The question I originally set out to answer was "Are the servos strong enough so as not to result in control surface slowing while moving for a snap roll". My goal was to measure the servo current draw during this condition and use that info to study the speed of the control surface on a bench rig.

The elevator / ail snap set-up measured was a DR step function, stick activated, with minimal AFR limiting and no expo - pretty simply on/off (switch safety) max travel, which then transitions to normal behavior after the initial "hit". Initial observations were that current was high on the elv and "normal" on everything else. Knife edge loops were then flown to observe max rudder current draw which was relatively low. Measurements during normal F3A sequences were considerably lower, in fact, lower than I expected.

The same servo was then mounted on a rig and connected to linkage geometry / control surface dimensions as the airplane. The same current sensor was installed and the servo driven from the airplane RX using the same TX program used in flight. A variety of springs were attached to the "control surface" until the current signal / amplitude matched that observed in the airplane.

Another rig fitted with linkage / pot so as to provide a voltage output was connected to the "trailing edge" of the "control surface". The voltage output from this rig was fed into a PC stripchart (poor man's storage scope) which provided a means to measure travel vs. rate and can be expanded sufficiently to measure / observe the time from start / stop of movement. With this it was apparent that the rate slowed noticeably with the elevator servo and not with the other servos during the snap inputs. Changing the elevator servo to one 20% or so more powerful eliminated the slowdown. That's really the info I was interested in.

Jim, as I recall the current draws measured in "normal" flight maneuvers were consistent with Ola's data. The rig draws were higher, as the load was set for max observed flight current. I'll send you the flight current files, I don't recall the details (been a year or so). Most of the rig current set-up was done with real time observations that I didn't record. When I get a chance I'll post some of the chart recorder files.

Also interesting to set the rig up for "normal" loads and explore the different starting / speed characteristics of different servo types - but that's another story