Instrument

Some problems were reported with the radial-velocity stability of the spectrograph when using the medium-resolution fiber (RMS$\sim$30 m/s) when compared to measurements done with the medium-resolution fiber on the old fiber bundle (RMS$\sim$10 m/s). Tests were made placing a radial velocity standard at different positions at the fiber entrance, with different zenith distances and with different rotator distance. No conclusive answer was found why the fiber in the current bundle seems to behave different than the fiber in the old bundle. However, it is clear that some care should be taken to center the target in the fiber as deviations of up to 50 m/s can be expected if not centered properly, while it also is advisable for the stability of the radial velocity to always observe with the rotator at the same angle so the relative position of the fiber with respect to the light coming from the secondary is the $\sim$same.

Similar observations with the high-resolution fiber did not show an increase in spread, giving similar RMS values to those obtained with the medium- and high-resolution fibers on the old bundle. It also seems that observers that use the medium-resolution fiber with simultaneous ThAr exposures do not see an increase in the spread of the data compared to the old bundle.

We have been looking at a design for pressure control around the FIES table to increase the stability of the spectrograph, but before we better understand the apparent change in stability with the medium-resolution fiber there seems little sense in pursuing this at the moment.

The current fiber bundle does have a working sky fiber, and we still lack a proper testing and assessment of this mode. Analysis of some blue-sky frames for the 2 medium-resolution fibers involved has yielded the following results:

• The target fiber appears to be more efficient than the sky fiber with a mean difference in throughput of 29%, increasing towards the red. The increase in throughput in the red is a bit strange, since the fibers to the best of our knowledge are exactly the same.

• Beyond 6700Å the inter-order distance gets so small that it starts to affect the scattered light subtraction significantly. At 6700Å the difference in the reduced spectrum using only the target fiber, or using the simultaneous sky mode is $>$2.5%.

The old FIES pickoff mirror was replaced with a new mirror with a durable ``enhanced-aluminium'' coating. The new mirror has about 10% better reflectivity than the old one.