Although it was originally an issue with the tolerance of the nylon washer in the locking nut of the swivel, this apparently was only one issue of this design. I'll not fault M&J for the design of the swivel, frankly, Joe is a great machinist. The primary problem is that the Divator uses a finger tight fitting from the hose to the threads on the second stage. I don't want to sound like I'm now blaming this on Interspiro, they specifically don't recommend these. With the swivel being a right angle fitting, it's prone to backing off of the threads when swivel doesn't move freely. The action of looking up and down causes the locking nut to back off. With the system being pressurized, this acts somewhat as a ratchet if there is anything that causes the nut to bind with the swivel it will literally "wrench" off of the mask. Things that could cause this binding may be mineral deposits, salts or debris as well as the original tight nylon washer. The units that caused the initial voluntary recall in 2005 were found to wrench off in a very short order.
There are 90 degree elbows being made that would have the same problems, although, I've not heard of any specifically related to the elbows, the mechanics would be the same. I heard of a few failures of the hemispheres separating on the swivels, but these appear to be related to improper disassembly and reassembly.
Some have tightened the swivels with the use of tools. This can damage the regulator if improperly done. It also means that the mask cannot be easily removed from the regulator assembly. For some, the solution to this has been to put a quick disconnect to the swivel. Now you're stacking additional points of possible failures, adapters on top of adapters. In my opinion, this is an extremely dangerous way to set up your primary system. Especially with some QD's that do not positively lock, not to say anything of the potential for gas restriction. Add a little depth and this issue is compounded exponentially.
I use to use a swivel and liked the way it allowed for streamlining of my system. Since these issues arose, I went back to a standard hose. It's clean, simple and not nearly as prone to failure.
OTS insisted on the original voluntary recall when the problem was finally realized not to be "user error". For several weeks after, ALL customers that ever bought a swivel from OTS were contacted by phone, e-mail, fax or snail mail. During this, it was discovered that there were incidents outside the dates of the recall (01/04 to 02/05). The manufacturer did not want to expand the dates of the recall and as a result, OTS stopped distributing the product.
What amazes me, is that with these failures, I haven't heard of any divers being injured or worse. I think this reflects on the skill of the divers that use this equipment. Keeping your system simple is by far the easiest and safest way to resolve this issue. The simplest solution to this problem is to go back to the standard hose. The standard hose works great and provides for little opportunity of failure. We've been lucky. We know of this possible failure point, yet...
John
Bridgediver, I scratch my head... Regardless, I am wondering if you can provide some additional background information. On the swivel is a stamped date of manufacture. I am curious to learn if you know which swivel came disconected from the mask and the date it was manufactured. 
