I agree whole heartedly agree that chassis dynos are not the best piece of equipment for measuring HP at the crank, that's what an engine dyno is for. However if you can do static speed runs, then the power you're measuring will be fairly close to crank HP because the loss of power you get accelerating all the moving parts, in the drive train, has been taken out of the equation. Our dyno can do what is called a step test, you're probably familiar with these, but I'll explain it for everyone else's sake. During a step test the operator holds the bike at a constant throttle position, usually 100%, the dyno will then hold the bike at a specified RPM with the load cell, then release the load and allow the bike to climb to the next specified RPM, where it will again hold the bike at that RPM. During these periods where the load cell is holding the bike at a constant RPM there is no acceleration happening, thus there is no power lost to forcing the drive train components to increase in speed. There is still some mechanical loss, since no system can be perfectly frictionless, but the numbers being logged will be fairly close to crank HP. An added benefit to this type of testing is an increased accuracy in the way power is being measured. The torque is being directly measured from the strain gauge attached to the load cell and HP can be directly calculated from the rotational speed of the drum; it takes 1 hp to raise 550lbs 1ft in 1 second, so when you know the weight of the drum you're spinning the equation to calculate HP at a static speed is pretty straight forward. Our dyno measures actual real life HP and torque, through the use of the strain gauge on the load cell and a highly accurate glas-on-glass encoder speed sensor on the drum axle. While our dyno has the capability to do inertia only runs, we always tune and do our dyno runs with the load cell engaged and operating in a speed squared function. meaning it exponentially applies a load to the drum as speed increases, this actually mimics wind drag. We shy away from inertial only runs because they are inherently less accurate in the way power is calculated, and when tuning there is not an accurate real world load being applied to the motor. Additionally our dyno has expansion modules that allow you to read any array of sensors you wanted to, oil temp, coolant temp, boost, secondary throttle plates, multiple O2 sensors, etc. If you want to measure something, our dyno can measure it, and measure it more accurately than anything else on the market.
I like that idea of using a servo to control the throttle, if you ever get that working let me know.