In Sept 2006 I paid $9,000 to a manufacturer (Bettany Gears Limited, Paraparaumu, New Zealand) to make 2 dog gear sets for my 1990 Ford Laser GTX (same running gear as Mazda 323 GTX).
After installation, the gear set failed after about 2km of half-throttle use. The failure occurred because first and third gears simultaneously engaged. This was directly due to inadequate clearances between gears because the manufacture did not take in to account the typical gearbox deflection under load. Subsequently other OEM (Original Equipment Manufacturer, Mazda in this instance case) parts in the gearbox were also damaged.
During investigation into the cause of failure, Mechanical inspection showed supporting evidence of dual engagement, and Metallurgical testing was also conducted and showed unprecedented damage (burning) of the surface of the gears incurred during manufacture.
The manufacturer, despite assurances before the failure that I would not have any trouble with their gear set, took a denial stance after failure. Even though the errors existed they did accept they were theirs nor did they take responsibility to fix them. They incorrectly blamed "housing flex" for the failure and arrogantly advised me to get another manufacturer to fix the problem by making the remaining drive gears spur gears also. Later they wrote "in hindsight" spur gears should not have been used. That was in fact a choice they made.
After about 3 months from failure I chose to employ a lawyer to seek reimbursement because it was clear that the manufacturer was not entertaining my request for reimbursement nor offering any assistance.
11 months after purchase I accepted an out-of court offer of settlement from the manufacturers solicitor. This however did not cover any of the $6,000 of damages and recovery costs that were additional to the purchase price.
Discussion
For those of you who are unaware: housing flex will always occur, because metals obey basic stress/strain engineering laws, whereby a load(stress) will induce a displacement (strain). To assume this will not occur is unrealistic and naive. The manufacturer was provided with a load criteria but chose not to apply it.
At the time of purchase I knew little about gear technology so naively and regrettably relied on this manufacturers’ experience and advice.
The manufacturer undertook design and fitting of gears into my supplied housing. I supplied them with all necessary OEM parts during the design and build process. My request to them was to make stronger gears than OEM Mazda parts, for an applied load of 400Nm.
They also offered to make 6 forward speeds. I said no to this because the 6th gear would be on the unsupported end of each shaft.
The Manufacturer chose to make spur gears (evidenced in their quote), whilst retaining the OEM helical drive pinion. They also did not allow sufficient running clearances between their dog-engagement hubs (only about 1mm). These 2 factors caused the dual engagement failure within the first 2 km of light operation: 1st gear was forced to engage whilst power was applied in 3rd gear. This was identified by manufacturers Metallurgist upon mechanical investigation of the failed gearbox. Evidence shows actual shaft displacement was about 1.5mm under the failure load.
Pre-failure problems were also evident with the gear set as supplied by manufacturer:
1. The gear set remained in stuck in second gear - to fix this I had modify the shaft selector.
2. The end of the input shaft was too large for the flywheel bearing - I fixed this by removing 0.1mm from the shaft.
3. The OEM reverse gear synchro rubbed inside the end cover (due to the size and location of the 5th driven gear) - to fix this I had to remove and lathe 1mm from it.
Metallurgical investigation showed micro structural damage to all the gears due to the manufacturers faulty in-house heat treatment (a process they have yet to disclose despite repeated requests). My metallurgist (and another in Auckland) advised me not to use the second gear set because of the damage the orientation of impurities in the material. Interestingly the manufacturers Metallurgist claimed the damage was acceptable because “it was there” (i.e. existed) and in his opinion (contradictory to supporting literature) this was typical of a carburized steel. I have never seen a report from him to support his conclusions.
The damage to the gears was inter-granular cracking typically to depths of 20 to 30 microns (again by some undisclosed process). In metallurgical literature, cracking to a depth of 10microns in carburized steel is described as "severe" and has significant effect on fatigue life and wear.
The second (unused) gear set has been returned to the manufacturer. I suspect they will try to sell it to another unsuspecting buyer.
The manufacturer never offered any documented design evidence to support their assurance the gears were strong enough and I would not have any trouble due to the burned surface. The manufacturer also indicated they do not believe there is any benefit in shot preening – however shotpeening I have been told by my Metallurgist is a proven technology available since the 1920's.
I have since purchased a quality made gear set from Albans Off-road gear in Australia. They supplied helical gears with sufficient design clearances. They were also sub-zeroed and shot preened. Another user has had this gear set in an identical car also running 400Nm for >3000km without any trouble or inspected damage.
