Chuck Schultz is a licensed engineer, Gear Technology Technical Editor, and Chief Engineer for Beyta Gear Service. He has written the "Gear Talk with Chuck" blog for Gear Technology since 2014.
Not to worry — no politics here. The “backlash” I want to talk about is the “lost motion” in gear systems. It sometimes seems to rile people up almost as much as politics and I have long been puzzled about the misunderstandings that persist.
At one time, AGMA actually included “backlash grades” in the quality standard, and this may be the root cause for confusing “low backlash” with “precision.”
While a clock or other timing mechanism needs a minimum of lost motion, the amount of backlash in a power transmission gear is not a reflection of its precision. It is entirely possible for a very high-quality-ground tooth gear to have a wide variation in tooth thickness — a key variable in set backlash.
For unidirectional service, where rotation and power flow do not change, backlash is only “active” upon start-up. As long as the teeth are still thick enough to carry the load and the spacing accuracy meets specification, “extra backlash” is not going to cause a failure.
Absence of sufficient backlash, on the other hand, can destroy a gearbox in very short order. The forces generated in a tight mesh, or a mesh with debris going through it, are incredibly high. There has to be enough “open space” to account for runout, spacing errors, center distance tolerance, and temperature changes.
So how much backlash is enough and how much is too much? It depends on tooth size and application.
Process capability will determine the range of tooth thickness for a particular part, but the designer has to establish a lower limit. You can calculate all the variables that could be involved in determining the actual operating center distance. Or, you could work from a “rule of thumb” and vary it with tooth size. I use .030 inches divided by the normal diametrial pitch and have yet to encounter a “tight mesh.” For an upper limit I specify 1.5 times the lower limit. No matter how small the teeth get, I do not go lower than .0015" for minimum backlash.
I do not remember ever seeing this information in a magazine article, but you could easily check in the Gear Technology archives.