Paul: Mainstream aviation considers the gyroplane to be the flying equivalent of a one-horse open sleigh -- fun perhaps, but hopelessly obsolete. So there's been very, very little professional engineering literature about them written since the 1930's.
There's a book by one Brooks (Peter?) about the Cierva gyros (look on Amazon). It recounts the efforts of the inventor of the gyro to eliminate torque roll, once he designed out the wings and ailerons that he'd used on the earlier models. It really tells you all you need to know -- about the nature of the problem and about how to fix it. Very briefly:
Yes, we're talking about the rolling torque that is caused by the reaction between the airframe and the prop.
As long as the rotor is pulling upward on the frame with a large and steady force, it's possible for this upward pull to be aimed a little to one side and thereby cancel the rolling tendency. Problems solved? Not.
Everything's fine until the rotor's upward pull decreases. It can do so either because of strong turbulence or an intentional low-G (an arc toward earth) maneuver by the pilot. The reduction in rotor pull (thrust) then allows the prop's torque to roll the gyro, even though the pilot did not command a roll. The roll can be violent enough in extreme cases to cause a crash; it's happened several times over the years, in many different types of gyro.
Ron Herron, designer of the Little Wing tractor gyro, retraced Cierva's steps as outlined in the old Brooks book, more or less showing the rest of us how it can be done.
We should have figured this out long ago, as the original Bensen gyro with 90 HP engine would lift one wheel during a ground runup, at least if there was no one in the seat. IOW, it tried to pull a torque roll without even being in the air.