typical for them, based on everything I have heard for many years. True, they have not had an answer to provide until just now, but it's still important to say something, even if just "we don't know".

If anyone has a question about the depth sensor on their Cobalt 2, compare the atmospheric pressure reading on the System Info page with actual local barometric pressure- it should be within about 20 mbar. Check atmospheric pressure readings before and after a dive- they shouldn't vary by more than about 20 mbar. For perspective, 33 mbar = 1 foot of sea water.

From what I understand, Cobalt 2 shipping with new sensors will resume next week, with priority going to those who need replacements. new production has been given the go-ahead. In defense of whoever your dealer talked to, I believe this decision was just made today.

I'm sorry I was not able to be more definitive about this earlier, but this was a really difficult problem to figure out.

Ron

 

[Dr. Lecter]

As far as AA customer service goes: I'm still waiting since April for my email, phone call or whatever. My dealer has spoken with them on the phone several times, at times right in front of me. On the last call we were told they'd try and get us a working pair in a week or two even though I'd purchased them at the end of February! That's extremely poor customer service.

Why would you work with a dealer unless forced to? Step one: call AA. Step two: your problem will be fixed.

 

[buddhasummer]

FWIW, It seems that Suunto had a series of computers with bad pressure sensors. I do not know the similarities and differences in the sensors.

Same thing happened with the Heinrichs Weikamp Frog and OSTC3. They stopped production of the Frog, recalled all OSTC3 and took almost a year to sort out. It does happen it seems.

 

[staggsSevin]

Hmm makes me wonder if they will end up issuing a recall on the cobalt 2. We got one for my wife but it hasn't been used yet. (I still use my cobalt 1)

Sent from my SAMSUNG-SGH-I337 using Tapatalk

 

[RonR]

Hmm makes me wonder if they will end up issuing a recall on the cobalt 2. We got one for my wife but it hasn't been used yet. (I still use my cobalt 1)

I don't know, but I'd consider that unlikely, as 1) if the problem manifests in a dive it makes the computer slightly more conservative, and 2) it is easily detected using the Cobalt's built in diagnostic features. The System Info page is designed to monitor the sensors. Major manifestations like what the OP experienced are of course very obvious.


A new firmware version, which should be posted to the Atomic site soon, contains additional "under the hood" monitoring software that can check the depth sensor if there is any question. For anyone with a Cobalt 2, I would just recommend checking before and after a dive as in my post above. If the sensor is out of spec, contact Atomic directly- they are setting up a mechanism to rapidly swap out problem sensors. I expect Atomic will make an effort to contact Cobalt 2 purchasers of the potentially affected computers and let them know how to check for sensor accuracy.

Ron