Since this subject comes up on various wing boards at about the same frequency that SUV-hate stories come up on the 6 o'clock news, I thought I would try to figure out exactly how mine works and why I find the turn signals acceptable.
First, just like "Motorcycle Consumer News", I'll detail the procedures and the high-tech equipment:
The timer was a 1000 one, 1000 two... silent counting system that was calibrated before the ride with a wall clock.
The distance traveled was computed by the odometer as it turned a new 1/10th mile number as a start for the turn signal time test, thus giving a high level of confidence in the concurrent recording of both time and distance in lock sync with the 1000 one, 1000 two clock.
The weather was 60°F with a light breeze. Traffic was light, but generally unpleasant during the constant speed test with turn signals flashing while no turns were made. Several senior citizens were giving a thumbs' up to a bike with an apparent lingering turn signal, while young men in pick-up trucks were using another finger to show their support.
The results:
There appears to be a inverse time to speed relationship with the Honda designed self canceling system where it stays on longer at lower speeds. The relationship of distance to speed was not linear, nor necessarily in sync with the time relationship. Overall, the Honda self-canceller was seen to be adequate for a high percentage of riders. However, in cases where a rider would put the signal on at a high speed and then slow down significantly before the curve or intersection stop, that rider would find that the signal expired before the turn. More about this later*.
Test numbers (note, all speedo & odometer numbers are in "Honda Miles" so please allow for a 6% error rate):
60 mph; 7 seconds; 615 feet
30-35 mph; 9-10 seconds; 400-450 feet
stopped at light/stop sign; unlimited time; zero distance
Other thoughts:
The system seems to work nicely for riding where you are changing lanes or giving turn warnings at "normal" advance distances such as one block or less at city speeds or 1/10th of a mile at cruising speeds.
*In those cases where it would seem to be to prudent to give an advance turning signal, such as when riding up-front in a large group or when auto/truck traffic is frenetic, one might have to move their thumb all the way to the signal button and push it yet a second time (a little exercise is a good thing).
To this tester's surprise, there was no difference in results if you pushed the turn signal button all the way over to feel the click, or if you just pushed it to the clicker resistance while under way. This was not checked when stationary except last night in the garage when it would not activate the signal unless it "clicked". I couldn't get my signals to ever click off in less than 7 seconds, as indicated in the Owner's Manual, but the 110 yard metric seems to have no role unless it is the slow speed time determinent.