The situation was removing a ball bearing from a shaft using a PosiLock puller.  The puller has three jaws that grab the outer race of the bearing and a screw that pushes against the shaft. The inner race is fixed to the shaft as far as rotation is concerned.  I was using an impact driver to turn the puller's screw.

The odd thing that happened was that as the screw turned clockwise (from my perspective) the outer race and rest of the puller turned counter clockwise at about the same rate.  A free body diagram would show no force acting in the counter clockwise direction except for the fixed shaft and the world.  I have had the puller turn along with the screw before, which requires something to hold it in place, but this was rather weird in that it seemed to be helping  instead of hindering.

What the heck was going on? Where was the counterclockwise force coming from?

The screw shaft?

Impact drivers use rotary motion inside them.  They use a screw principle to convert impact to torque.    So there IS rotational energy in the system.

Of course, otherwise they wouldn't work.  But, where is the counter rotation coming from?

When the screw moves clockwise, the rest (which includes the jaws, frame, nut and bearing) all rotate the other direction.  Most often, the rest will move in the same direction but at a slower rate, and frequently has to be restrained otherwise no effective work is done.

The puller jaws or maybe even the bearing race are uneven adding a "ramp" in the opposite direction.

The driver acts in both directions.   Thump, recoil, thump, recoil.   So maybe the recoil impulse was adding the backwards momentum?

Puller screw flexes shaft clockwise too fast for the friction in the bearings to overcome the static inertia of the puller body, arms, and outer race. Between impacts, the shaft returns to its previous position at a rate that allows the friction in the bearings to accelerate the puller body, arms, and outer race counterclockwise.

Most likely what is happening is the screw is not only rotating, but being twisted by impact wrench's very high initial torque, and the end of the shaft is turning less than that point which goes through the puller body.  As the impact wrench's torque is removed, the shaft wants to twist back to its zero torque point, and vibration and/or movement of the bearing reduces the compression on the shaft enough to "break" the friction holding the jaws in place on the race, allowing the (now "un-twisting) shaft to carry the body of the puller along with it, in the opposite direction of the original input torque.

Mike

Maybe epicyclic fretting precession.

harmonic resonance caused the counter rotation.  i have seen this happen before, if you used a different impact wrench the phenomena you witnessed would not of accrued.

The balls in the bearing acted like planetary gears.

Ever noticed what direction a tire spins as you torque down the lug nuts?

The friction forces are not in the direction you think.

That sounds more plausible but I'd like to see a diagram.

Magnets