Things to know about power measurements on a test bench

We have been dealing with this topic for more than 25 years now. One thing is clear from the outset: A tuning or engine repair store that deals with the tuning of motorcycle engines must have a performance test bench as basic equipment. How else are performance-enhancing measures to be developed and results to be verified? Power and torque figures should be measured and provable data, not optimistic estimates.
The starting point is the engine power of the production motorcycle and this is where the misjudgement begins. On the one hand, the manufacturers' performance specifications are often not achieved, and on the other hand, different test rigs often measure different performances. And what power is specified at all? The rear wheel power, the clutch power or the crankshaft power? A lot of room for uncertainty.
As a rule, roller dynamometers are used to determine the rear wheel power. By rolling out the motorcycle, the power and torque measurement is followed by the drag power, which is then added to the rear wheel power to give the clutch power. The engine power can only be measured directly at the crankshaft, which can also only be determined on a pure engine test bench. For this purpose, a disassembled engine is adapted and measured on the same. A procedure that is usually only used by manufacturers. In a BMW boxer engine, the clutch sits directly on the crankshaft; in four-cylinder engines, at most, the primary drive is still interposed. The loss is not high and is assigned a factor of 2%.

We have illustrated here measurements of BMW 2-valve boxer engines in series production condition. It is not so easy to find original vehicles in a condition that is suitable for a reference data determination. But in general we can say: All R80 models have a clutch power of about 45 hp and all R100 models about 53 hp. If you now add the 2% to the crankshaft, BMW is close to its factory specifications.

However, this does not apply to the old Twinshock models. Especially the R90S, R100S, R100T, R100CS, etc., which were all specified with power between 65hp and 72hp, do not reach more than the normal R100, so also only about 55hp at the clutch.
As a rule, it is normal for the performance measured on our dyno to be lower than the manufacturer's specification. This is also the case with today's vehicles and was no different almost 30 years ago when we bought our first dynamometer. We have always used the same manufacturer and it may be that other dynos show higher values. For us, the pure numerical value is not decisive, it is more about the comparison of the measurements with each other. Only in this way can the effectiveness of our measures be reliably recognised and proven.