Automotive engineering firm Ricardo and failure analysis specialists Axiom have helped identify a crankshaft problem causing engine failures for one of the world’s biggest car manufacturers.

The unnamed OEM appointed Ricardo to investigate why crankshafts were failing across one of its engine families, and Ricardo brought in Axiom to consult on the issue. Stockton-based Axiom sectioned the broken crankshafts to examine the actual machined radius profiles of the journals, which can be a factor in the initiation of fatigue issues.
“We examined the extent of the cracking, charted the position of the failures and the morphology of the cracks,” explained Axiom co-founder and metallurgist Dr Rene Hoyle.
“The crankshafts were sectioned so our experts could study the actual crack face. In some cases we will actually ‘break open’ a crack and examine it under a high-powered microscope; the peaks and troughs of the crack are like an open book to experts in failure analysis. Once the extent and type of failure was plotted it was a relatively simple matter to identify the cause, which in this case was due to metal fatigue.”
Although the engines were already operating on the road, the problem had not been identified during pre-production testing. Instead, it was discovered during on-going endurance tests carried out by Ricardo. Following extensive running on a test bench, the engines were examined and the crankshaft problem was revealed.
“No one wants to suffer a catastrophic engine failure, such as a snapped crankshaft, in the outside lane of the M1,” said Hoyle.
“Manufacturers go to great lengths to make sure their latest designs are suitable for the intended application, but if things go wrong and cracking is found, they need to understand what is causing the problem.”
While the engines in question were already fitted to thousands of vehicles, Axiom says the early identification of the problem has potentially saved the manufacturer tens of millions of pounds in recall and compensation costs.
The Jowett Javelin crankshaft suffered a similar problem. The cause was torsional vibrations from the Flat Four configuration, which would break the crank when sustaining a speed of~ 80MPH, about 4,700 RPM. Laystall solved it by its oval web design and careful radiussing. In the R4 Jupiter the re-equipped engine would rev to 8,000RPM without damage. Porche offerings in the same class were trounced by the R4 when they met in international competition.
No crankshaft and for that matter no big ends, conrods, pistons, valves, valve springs, collets, timing chain, oil, oil filter, radiator.etc etc to wear out or fail in an electric car motor which has essentially one moving part and battery warranties are now 8 years from some OEMs and there is a also a growing after-use market for EV batteries for static energy storage. Not that EVs are immune from recalls. Mine has been recalled twice in four years for an Engine Control Unit change, not that I have noticed any difference before or after the change.
8 years is not a lot for a conventional car engine. And the cost of replacing the batteries is more than that for fitting a reconditioned conventional engine, which will last for double the time.
The electric motor of an electric car with very few moving parts can be expected to last much longer and cost far less in maintenance over an eight year period than an internal combustion engine. No oil, air and fuel filters and lubricating oil to be changed at regular intervals, new exhaust system, new timing chain etc adds up to quite a sum over eight years.
The battery cost of an EV should be compared more to the fuel cost of a conventional ICE vehicle. My EV with a 16 kWh battery costs less than £1 for over 50 miles of motoring when charged on cheap rate power at night and nothing when charged on a sunny day from the PV panels on my roof.
For say 10,000 miles a year the cost to recharge my EV over an eight year period at £1 for 50 miles is £1600. For a petrol or diesel car doing say 50 miles per gallon at say £5 per gallon covering 10,000 miles over 8 years the cost is ££8000 so the saving in fuel cost is £6400 in favour of the electric vehicle.
There is a growing market for used EV batteries for use in domestic and industrial energy storage schemes so the old battery will have some value at the end of 8 years while the fuel consumed by the ICE vehicle will have been burnt and lost forever and will have a negative value if the effect of tailpipe pollution on the health of people breathing these fumes is taken into account.
As with computers and other mass produced electronic items it can be expected that the cost of batteries for electric vehicles will fall over time which may not be the case for petrol and diesel fuels although time will tell.
Batteries on electric cars are water cooled, in the event of the water jacket leaking this can cause a fire or even an explosion . A mechanical failure in an I.C engine results in a stopped vehicle.
Good call. Nice to know our chariots are being thoroughly checked!
Spontaneously combusting mobile phones released onto the market however, shows not all manufacturers so diligently catch problems prior to release. Recalls & subsequent litigation could easily bankrupt some companies.
amazing why the fault never shew in prototypes-I would have thought that at least one engine would be tested to destruction or near–hard enough to have such a fault show up. Surely no one drives a car anywhere near to manufacturers prototypes?
I think the producere for testing of the crankshaft needs major overhaul .
Yusuf
Bit worrying that cards seem to be driven harder on the road than the makers engine testing procedures
Amazing, fatigue and torsional vibration of crankshafts has long been know about and both classic and CAE analysis methodologies are abundant and accurate, providing verifiable solutions. They would have been applied on this design, surely. So what was different about this mode of failure that was not predictable with those tools? That would surely be worth knowing.
The ‘unnamed’ manuufacturer needs to be named. It is in everyone’s interest otherwise unexpected failures will occur. By failing to take action the manufacturer will significantly increase their liability costs to include subsequent damages, not only that but hiding their name will ultimately lead to majjor loss of credibility. Being open and honest about problems improves public perception. And of there is a dangerous design defect they are doing nothing to restore public confidence!
