We are getting more and more mails asking for accuracy and validation and as well stressing the point, that power meters are much more accurate that NIRS .
So I decided to add some more information to this discussion. The beauty of this discussion is, that more and more power users start to do some thinking on what may go on with NIRS, It is always a time process to learn to integrate new ideas and accept , that we can make progress by looking at optimal changes.
Here for all this questions to independent answers on the discussion on accuracy.
Bike Power Meter Accuracy
The mobile bike power measurement systems from products like SRM and Power Tap use strain gages to measure torque or twisting force on the crank arm or on the rear hub. Torque is multiplied by axle RPM to determine rider power. Strain gage technology and accuracy are discussed on an internet site2. Strain gages are subject to both zero drift and span drift. To understand these two types of drift, think of a bathroom scale. Zero drift is the failure of the scale to read zero when you get off. Span drift is an incorrect reading of weight when you step on the scale. Both forms of drift, zero and span, result from change in strain gage properties with temperature and with aging of the glue used to attach the gage element to the measurement point. Strain gage systems can be calibrated using weights to produce a known force on the bike pedals. There are potential errors in weight calibration because the force can both twist the measurement element and also bend it. Strain gage signals due to twist or torque is data and any due to bending is an error. Reported accuracy in terms of mean error scores for SRM and Power Tap factory calibration over a range of 50 - 1000 W were 2.3 +/- 4.9% and -2.5 +/- 0.5%, respectively3. Accuracy for SRM and PT was not largely influenced by time and cadence; however, power output readings were noticeably influenced by temperature (5.2% for SRM and 8.4% for PT). During field trials, SRM average and max power were 4.8% and 7.3% lower, respectively, compared with PT. Calibration and strain gage errors are also discussed in reference 4, which reports also a comparison of SRM, Power Tap and Polar mobile bike power measurement systems. This article4 suggests checking zero on each ride and checking span calibration at frequent intervals.
The CompuTrainer system uses the bike rear wheel to drive a copper flywheel, spinning in the field of an electromagnet. The accuracy depends on knowing the rolling drag of the bike wheel driving the flywheel and the accuracy of calibration of the drag versus rpm versus magnet electrical current. The rolling drag is determined by a calibration procedure from the rate of slowing of the known mass flywheel at a given force on the friction roller and determined by user test. The drag generated by the electromagnet on the spinning copper disk depends only on the electrical properties of copper, RPM and the intensity of the magnetic field. The electrical and magnetic properties of copper are predictable functions of temperature and compensated for in software. The drag versus RPM versus current are constant because the geometry of the electromagnet and the location of the copper flywheel are unchanging. The drag versus RPM versus current were initially calibrated during development of this product with literally thousands of measurements for an accuracy of better than +/-2.5%.
Near Infrared Spectroscopy (NIRS)
C. Dean Kurth MD
Cincinnati Children’s Hospital
Professor of Anesthesia and Pediatrics
University of Cincinnati College of Medicine
Determination of the accuracy of the NIRS devices has been problematic. In order to determine accuracy, the device must be compared with a gold standard. Because there is no gold standard for NIRS (ie, no other device measures O2 saturation in the tissue circulation), determination of accuracy remains an estimate. There is currently one FDA approved device, made by Somanetics. Its accuracy on the FDA application was compared relative to a weighted average (SwO2) of arterial and jugular bulb O2 saturation. In adults and children, the device is not that accurate (+10-15%) on an absolute level of oxygenation (rSO2 vs SwO2), but is fairly accurate (+5%) on a change in oxygenation ( rSO2 vs SwO2). In other words, the device indicates a change in oxygenation accurately but does not indicate accurately what the oxygenation actually is.
Other devices purport greater accuracy than the Somanetics device. However, these devices are not FDA approved, not commercially available, and have not been subject to accuracy testing on large scale. Nevertheless, several of these devices have been tested in animal models and have been found to be accurate +3% on an absolute.