Overview

See the KPF Web Page section on Calibrations for the overview content.

Notes About Cal Lamp Operations

Warmup Time

Some lamps (the ThAr, UNe, and Broadband lamp) require a warm up time. The kpflamps.%_THRESHOLD keywords indicate how long they must be on before taking cals. Currently (Jan 2024), these values are set to 30 minutes:

The FF_FIBER source mentioned in the keywords above is another broadband light source, but has not had the warm up time set up because it is not currently in use.

A calibration OB which invokes any of these lamps, will check how long each lamp has been on (the relevant kpflamps.%_TIMEON keyword tracks how long each lamp has been on) before taking data. If needed, the OB will simply wait until the _TIMEON value exceeds the _THRESHOLD value.

Sequencing in OB Execution

Note that this cycling of a lamp on and off is done at the beginning and end of each OB. It is not done for each observation within an OB. Thus if an OB contains one set of Thorium exposures and one set of Uranium exposures, both lamps will be powered on at the beginning, all exposures taken, then both are turned off at the end of the OB. As a result, we tend to write OBs with only one of the consumable lamps to avoid excess power on time.

Types of Calibrations

Wavelength Calibrations

Wavelength calibrations are the most critical daily calibration. Absolute wavelength calibrations use the laser frequency comb (LFC) or one of the hollow cathode lamps (ThAr or UNe).

Absolute References: The LFC, and the Thorium and Uranium Lamps

The LFC is the best calibrator due to the density of lines which allows a very precise measurement of the wavelength solution, but it is currently (early 2024) limited due to a lack of flux in the bluest orders. For those, the traditional Thorium and Uranium lamps provide absolute calibration.

The disadvantage of the LFC and the Thorium and Uranium lamps is that they have limited lifetimes. There are components in each system which are “consumables”, so we try to minimize their use. For the LFC, this is motivated by cost – the consumable is a “photonic crystal fiber” (PCF) which is quite expensive.

For the Thorium and Uranium lamps, the lamps themselves fail over time and while replacement lamps are not very expensive, they are not identical, so replacing a lamp requires some manual checking of the line lists.

The LFC is turned on or off (by changing the control system from “StandbyHigh” mode to “AstroComb” mode or back) by each calibration OB which uses the LFC. The Thorium and Uranium lamps are simply powered on and off at a power strip by the script executing the calibration OB.

Relative Reference: The Etalon

The etalon is also used for wavelength calibration, but in a slightly different way. The etalon spectrum is not fixed to absolute wavelengths as the LFC and the Thorium and Uranium lamps are, but it has the advantage of being far more robust. The etalon itself is a passive filter, but it is fed by a broadband light source (a “SuperK” from NKT) which has a much longer life than the absolute references we discussed above. The SuperK is left on at all times which helps stability of the output spectrum, thus etalon light is always available.

The positions of the etalon lines can evolve slowly with time, but on short (<1 day) timescales they are stable. Thus the etalon calibration frames can be used to track the evolution of the instrument over daily timescales (anchored by the morning and afternoon wavelength cals). Thus it is valuable to have etalon observations taken periodically throughout the day and night.

Broadband Flats

KPF takes relatively standard spectral flats using a broadband lamp. Because of the extreme requirements for PRV measurements, the flats must be extraordinarily high signal to noise and be taken daily to track any changes.

Biases and Darks

KPF takes biases and darks like other instruments. We try to schedule these in to the lamp warmup time built in to other calibrations, so they end up being spread out throughout the day during other calibration OBs.

Other Calibrations

We currently also take two other calibrations: a set of biases for the Exposure Meter every day and a set of CRED2 guide camera biases a couple times per week.

Morning Wavelength Calibrations

The morning wavelength calibrations are executed at 7AM HST each day by the run_morning_cals script. The script executes three separate OBs (detailed below). Before each OB is executed, the script verifies that the kpfconfig.ALLOWSCHEDULEDCALS keyword is set to “Yes”.

MorningCals_ThAr.yaml

This OB contains five sequences. First, a set of biases are taken, then a 20 minute long dark frame. Then a few etalon frames are collected. This takes up most of the lamp warm up time for the Thorium lamp.

Once the Thorium lamp is warm a set of Thorium frames are taken which are set to illuminate all traces in the pseudo slit (all three science slices, the sky fiber, and the simultaneous calibration fiber). Then a set of longer Thorium exposures are taken in order to provide exposures of reasonable length for calibrating the Ca HK detector.

MorningCals_UNe.yaml

This OB contains three sequences. First, a 20 minute long dark frame is taken to spend most of the Uranium lamp warm up time.

Once the Uranium lamp is warm a set of Uranium frames are taken which are set to illuminate all traces in the pseudo slit (all three science slices, the sky fiber, and the simultaneous calibration fiber). Then a set of longer Uranium exposures are taken in order to provide exposures of reasonable length for calibrating the Ca HK detector.

MorningCals_UNe.yaml

This OB contains three sequences. First, a 20 minute long dark frame is taken to spend most of the Uranium lamp warm up time.

Once the Uranium lamp is warm a set of Uranium frames are taken which are set to illuminate all traces in the pseudo slit (all three science slices, the sky fiber, and the simultaneous calibration fiber). Then a set of longer Uranium exposures are taken in order to provide exposures of reasonable length for calibrating the Ca HK detector.

MorningCals_LFC.yaml

This OB contains a single sequence to get LFC spectra and is thus much shorter than the other two OBs in the morning calibration script.

Afternoon Wavelength Calibrations

The afternoon cals run starting at 3pm HST. The run_afternoon_cals script runs 3 OBs, but those OBs contain more sequences and instrument configurations than the equivalent morning cals as detailed below.

AfternoonCals_ThAr.yaml

This OB contains eight sequences. It begins with biases, a dark, and some etalon frames to use the lamp warm up time similar to the morning OB. After that, it cycles through several observation sequences illuminating each trace in the pseudo-slit individually: science fiber only, then sky fiber only, the simultaneous calibration fiber only, then all fibers at once, and finally a set of exposures for Ca HK.

AfternoonCals_UNe.yaml

This OB contains six sequences. First, a 20 minute long dark frame and some etalon frames are taken to spend most of the Uranium lamp warm up time.

Once the Uranium lamp is warm it cycles through several observation sequences illuminating each trace in the pseudo-slit individually: science fiber only, then sky fiber only, the simultaneous calibration fiber only, then finally all fibers at once.

AfternoonCals_LFC.yaml

This OB contains a single sequence to get LFC spectra and is thus much shorter than the other two OBs in the afternoon calibration script.

Flats

The afternoon flats script starts at noon HST and contains a single OB. This OB contains three sequences. It begins with a 20 minute dark and some etalon frames to occupy the broadband lamp warm up time. Then it takes a long set of 100 flat exposures.

SoCal

The solar calibrator operations are still being finalized. The SoCal OB monitors the Sun over several hours after morning cals are complete and before afternoon flats begin.

Overnight Etalon

We have scheduled four long etalon sequences to happen overnight. The start times are chosen so that they should not be in progress when an observer might be starting use of the instrument (e.g. at a quarter night boundary). The scheduled etalon frames will not begin during observations because the kpfStartOfNight script sets the kpfconfig.ALLOWSCHEDULEDCALS keyword to “No” and the script to run the overnight calibrations respects that keyword.

If for some reason, they are running when an observer needs to being observing, the cals can be stopped using the standard tools to halt a script in progress (e.g. the “Request Script Stop” button in the OB GUI).