I believe action has been taken to address the problem Nick, and no one is in danger. The decision to leave the OEM unnamed is not that of The Engineer, and we ourselves don’t even know who it is.
Appreciated that The Engineer didn’t do that, but customer confidence is not helped by anonymity! There is a short term hit because of a failure, but in the long term their credibility is enhanced. And how many people are running around with engines likely to fail? They have a right to know that something they have bought is not perhaps as reliable as they were led to believe. Though I recall many years ago (early 70s) I had a mini that had the same fault and while it was noisy it still ran after a fashion.
All R & D Programmes must have one key verb “Validate” through rigorous testing.
This is vital for British industry in this competitive Global marketplace
I would liked to have seen drgs/details of where the cracks started an propagated, and what radius-increasing measures were adopted. Fatigue testing of aircraft structures started in the 1960s after the Comet failure, and I was involved with this work at an aircraft manufacturer – Handley Page Ltd – for a few years. Also did some photoelastic work. Very interesting work it was, and amazingly informative. I then moved to the nuclear industry for the rest of my working life.
A failure in the validation process could be as simple as a difference in machining between prototypes and later processes, leading up to full-on production. A marginal design that was machined favorably for fatigue in prototypes could have passed all the design analysis & testing. A later, higher-volume machining process which gave a very slightly different profile or finish to a critical radius could still have “met print” and put the marginal design over the edge and into unacceptable stresses for fatigue life. Of course, it’s the limit-seeking drivers of the world that will find weak points in the execution of a marginal design.
If this failure mode wasn’t observed during the development testing, which in my experience would have included several engines being tested “too destruction”, despite several commenters here observations, then the reason it wasn’t picked up is likely due to a quality variation. I once was involved in the development of a front torsion bar that passed all the development testing on prototypes but failed during testing of off tool samples from the factory. The fatigue failure was as a result of variations in the reducing atmosphere of the heat treatment oven, between that used to make the prototypes and that used to make volume production parts at the factory. The variation in the oven’s atmosphere led to a very thin decarburised layer on the surface (less than the historically specified minimum), so no amount of shot peening was going to improve the residual compressive stress in the surface beyond the fatigue limit. Machining radii like the one described can be a challenge, as a slight undercut (like microns) at the tangent of the radius is essentially enough to create a stress concentration and reduce the fatigue life significantly. Subsequent journal grinding then has to be sufficient to eliminate any undercut. These things are very near the limits of what can be achieved and quality control is critical. SInce in this case the development testing is said to have validated the design, the failed parts obviously didn’t match the design in some aspect. Either that or the tested prototypes weren’t at the worst case allowable tolerance, which is why several engines would have been tested under very arduous conditions, exceeding those expected to be experienced by actual engines in cars in the field. This is a rare case where the historically derived testing regime, which is meant to ensure some “design margin” didn’t quite account for the huge non-linearities in fatigue life that can occur with slight surface imperfections. I’m surprised it doesn’t happen more often given the commercial imperatives to make things cheaper and lighter all the time.
This subject is not “news”… Cranckshaft fatigue failures were observed, analyzed and corrected as far back as the 50’s, at least in aviation engines. The improper radiusing togethre with improper surface treatment (case hardening) caused many failures of this type. And they didn’t have very “advanced” tools and methods. That present day manufacturers suffer this kind of problem now, points out the lack of engineering culture of young engine designers. For interesting case studies on Crankshaft failures see the many Accident reports that contain extremely detailed analysis on this subject. (Automotive “designers” should read them too).
All comments with respect to the sensitivity of parts to minute changes are correct.
What surprises me is the fact that a major manufacturer did not have a data bank where all those aspects which increase the risk of fatigue failure were recorded in order to verify every new design and avoid such situations which could cost the future of a company. The future is unknown but we know the past and we can use it to make better prognosis for the future.
In my days as Risk Engineer with the old Rover Group, I would often sit with the road test drivers in their canteen. Their frequent comment, especially after an uneventful drive in one prototype or another, was that they didn’t really understand why they were employed because Joe Public would find a way of breaking the car that no one had ever thought of…
For me, it sorely highlights the overall incompetence of young, present day designers, managers and CEOs… That a damn Crankshaft, that is an item produced by the millions, since loooong ago, STILL HAS PROBLEMS is almost unbelievable! This points out the important need to better educate engineers. By Gosh!, my students at the University get a hard reprimand whenever they fail with something we have already studied and left behind…
There may be another factor at work. With increased recycling of alloyed steel scrap the more difficult it must become to ensure the precise metallurgical characteristics of alloyed steels. Perhaps minor changes in the metallurgical composition of the crankshaft material lead to slight differences in strength resulting in weaknesses during high tolerance machining. Has any research been conducted in this area?
After spending most of my working life in testing and test facilities for proptype and production vehicle drivelines (now aerospace ) I still get increasingly frustrated by the attitude ‘no need to test , we can simulate it all on a p.c. model ‘
This is unfortunately the x box generation of designers way of thinking !