On August 22nd, 2020, I gave a talk on the drift scanning method during an AAVSO Spectroscopy Observing Section webinar.
You can watch a recording of that presentation here:

Recording high quality spectra of the stars is easier than you ever thought it was! Drift scanning is a method which can be used to record scientific spectra using any telescope, even if it doesn’t have a tracking drive. Dark skies aren’t required– I shoot all of my spectra from inside the city (a “white zone”). The large aperture of most telescopes means that your limiting magnitude will be much deeper than with the standard DSLR + objective grating method.

Gather Your Equipment and Supplies

To get started, you will need:

    • A telescope (a dobsonian is
      ideal due to its large aperture)
    • A camera (this guide was written assuming an astro camera; a
      webcam should work, though)
    • A laptop or cellphone
    • A diffraction grating

Items to potentially increase resolution
(not required, but nice to have):

    • Wedge prism
    • Extension tubes / barlow housings
    • Focal reducer/Barlow (if scope is excessively fast (<f/3) or slow(>f/8)
    • Aperture mask (could be useful if you want to try very bright stars)

Don’t forget these support essentials:

    • Camera cord(s) (data and power)
    • Dew shields / hairdryer (if observing in a very dewy location)
    • Extension cord or battery pack (if using laptop or cooled camera)
    • Table (if using laptop)
    • Chair or stool

Before you set up to record spectra for the first time, make sure that you have all of the correct camera drivers installed, as well as a good capture program to control the camera with. I use Sharpcap.

Before You Go Out

The Golden Rule: Know Your Stuff
The worst possible time to discover that you don’t have the correct drivers installed is when you’re plugging in the camera to take a spectrum. Ditto for realizing that you’re missing an essential adapter or cord. It’s oh-so-easy to frustrate a beautiful clear night away trying to deal with technical issues in the dark (ask me how I know!), so prepare as much as you can ahead of time. Make certain that you know how to operate your personal telescope and camera before you set up for your first spectro run, lest you find yourself fumbling through assembly for the very first time in the dark.

Select Your Targets
It’s a good idea to have a plan before you go outside to set up, even if your plan is just “I’ll browse around the sky and point at random stars”. When you know ahead of time the general brightness of the targets you’ll be looking at, you can pick the correct spacing to optimize your spectra.

More spacing = higher resolution, but the more you let the spectrum spread out, the fainter it’ll get. If you’re planning to shoot spectra of bright stars like Vega or Betelgeuse, you can get great results with high spacing; but if you want to shoot faint stars like WZ Cas or WR 137, you should probably use a low spacing.

If you want to find specific targets to shoot, but aren’t sure where to find stars with interesting spectra, SIMBAD is a good resource. If you’re interested in Be class stars, you can query a large catalog of them using the Aras-BeAm database.

Name Magnitude
Omicron Cet 2 to 10
Zeta Tau 3
EZ CMa 6.9
Zeta Pup 2.3
Y CVn 5 to 6
RT Vir 7.4
Alpha Her 3.0
Beta Lyr 3.4
V1679 Cyg 8.0
P Cyg 4.8
TX Psc 5 to 6

Set Up Your Equipment

 1. Set up your telescope. If you’re using a dobsonian, you may want to let it sit outside for an hour or more ahead of time, to allow time for the primary mirror to reach ambient temperature. (A primary mirror which is warmer than the air will give off shimmery heat waves just like a car does on a hot day, creating its own bad seeing.)

You should also set up a table with a laptop on it, to control the camera with. Don’t forget a power cord!

TieDyeAstronomer Drift-scanning setup

IMPORTANT NOTE

ZWO USB 3.0 cameras can be controlled via a smartphone using the ASICAP app instead of a laptop, which is very cool and works in a pinch, but the battery life is abysmal. If you’re using ASICAP on your phone, I have two important tips: make sure that you’re shooting raws, and record video instead of stills to avoid an issue where your highlights will clip to black!

2. Assemble the camera, nosepiece, and grating. If you have a wedge prism to correct for chromatic coma, you’ll probably want to attach it to the grating first, using a piece of tape to hold it in the correct orientation.

(If you’re using a Star Analyzer diffraction grating and a Paton Hawksley prism, you’ll know that the prism is in the correct orientation when the two short while lines on the housings line up.)

Next, screw the nosepiece onto the camera, along with any spacers which you want to use.

Finally, attach the grating.

