Aluminium Cylinder Failures
8th Sept 2009

On Sunday 16th August 2009, a SCUBA cylinder ruptured causing serious injuries (loss of right hand and compound fracture of lower right leg) to the person who was filling the cylinder. The cylinder was 88cuft aluminium serial number #H2225 manufactured to Australian Standard AS1777 by CIG in 1983/84 and was in current test. The cylinder had been tested yearly as required by Australian Standards and this cylinder had been annually tested for at least the last 5 years.

The cylinder had a working pressure (WP) 22.4 mpa (3248.84 psi) and a test pressure (TP) 35.2mpa (5105.328 psi). A number of SCUBA cylinders that had been filled during the same time were checked and were gauged at 220 bar (3190.83 psi). The compressor a REVAL model 5409 had a blow off (safety valve) fitted and set at 285-290 bar (4133.575 – 4206 psi) I had the shop owner reroute the air delivery lines to a blank fitting and then he operated the compressor until the safety valve activated which showed 4100psi (282.68 bar) on the gauge on the compressor.

I believe that the cylinder could be over filled by a person not taking notice of the pressure showing on the fill panel gauge, however this pressure is below the test pressure of this cylinder, and based on the fact that 10 other cylinders in this same group of cylinders all showed a pressure of 220 bar (3190.83 psi) I have no reason to believe that this cylinder was in fact over filled.

The cylinder when it ruptured exploded into four pieces and the valve. In two of the pieces there was evidence of cracks through the neck and threaded area of the cylinder.

There are records available showing a number of aluminium cylinders manufactured world wide prior to 1990 from 6351-T6 alloys have ruptured, in some cases with serious consequences.

SCUBA COM P/L - SAFETY POLICY / PRE 1990 Aluminium SCUBA Cylinders

Due to the latest terrible accident (as above) and a number of failures in Australia and worldwide of scuba cylinders made from 6351-T6 alloy (pre 1990), Scuba Com P/L will no longer fill or test these cylinders.

I have been a cylinder test station manager for 30 years and have been increasingly concerned about the safety of these cylinders for the following reasons.

1. When cracks started to appear in these cylinders many years ago test stations were instructed to take special care to inspect the shoulder and thread areas of the cylinder with a double magnification mirror. These cylinders often had deep folds in the shoulder area which made it impossible to detect fine cracks there. At that time if the crack was not able to be seen it was not thought to be a problem.
2. As these cracks became more of a problem, a system of "eddy current" which involves inspecting the thread area every two years, was introduced. This new test was very good at detecting cracks in the thread and as a result many of these cylinders were condemned and withdrawn from service. That was great if the crack had progressed to the thread area but the shoulder area was still only being checked with a mirror. As stated previously it was very difficult to detect fine cracks in the shoulder area so many cylinders stayed in service that could still have a number of fine cracks in the deep folds inside the shoulder area.
3. Considering cylinders being tested may still have undetected fine cracks in the shoulder the cylinder then must be pressure tested to around five thirds of the filling pressure. Is it possible that testing the cylinder to near five thirds of the fill pressure may actually make the fine cracks worse or cause a new crack to start. Considering this the cylinder is then passed according to standards, filled and given to the client to use.
4. The facts, as I understand, are aluminium cylinder manufactured pre 1990 from 6351-T6 alloy have had many problems over the years with cracking in the neck and shoulder areas. Cylinders manufactured after 1990 from the new alloy have had no problems at all and from my experience of testing thousands of aluminium cylinders have never seen a crack in the post 1990 cylinders.
5. These questionable cylinders are now nearly in the 20 to 30 year old bracket. My main concern is that due to their age, is it possible that metal fatigue may now be increasing the chance of cracking and now causing a higher risk of catastrophic failure.
6. We are well aware of the cracking problems with these cylinders so how many times can we fill and empty (expanding and contracting the metal) them before they pass their “use by” date?

Identification of the suspect cylinders as below:
1. CIG/Luxfer cylinders marked with AS1777 manufactured between 1975 and 1990 , where the serial number does not start with an S.
2. Luxfer cylinders marked DOT3AL or DOTE6498 manufactured between 1972 and 1988.

Could you imagine the devastation if one of these cylinders failed in the back of your car while driving to a dive site. I would ask safety conscious divers to not risk themselves or family and remove these cylinders from service.
Divers beware of buying the suspect cylinders on the web or from second hand dealers because I have heard of people selling them in this manner.
Also beware of using these cylinder while doing dive courses and check before hiring them.

Do not hesitate to contact us for further information or bring your cylinders in for identification.
There are no problems with aluminium cylinders manufactured after 1990, we will test and fill those without any hesitation.