Overall I feel the benefits gained from having an always on lighting system that will not run out of batteries outweighs any loss. But I ride to get around and not for sport or exercise. As my bike alone is already ~30 kg, I don’t worry about added weight.

If this test did aim to same light brightness in all cases, and light output was measured with light meter, then it actually DID measure efficiency. If it didn’t measure light brightness then it didn’t give info on efficiency. But simple resistor don’t emit light! Current and voltage information together would tell power (watts) IF it was measured. Was it? If not measured then output is a mystery.

You’re right, the output power comparison would be interesting. I guess the reason they didn’t test it was that output power is typically VERY close to 3W on most of the hubs (1.5W on the Shimano 1N70). You can boost the output through electronic circuitry, but it of course increases the resistance at the hub. This comparison uses the same light on multiple hubs at varying speeds.

With regards to the usage of the word ‘efficiency’, my terminology across the site tends not to quite align with engineering terminology. This is to make it easier to read and understand for people with a less technical background. In this case, it’s the power difference between pushing the pedals and the light output. Less pedalling effort to achieve the same light output being the more efficient.

Thanks. Good to hear. The warranty replacement looks like a new generation product, so I trust it will last much longer. I haven’t used it since the rebuild. Still using the OEM wheel that came on my Surly Trucker. Much less need for a dynamo to power lights and recharge USB devices when not doing multi-day tours.

I’ve had a PV-8 and PD-8 with premature failed bearings. Strangely the replacements have been good to date. I’m guessing bearing or assembly problems. With the noticeable drag you experienced could it be your bearings, whilst slowly deteriorating, were increasing drag subtly without being aware of it?

I am a keen cyclist and use a SON 28. I also used to work as an electronics engineer. So if I may be so bold as to offer a suggestion:

Most dynamo lighting these days use LEDs which are normally with switched regulators; important otherwise excessive speed would result in overvoltage and blown LEDs.

By performing your experiment using a simple 12 Ohm resistor as a load you will get slightly higher power loss than you would with a switched regulator which would limit the power going to the LED in an efficient way.

So above 20kph they would probably be less power consumed resulting in less drag.

Cheers