TieDyeAstronomer Spectrography Camera Assembly

The Barlow Tube Trick
Unscrew the lens assembly from a barlow lens (put it in a safe place!), then screw the grating into the barlow tube in its place. Ta-dah! An extra extension tube! But wait, there’s more to this trick– because the barlow tube isn’t threaded, you can rotate it smoothly and precisely around the camera nosepiece in order to square the grating (step #5). If there isn’t enough friction to hold the barlow in place firmly, place a strip of tape on the inside of the barlow tube, then peel it off. The slightly tacky surface left behind by the tape will create a slight amount of friction between the barlow tube and the camera nosepiece, allowing you to rotate the grating very precisely without unscrewing it.

 3. Align the finderscope(s) using a high magnification eyepiece– you want them to be as precise as possible.

 4. Find and center a bright star to use for the next two steps.

 5. Square the grating to the camera sensor. This step makes the spectrum run perfectly horizontally across the field of view. Squaring the grating typically takes trial and error; you can get close by aligning the white line indicator on the outside of a Star Analyzer with the horizontal dimension of the camera sensor, but to get it perfect you’ll probably need to repeatedly make fine adjustments to the rotation of the grating, then insert the camera into the focuser just far enough to check alignment. You may want to use a small piece of tape on the outside to hold the grating in the correct orientation on the camera nosepiece.

Each star will have a spectrum on either side; one bright and one faint. Make sure that when you square the grating, you’re placing the bright spectrum to the right of the star!

It’s best not to rotate a spectrum in software, as even a rotation by a fraction of a degree can introduce artifacts. That makes squaring the grating an important step.

 6.  Focus the camera. The spectrum lies at a slightly different focus than the zero-order star, so make sure that you focus on a feature in the spectrum, and not the star. Do not set your focus by making the spectrum as skinny as possible– there is some astigmatism present in the spectrum, and the sharpest focus is likely achieved with a spectral strip that’s taller vertically.

As an example of the effect of astigmatism, check out this comparison that I made by taking two spectra of P Cyg on the same night; one with what I thought was a poor focus, and one taken after I refocused for a “better” focus. The one with the “poor” focus was nearly twice the resolution!!

Recording Your Data

1. Starhop to the target. As an experienced visual observer, starhopping typically takes me anywhere from ten seconds to 5 minutes. The smaller your camera’s sensor is, the tinier the chunk of sky it can take in, and the harder it’ll be to starhop. It helps a lot to have a precisely aligned magnifying finder. Before I started using my magnifying finder, it would sometimes take me 15 minutes to starhop to a target. 

Starhopping tips:

  • Turn your gain up to maximum
  • Use a display stretch to increase contrast on the preview screen
  • Use a digital chart app such as SkySafari which will overlay a field-of-view circle on the chart

Video: Short instruction of drift scanning process step-by-step.

 2. Rotate the camera within the focuser until the star drifts across the sensor from top to bottom. (This allows plenty of room to fit both the zero order and the spectrum along the horizontal axis of the sensor.)

3. Set the gain and exposure. Your goal is to get the spectrum somewhere in the center of the histogram; not scrunched up all the way over to the left (in which case it will be faint and noisy), and not too close to the right (in which case it might saturate and clip to white). Err on the side of underexposure — It’s MUCH better to have a noisy spectrum than a saturated one. When you stack your frames later on, it will remove most of the noise.

  • For very bright stars like Vega you’ll probably want to use unity gain, and a short exposure. (For my setup, that means gain=111 and exposure less than 0.2s.) In theory, unity gain yields the highest dynamic range while not discarding any light.

Video: How to record stellar spectra with the drift scanning method.

  • For faint stars, you’ll want to use maximum gain, then raise the exposure until the spectrum starts to trail vertically. If the spectrum is close to being overexposed, you can lower either the gain or exposure.

Maximum gain may look like it’s noisier because it makes the noise look so much brighter on the preview screen, but in situations like drift scanning where you aren’t able to increase the exposure to collect more light from the target, maximum gain yields the best signal to noise ratio.

 4. Position the star in the upper left corner of the field (or just above), so that the Earth’s rotation will carry both the star and spectrum across most or all of the sensor.

 5. Start recording a .ser file once the zero order star and spectrum are both fully within the frame. Stop the recording before the zero order and spectrum start to run off of the frame. (It’ll make it easier later on if you don’t have to discard any frames due to the spectrum running off the edge.)

I use Sharpcap to capture my SER files. SER is an uncompressed video format; the quality is like FITS, but it’s much easier to process a single SER file later than it is several hundred FITS files.

 6. If you’re using an uncooled camera, you should now cover the front of the scope, and without changing any settings, record another .ser file with anywhere between 10-100 frames. This second SER file contains the dark frames for this object. You’ll use these dark frames to remove false signals inside the camera, such as hot pixels and amp glow.

