October 5, 2023
Regular readers know of my concerns about the over reliance on commercial gear design software. Some of what you are about to hear may sound familiar. If so, I have done my job. Gear design is much easier these days thanks to the millions of hours that other people, around the world, have invested in codifying the rating standards and technical papers other folks have developed, tested, and published. You can input a minimum amount of information, wait a few seconds, and look at five or six pages of results, a beautiful 3-D image, and even a file to print out an actual part. But is it the right design for your application? You do not know until you thoroughly review that multi-page printout and assure yourself that the detail geometry meets your standards. Back when we wrote our own rating programs, you had to immerse yourself in the standard. Each formula had to be tested to prevent a typo or a misinterpretation from distorting the results. Often, the “old family recipe” got cooked into the code without anyone being the wiser. Your program would give error messages over things like a helix angle being over 20 degrees, for example, while the standard carried no such prohibition. This was apparent when AGMA published its own software to aid in calculating ISO ratings for gears. We were in the midst of designing a new product line and had delusions of global distribution, so we bought a copy and signed up for the training sessions; we even followed the pre-training guidance and made sure we understood the program by creating math models for all our gear sets. The training session itself would have been anticlimactic but for my sitting between two bright young engineers from a former employer and the “chief” American gear engineer for an offshore competitor. Both tables had their problems with the software. The “chief” could not even find “rack offset” on his own drawings because they called it something different. Even more confused were my young friends, as they had no idea that their “old family recipe” started with a transverse tooth form, an unusual pressure angle, and helix angle calculations that were based upon adding a partial tooth to the total so the pinion would be stronger. For both companies, you just entered the number of teeth and the center distances into the “black box” and used the output without further review. Sorry folks — filling in data entry forms does not make you a gear designer! We all hated writing computer code, but a great side effect was gaining a deep understanding of how one variable was related to another and how they combined to do good or bad things to your part. No matter how pretty the screenshot, your work is not done until you study those variables.