With new thicker solid mu-metal magnetic shields (2 layers), the clock transition in our 88Sr+ experiments shows a transform-limited linewidth of ca 10 Hz using a 72ms probe pulse. More testing to follow, as we try to lengthen the probe-pulse and search for the ultimate linewidth limit.
Farnell is now selling a dirt-cheap (28 euros or so) Multicomp fiber optic power meter, so here's a quick comparison against other power meters in the lab.
It looks like the 28 euro meter performs good enough for most tasks in the lab. The Multicomp only has a few fixed wavelength settings, while the Thorlabs is adjustable in 1nm steps.