Does anyone have any rule of thumb values for the gearbox efficiency ratings in JVN? We use a good number of VersaPlanetary in both single and two stage but I cannot find a reference the efficiency ratings for these.

According to Wikipedia* each planetary stage is about 97% efficient. I’d maybe decrease that to 95% or a bit lower since the VPs are modular and self-assembled.

So the efficiency of the whole gearbox will be 0.95^n, where n is the number of stages.

Great find! I actually went and read the paper in the references. I do not always trust Wikipedia so I have to verify. While the paper makes a generalization, I have a bit of trust in the UC Berkley system!

Now, as an extension of this question, is there any documentation of the efficiencies for other FRC gearboxes?

For spur gear meshes, a typical rule of thumb on efficiency is 1% loss per interaction.

I’m not buying that paper, there is zero support for that number and they actually claim that planetary gear sets are an efficient method of power transmission. If you look at the NASA work on the subject of spur gear interfaces it supports that ~1% per interface number that Tom mentions. It is important to note that their work assumes the shafts are supported by ball bearings and the lubrication is by oil jet and the gears not in an oil bath, so there is no churning loss.

Now lets apply that to a typical planetary gear train used in FRC. Depending on the ratio you have 3 or 4 planets which means 6 or 8 spur gear interfaces, or 6-8% loss due to the gears. However those planets are not supported by ball bearings they are a bushing interface which will not be as efficient. We also don’t have the luxury of a oil jet for lubrication, we use grease which is more viscous, furthermore that grease is trapped in system so with excessive lubrication the churning loss can increase significantly.

Personally I use 10% for each stage of a planetary gear set which is probably actually on the low side for the typical FRC planetary gear train and how it is used.

Edit Looking at the Mead work that g2tkcf posted they come up with a 75%-85% effeciency at a 30-1 ratio which would include at least 2 stages, or ~10% per stage.