FLORES Night Procedure

Open the windows

No terminal. Both clients are in the window manager menu.

Start FLORES and Fiber pointing client EXPERIMENTAL 2026 from the menu, and have TOPTEC and ObserveClient 630 where you can reach them. Arrange them so both images are visible at once — you will be looking back and forth all night.

The ordinary pointer centres on whatever is brightest in the field. The fibre pointing client lets you centre on your object, which is the whole difference between observing a bright star and observing something faint enough to be interesting.

Spectral filter to 4

ObserveClient 630 → Spectrograph tab. This is what sends the light into FLORES.

Pointing camera on, and check the focus

TOPTEC. Two things, one of which people forget until the frames are ruined.

• In Cameras, select G1 ON (G2 OFF). G1 is the pointing camera — the one the fibre pointing client exposes with. Without it there is no acquisition image.
• At the top, in Focus: do check your focus. It is the first control in the window for a reason, and a night of beautifully centred, badly focused spectra is a sad thing.

Check the binning is 2×8

FLORES client → Camera settings. It usually is. Nothing enforces it, so look.

2×8 is the default and the sweet spot: good readout noise, and it does not yet destroy the image. Everything reduced so far was taken this way — the frames carry BINNING = '2x8'. But it is only a default. Anyone who changed it left it changed, and it will still be changed when you arrive.

The comp — and the number that saves your night

TOPTEC → Cameras position, then Relay 1 COMP. The one thing nobody could guess.

The stage position in TOPTEC's Cameras panel decides what the chip sees: which fibre, and whether it sees the sky or a lamp. The named presets are:

So type the position in by hand. 13290 is exactly the average of the two comp presets, so the lamp lands between the fibres and both get a moderate, properly exposed comp spectrum at once:

Then Relay 1 COMP → ON, expose in the FLORES client, Relay 1 COMP → OFF. A full calibration is a pair of exposures, back to back without touching anything in between:

• 60 s — the full comp. Carries the wavelength solution across most of the range. You want two rows of sharp lines, not a white slab.
• 1 s — immediately after it. In a 60 s exposure the strong argon lines in the near-infrared saturate, and a saturated line is no use for a wavelength solution. The 1 s frame resolves those same lines cleanly, and combining the two in processing is what stabilises the solution at the red end.

The same 1 s exposure has a second, separate job: taken between targets during the night, it is your cheap check comp — it tells you whether the solution has moved, without costing a minute of sky.

1 s comp
  target            (long-ish)
  standard          (Vega, or another spectrophotometric standard)
1 s comp
  target
  standard
  ...
1 s comp
60 s comp + 1 s     ← the full pair, session closed

Darks and flats: currently, neither

Both are available. Neither is part of the routine yet.

• Darks are not being taken at the moment. An evening sequence is planned, which will collect enough to build a proper dark model. planned Until then take your own whenever you like — set Shutter to DARK in the FLORES client, same exposure time and binning as your science frames. The chip is cold all night anyway, and darks do not interfere with whatever else the Perek is doing.
• Flats are not used in practice — nothing in the reduction needs them. If you want one: position 14930 (flat, both fibers), Relay 2 FLAT → ON, expose 1 s or less, relay OFF.

Send the telescope to the target

ObserveClient 630 → Telescope tab.

1. Prepare the object in the web database beforehand.
2. Get prepared star coordinates loads that position into the RA and DEC fields.
3. GO sends the telescope there.

From here on, the coordinates travel with every image the telescope takes — which is what makes the next steps work.

Choose the fibre: position 1050

TOPTEC. Fibre 2 is the untilted one — put the object there.

Every frame carries two fibres, two sky positions about 18 binned rows apart. One of them is geometrically tilted: its lines are not perpendicular to the dispersion axis.

Fibre 1 is perfectly good in every other respect — it is only the tilt, and the tilt only complicates the reduction. The pipeline does correct for it, so a spectrum in fibre 1 is not lost; it is just more work. Whenever you have a free hand, use fibre 2.

You can see the difference on any comp frame: one set of lines is dead straight, the other visibly leans.

Get the star into the fibre

Fiber pointing client. Green box onto red box — that is the whole game.

1. Expose with Camera: G1 pointing — a second or so to start with.
2. Use pointing fills the Target RA and DEC from the telescope, via the image header.
3. Solve frame sends the frame to astrometry.net. WCS: off becomes a solution, and the client then knows where everything in the field is.
4. The red box is the target — the fibre — at Target X / Target Y. The green box is the source, and it is always on screen; what a WCS solution does is move it onto your object.
5. Move source to target, or Center, closes the gap. Repeat until the star sits in the red box.

Expose

FLORES client. Sixty seconds at a time, for as long as you like.

• Exposure time: 60 s is typical. Long targets are a series of sub-exposures, not one heroic shot — you combine them later, deliberately. For how long yours needs, see how long to expose below.
• Repeat: do not type a huge number. 0 means infinite. Run starts the series and the button becomes stop, which ends it after the current frame.
• Shutter: LIGHT.
• Save to disk must read Saving: ON. If it does not, you are running a very expensive light show.

Then a spectrophotometric standard

If you want fluxes rather than shapes.

Flux calibration needs a standard observed in the same session, and the usual rhythm is one after each target. The pipeline already carries reference spectra for Vega, η UMa, ζ Cas and HR 7596, so those are the convenient ones — and Vega at mag 0 needs 0.1–0.5 s, not a minute.

Log the fibre — and the session

Nothing downstream knows either of these but you.

Both fibres are reduced and written out separately and clearly labelled — but which one held the star and which held sky is your record, not something the pipeline can guess.

Note the session too: one line saying which comp belongs to this frame, and whether the wheel has moved since. When you come back to the data weeks later, that line is the difference between a spectrum and a picture of some lines. Then go back to B1 for the next object.