A large-wheeled construction truck lost braking power AND went careering down a steep slope. The operator stayed with the truck, but was killed when it tipped over. I was called by a lawyer-client who handled the truck manufacturer's insurance work in the West Virginia area.
Scene of the Crime
A 5/8-inch OD by 0.050-inch wall thickness brake tubing fractured through very near a mounting. The line was designed to withstand 550 psi in what was called an "air over oil" braking system. I am not sure which fluid was carried by the tubing in question.
My study had to be strictly non-destructive, which precluded taking of sections for metallographic study of the microstructure and hardness testing and for chemical analysis. My study was limited to the unaided eye and hand lens, radiography and Scanning Electron Microscopy (SEM).
Radiography showed no cracks in the line, other than the fracture. Study with the naked eye and low-power hand lens revealed only a rough fracture surface. The SEM was the key to the study.
Manufacturer literature described the tubing variously as being resistance-welded steel, silver-soldered, brazed steel and double-wall, copper-jelly-roll, brazed steel. Study in the SEM revealed a pure copper layer and no sign of silver solder. The tubing resembled a thin layer cake that is spread with jelly, rolled and sliced. Each slice has attractive spirals of white cake and colored jelly. The steel of the tubing had been copper plated, then heated to melt the copper and braze the layers of steel together. If sectioned, the tubing resembled the dessert with the copper playing the adhesive role of the jelly.
Failure of the line had catastrophic consequences, so the chances for such failure should be minimized. Why jelly roll? I here cite "Pelloux's law*" on fracture which states, "it always breaks where it's welded." A single small region of defective resistance weld could result in failure whereas the copper braze in jelly roll would have to fail in a continuous path all the way round the circumference of the tubing to give a leak. The jelly roll seems the safer choice of brake tubing.
The chief strength of the SEM in this case was its depth of focus — the ability to study irregular fracture surfaces at high magnifications. The SEM could also analyze the X-rays induced in the sample by the bombarding electrons and thereby give a rough idea of the composition of the small region of the surface under study. The smaller fractured piece was about the size of the end joint on my thumb, and could be studied without cutting in the relatively primitive SEMs of 30 years ago.
The SEM study showed abundant fatigue striations, as seen in Figure A (about 2,000 times magnification). The failure was due to repeated stresses from vibration and pressurization. But why did the line fatigue? Faulty design or manufacture of the line should have given a bunch of similar failures. Such was not the case. There was no sign of damage to the line as might occur during operation under difficult conditions. Nor did the truck have an exceptional number of hours.
The SEM study showed the steel to have a very high concentration of non-metallic inclusions, the angular particles seen in Figure B. Some non-metallic inclusions are accepted as being part of the steel. These particles are mostly oxides formed during de-oxidation of the molten steel with either silicon or aluminum. The particles are very brittle and tend to fracture during rolling or drawing. The cluster of inclusions seen in the micrograph is probably the fragments of an earlier, large inclusion. There was a large cluster in the region where the fatigue crack started. The brittle inclusions are often cracked, which in turn helps to nucleate fatigue cracks in the steel, leading to early failure.
Why would a major manufacturer known for quality machinery use such poor quality steel in a critical application? I do know that a price freeze back in the 1970s made certain grades of steel unavailable. The company could have been desperate and turned to a low-quality steel to meet production quotas.
*Prof. R. M. N. Pelloux, MIT.