October 12, 2022
I was trained in the cult of the integer face contact ratio. A few adherents to this belief are still above ground, but the introduction of personal computers has made it difficult to attract new recruits. We were not quite “flat-earthers” — but close. It started off simply enough: if you wanted true helical action, you needed a face contact ratio of more than 1.00. If 1.00 was the minimum, 2.00 ought to be even better. And if 2.00 was the goal, why not select helix angles that got you exactly 2? Or if your helix angle was fixed by the cutting machine, change the face width to get that magic “2.” Lengthy technical papers were written supporting this philosophy. Elaborate charts were prepared to aid our company’s design teams. So, imagine my shock when I left that firm and discovered that the industry was full of heretics who just used seemingly random helix angles; and their gears worked just fine. Not that I did not encounter other strongly held beliefs which, upon later examination, were not truly necessary. Among the “commandments” I heard about were: 1. Helix angles over 20° must be avoided. 2. Helix angles must be integer values like 8°; 10°; 12°; 16°; 18°; 20°; 23°; or 30°. 3. Face contact ratios over 2 are bad. 4. Face contact ratios under 1.0 are nothing but distorted spur gears. Each of these commandments originally had some basis in science; most likely some well-intentioned student either misunderstood, misremembered, or took something out of context and had it carved in granite so they wouldn’t forget it. Actually, they inscribed it in the company’s computer code and future designers did not know their output was compromised. Once the gear trade became international and computer programs were commercially available, it became impossible to ignore the biases trained into your designers. Many things our system said should be avoided worked well in competitors’ machines. With regard to helix angles, I learned that most values will function just fine provided the mounting system is sufficient.