we have chosen not to make it adjustable because in our view this is no longer appropriate. You should be able to purchase a power meter that is true and precise and remains so.
Let me compare to cars: 30 years ago you could fiddle with the carburetor of your car. Nowadays you can't access the electronic fuel injection of your engine because there is no need to - it works as it should without fiddling and adjustment.
It will take a while for the mindset to shift and expectations to change.
To complete the analogy - modern OBD (on-board-diagnostics) systems will automatically alert you (via a check engine light) if EFI is malfunctioning or there is a cylinder misfire or some other malfunction that needs attention. You can also plug in a scanner tool into your car's OBDII port and it will run a diagnostic check and display any error codes that were thrown by your car. This is standard across all modern cars produced within the past few decades and anyone can do it. There's no mechanical expertise needed - you literally just plug the scanner it into the port, push a few buttons and that's it. Where I live, you can go to your local Auto Zone or O'Reilly (auto parts shop) and ask to use their scanner. Or you can buy your own.
If your check engine light is not on and a scan tool indicates no error codes, that means your car is operating normally within the acceptable range of error defined by whoever came up with the OBD standard. If your check engine light is on, you can still probably drive your car (one of my previous cars was missing some key emissions control equipment so the check engine light was always on). But it's good to have a standard way to tell if any car is "out of spec".
For power meters - the standard method that everyone has always used is a static torque test. You can take a known, verified weight (I use a 20kg NIST-certified weight purchased for about $130 from McMasters-Carr, plus a 114g metal hanging tool that allows me to easily hang it on the crank spindle) and calculate the expected torque given: the total force that you're applying (in my case - 20kg + 114g), the force of gravity (in everyone's case - approx. 9.8 m/s^2), and the crank length. Apply this force to the power meter (by hanging the weight from the crank arm or the chainrings). Then measure the difference between the actual torque applied and the torque reported by the power meter and determine the % error. Very simple, and this is a method you can use for any power meter to ensure that all of your devices are within specification.
Regarding your note about BB resistance affecting the results of a static torque test:
Bottom brackets do provide some resistance, but the resistance is so small that it doesn't affect the validity of the test. This same topic came up in March or April on the Google Wattage forum (https://groups.google.com/...aI7Yp1o/otyfC_KmB9cJ
). Someone did a test and found that, with the chain detached, a 10g weight applied to the pedals was sufficient to overcome any resistance from the bottom bracket and move the crank arm forward. With a 20kg weight, this is very insignificant (0.05% of the total weight used in the test). 0.05% might even be within the +/- range of error for an NIST-certified weight. So unless your bottom bracket is completely seized or has some other serious problems that prevent the crank from turning, BB resistance just isn't a significant factor in the static torque test.
In the dynamic situation (moving drivetrain), Friction Facts has found that the bottom bracket uses up about 0.4 to 0.6% of the total input force. Again - not a significant factor. There is a recent ST thread ("Best Bottom Bracket Ever") which touches upon this. In other words - the best bottom bracket ever will only net you about 0.5% extra watts.
In any case, if you were hypothetically building a power meter that promised accuracy to 0.01% and wanted to eliminate bottom bracket resistance - you could rig up a static torque test to bypass the bottom bracket / crank arm and hang the weight directly from the chain. Below is a photo of SRM's calibration rig in their Colorado Springs facility. In the same thread on Google Wattage (linked above), someone did just this (hung the weight directly from the chain) and received identical results compared to hanging weight from the crank arm.
Please understand that I'm not trying to unnecessarily bust your chops, Nicolas. All of us do appreciate you responding directly to our concerns and requests in a prompt manner. This is very refreshing and is above-and-beyond what other competitors provide.
But I did want to just state the case for why static torque analysis (or rather - any standard, repeatable testing procedure) should be used to verify claims of accuracy for any power meter. In the past, users who have done static torque analysis on even SRM units have found that their factory calibration is out-of-spec.
Trying to recall the phrase that RChung and Tom A. use all the time - "trust but verify"? If you can accurately verify, then you don't even need to trust. Well - other than trusting that your certified weight actually weighs what it's supposed to weigh.