Upgrading the FibreHead
Since moving to the MPG/ESO-2.20m in 2002-Oct, FEROS has been
performing with suboptimal efficiency. The benchmark efficiency set at
the ESO-1.52m telescope is 15% (minimum) - 20% (desirable). During
Commissioning-III (2002-Oct), the maximum measured peak efficiency was
only ~6%. Modifications to the fibrehead/microlens/fibre interfaces
during late October/early November allowed a maximum peak efficiency of
~11% to be measured during Commissioning-IV (2002-Nov). Subsequently,
during a technical night of excellent quality in 2003-Jan a maximum
peak efficiency of ~12% was measured. Further modifications to the
fibrehead/microlens/fibre interfaces during early March allowed a
maximum peak efficiency of ~13% to be measured during the
(mini-)Commissioning-V (2003-Mar).
It was quickly realised that the suboptimal efficiency in 2002-Oct was
due to poor alignment of the fibrehead/microlens/fibre interfaces and
all efforts between 2002-Oct & 2003-Mar were concentrated on
achieving optimal alignment of these three components, culminating in a
fully adjustable installation of the fibrehead/microlens/fibre system
in early March. Obviously these efforts however did not fully restored
FEROS to its full potential.
Having eliminated misalignment of the fibrehead/microlens/fibre system
as the cause of the loss of efficiency, the problem was re-examined. It
was thus realised that the remaining efficiency losses result from an
optical mismatch at the fibre exit to the Spectrograph -- the ouput
f/ratio from the fibres is simply larger than what FEROS is designed
for. There are two, possibly three reasons for this:
- Ball Lenses: The MicroLenses used at the fibrehead/fibre interface are of a spherical design. They produce the wrong input f/ratio to the fibres. Since the initial stages of planning the move from the ESO-1.52m to the MPG/ESO-2.20m it was known that the ball lens solution was a compromise and that a rod-lens design (as used at the ESO-1.52m) would have been a better solution. However at the time of planning and designing for the move we were unable to find a manufacturer who could fabricate the rod-lenses required by the optical design of that time.
- Tight Curves: It is well know that any fibre will degrade the input focal ratio, even in the perfect case with no stress on the fibre. This phenomenon is called Focal Ratio Degradation (FRD). We now know that the fibres are under a significant amount of stress such that the FRD is significantly worse than the nominal situation. Tests have revealed that the current route of the fibres taken from the input plate of the FEROS adapter to the fibre head is far from ideal; it involves a number of small radius curves which significantly increase the FRD. Aside from the fact that solving this problem will require some machining of the adpater, the fibres are simply not long enough to allow a longer, less stressful route.
- Twisted Fibres(?):We
suspect but can not now test, and in any case can also not do anything
about for the moment, is that the fibres maybe twisted between the
instrument and the fibre head. We can not do anything about this
because the fibres are now fixed at BOTH ends.
FEROS FibreHead Upgrade Number 2
Based on the new rod-lens based optical design by AGI and GAV new fibres and a new fibrehead will be prepared. The fibrehead will of course be fabricated by LS-MEC and polished by Kugler as for the previous heads. The fibres and rod-lenses will be prepared and installed into the new fibrehead by GAV & AGI in Garching. The entire fibres+lenses+fibreheadsystem with then be installed into FEROS and the adpater.
With this plan we expect to address all three problems identified above and finally restore FEROS to its full efficiency.
- The new optical design will provide the correct input focal ratio.
- The new fibres will be long enough to allow a stress free routing through their entire length (especially within the adapter) and care will be taken to ensure the fibres are installed in a twist-free, minimum-stress manner.