Automotive dynamometers, accuracy and repeatability.

These are two totally different things so lets separate them!

• Repeatability
  Its entirely possible to have a Dynamometer that is calibrated wrongly, but can give very accurate repeatable back to back runs!   And this is frequently the case.  It does not really matter if the actual power and torque figures are 20% out, if all you are looking for is gains in performance, or trying to iron out a fuelling problem for example.  Provided that the dyno gives truly repeatable results then there is no problem.  This requires LOTS of airflow, and very controlled conditions and so can be aided by a good "setup" and experienced operator that knows these things.
  
  
  
  And in most cases a good well designed inertial Dynamometer gives by far the most repeatable results here.  This is because they are simply a constant mass to be accelerated. This mass cannot change from one week to the next, and its rate of acceleration is timed on each revolution directly and digitally. A braked Dynamometer generally has a few things against it here.  For a start they use some kind of brake (water, electrical EMF or electro-motive-feedback, Disk brake, etc) to hold the revs steady or rising slowly.
  
  These brakes are subject to variation due to friction, resistance changes due to temperature, etc.  And the torque measurement is measured generally by a "load cell" which in turn can be both calibrated wrongly, and is subject to a degree of Hysterisis, and or a slightly "non proportional or linear" output.  In addition this analog dc voltage output is then converted to digital by an analog digital converter introducing further small errors.   From this torque reading Horsepower is calculated as Power = Torque x RPM  (Or torque x road speed = horsepower can be plotted and used but now we have no RPM scale for the graph, or engine torque reading.

• Accuracy
  Sometimes really accurate power figures are important. For example when I was was working for Bike Magazines like Performance Bikes, they wanted to see the difference accurately between last years model, and the latest one.  This maybe a 1bhp or 1 percent difference. Generally only inertial types can be truly this accurate, especially over time.  Here (inertia type) there is and can be no "calibration" or drum mass changes, whereas there are DC voltages from sensors, and many other variables on most braked dyno setups.   These in fact have recommended calibration checks by most of the big manufacturers.   But the very fact that they CAN be calibrated unlike good inertial dynos gives rise to great concern over any quoted figures!  People "like" big numbers.

• NO "Engine" or flywheel figures can ever be completely accurate for a variety of reasons, discussed here so all of this relates to REAR WHEEL measured, corrected power.
   

• Conclusion
  If you want really truly accurate, power figures that can be relied upon, an inertia dynamometer will always be superior subject to it being from a good manufacturer such as the US based Dyno-Jet outfit., or my own(!) (for a variety of reasons) and you also need to read most of the rest of this site!  They are not all created equal...

Here is an example of a bike (my own bike and dyno as it happens) running on an  inertial dynamometer. The blue line is with nitrous boost from 7400 rpm onwards.  But what's interesting is how close the two curves are before this point. Especially as one run was done in 4th gear, and the Nitrous run done in 5th gear to prevent the nitrous from causing wheel spin.  The result (Power in hp plotted against engine RPM) shows that the two traces for engine power v RPM are almost identical until the nitrous suddenly takes over.  These runs were weeks apart, with different pressures and temperatures and humidity AND in different gears and on entirely different Dynos!  One was on my customers dynamometer (one of my own "make") the other was on my mobile dynamometer at Bruntingthorpe.

I doubt that you could get this kind of consistency and repeatability from any two braked dyno systems.   This is needed to see the TRUE results of tiny changes on different days and or dynamometers.   

All content, design, images, text, information copyright DynoPower Dynamometer Systems & John C Williamson