It’s important to record a new set of darks for every target because darks must be taken at the exact same temperature as the frames they’ll be used on, and with an uncooled camera, the sensor temperature is always changing. On the other hand, if you’re using a cooled camera, then you can keep your camera at a set temperature (for example: always capturing spectra at -10°C) and reuse the same darks over and over.

Processing Your Data

Create the master dark frame:

  1. Load the SER containing the darks into SiriL. Don’t use the ‘File conversion’ tab; SiriL can load .ser files directly from the ‘Sequence’ tab, and that saves disk space.
  2. Stack the darks SER under the ‘Stacking’ tab. Use “Average stacking with rejection” mode, specify no normalization and no rejection, and save the result as a master dark.

Apply the master dark:

  1. Load the .ser containing the spectrum into SiriL using the ‘Sequence’ tab.
  2. Using the ‘Pre-processing’ tab, apply the master dark to the sequence, disabling cosmetic correction.
  3. Check under the ‘Sequence’ tab to make sure that the pre-processed sequence (prefix “pp_”) was automatically loaded.

Align the frames:

  1. In the image preview window, draw a small selection box around the zero order star image.
  2. Under the ‘Registration’ tab, select “One Star Registration (deep sky)”, make sure that “Follow star movement” is checked, and click “Go register”.
  3. Check that the sequence registered correctly, by switching to the ‘Sequence’ tab, holding down the “+” button (under “Image selection in sequence”), and watching the image preview screen.

The detected zero order will have a circle around it in each frame. If the circle around the zero order disappears or jumps around, there might be a problem with the registration, and you should go back to frame #0 and try using a different selection box size.

Video: How to stack a deift scanned spectrum in SiriL!

It can happen with faint targets that some frames will not register correctly, no matter what you try. As a last resort, you can disable certain frames entirely by enabling the checkbox “Exclude” while the undesired frames are loaded, then changing the “Stack this set of images:” dropdown from “All” to “Selected” under the ‘Stacking’ tab.

Stack the frames:

  1. Using the ‘Stacking’ tab, stack the sequence, using the same settings as for the master dark (Stacking methods: ‘Average stacking with rejection’, Normalization: ‘No normalization’, Rejection: ‘No rejection’).
  2. Check for ghosting or stacking errors. After the image is stacked, it should automatically open in the image preview screen. Mess around with the stretch (“Linear” by default), white point slider, and black point slider, inspecting the image for any ghosting or stacking errors.

A drift scanning pass which stacked well should look like a less-noisy version of a single frame, with a dark gradient towards the edge of frame, but no gradient over the zone containing the target spectrum or zero order.

Graphing the Results

Congratulations, you now have a fully processed 2D spectrum image! Isn’t it pretty?? Now it’s time to make a graph from the image (a.k.a. ‘extract the 1D profile’). There are several pieces of software which will do the trick, but no matter which software you use, the general procedure is the same:

  • Profiling: Graph the raw brightness values of the spectrum. This is usually done by selecting the horizontal rows of the image which contain the spectrum you’re interested in. The resulting graph is called the “profile”.
  • Background Subtraction: Subtract the background sky brightness from the profile. You’ll usually do this by selecting the rows of the image which contain only empty sky.
  • Calibration: Convert the X axis of the graph from pixels to wavelength (Angstroms). You’ll probably do this by identifying features at known wavelengths within the spectrum.
  • Instrument response correction: Using an instrument response curve (which can be generated from a spectrum of a standard star taken under similar conditions as your target), remove the effects of your telescope, camera, and the atmosphere from the spectrum.

Once you’ve completed all of the above steps, you’ll have produced a scientific spectrum! That’s great! So what software should I use? I personally use RSpec. RSpec is expensive ($109), but it’s also the easiest-to-use spectrography software currently available. The other spectrography programs out there will perform all of the necessary functions, but the interface is usually clunky and the learning curve steep.

Tom Field, the creator of RSpec, has already created a series of excellent (short!) video tutorials, so I won’t rewrite the manual here. Instead, if you’re using RSpec, you should watch Tom’s videos. You don’t have to watch all of them to learn what you need; I would recommend watching, in the following order:

  • “One-point, Linear Calibration Revisited” (Loading an image, linear calibration),
  • “Non-linear, One-Point Calibration – (part of Update 8 )” (Nonlinear calibration, essential if you’re using a wedge prism),
  • “Update 12 – Real-time flux calibration, discontinuous background removal, etc.” (Selecting a spectrum, background subtraction, cropping a spectrum),
  • “Adjustment for Instrument Response (part of Update 3)” (Instrument response correction)

If you’d rather not buy RSpec, there are several free spectrography software options. BASS (Basic Astronomical Spectrography Software) and Visual Spec are two free options which can yield spectra which are just as high quality as those from RSpec, once you learn how to use them.