Edwin Hubble's "VAR!" - First Data Point

[oopfan]

On October 6, 1923 renowned astronomer Edwin Hubble discovered a pulsating star in the Andromeda Galaxy which quickly led to the revolutionary discovery that M31 is a galaxy unto itself 2.5 million light-years away, and not a gaseous cloud of stars within our own Milky Way...

I first wrote about this topic a few months ago. You can read more here:
viewtopic.php?f=9&t=2575

As promised I've begun the process of capturing data points in an effort to construct the star's light curve. I was hoping to use my lowly 71mm f/5.9 refractor and Bortle 5 skies but alas it is not feasible according to my calculations. At minimum light (magnitude 19.8) the required integration time would take days to complete.

I turned to a remote site in Spain to help out: an RC Optical Systems (RCOS) 12.5-inch Ritchey-Cretien. Last night I grabbed a one-hour time slot and captured 12x 5-minute subs. According to my calculations the 60-minute stack would have yielded a good signal-to-noise ratio. Unfortunately M31 was not high enough in the sky to yield that kind of quality so I discarded 10 of the 12 subs, and stacked the remaining two for a total integration time of only 10 minutes.

The screenshot below is from a program called AstroImageJ (AIJ). I use it for all photometry and exoplanet work. There is one green bulls-eye and two red ones. The red "apertures" are known constant brightness comparison stars: C2 is magnitude 15.522 and C3 is magnitude 16.01. The green aperture I placed on top of Edwin Hubble's variable star. You can see from the screenshot that AIJ calculated the magnitude as 19.168. Just now I looked at an ephemeris and see that the magnitude was expected to be around 19.2. So it works!

This is pretty amazing considering only 10 minutes integration time but the sad part is the accuracy is not very good. AIJ computed the error to be 0.155 magnitudes. Well, I wouldn't say that is unusable but I know I can do better. What's worse is that the star at minimum light of 19.8 may not even be detectable with only 10 minutes integration time. Thankfully, M31 grows higher in the sky with each passing night so it is increasingly possible to obtain better quality data. The first frame of last night's data was captured at only 34 degrees above the horizon, and the last frame at 46 degrees. The frame quality increased as time went on.

Brian

M31 V0619 AIJ Screenshot.jpg (229.67 KiB) Viewed 1892 times

[oopfan]

Here is Robert Gendler's image, rotated to match my field.

Brian

M31 V0619 Robert Gendler.jpg (83.94 KiB) Viewed 1878 times

Courtesy Robert Gendler
RCOS 20-inch Ritchey-Cretien
http://robgendlerastropics.com/M31NMmosaicglobs.html

[oopfan]

One thing I failed to take into consideration when calculating signal-to-noise ratio is that the star is embedded within dust. (Notice in Robert Gendler's image that there are dozens of nearby stars that are just above the detection threshold. Those stars are more deeply embedded but may be just as intrinsically bright as the variable star but because their light is diminished by the dust they are barely detected.)

Of the many levers I can pull in my SNR calculator there is one that addresses this case most directly: Sky Brightness. Currently I have sky brightness set to Bortle 2 but that would only apply to stars against dark space. Since it is not then I will bump up sky brightness to Bortle 3 or perhaps 4. That should give me more realistic estimates of SNR when evaluating other scopes.

Brian

[turfpit]

Very interesting write-up Brian. Thanks.

Dave

[admin]

Hi Brian,

another excellent and interesting write-up Does anybody know how long Hubble had to expose for and with what equipment to be able to perform his measurements? It would be interesting to compare and contrast to see how sensitivity of our equipment has changed over the last 100 years.

Cheers, Robin

[SteveInNZ]

Hubble's plate was 45 minutes on the 100" Hooker telescope.

Steve.

[oopfan]

Thanks, Steve, for that info: 45 minutes exposure. So here I go plugging numbers into my SNR Calculator:

Aperture: 2540mm (i.e. 100 * 25.4)
Focal Length: 22860mm (assuming f/9, it looks that way in photos)
Central Obstruction: 838.2mm (diameter of secondary mirror based on 33% of primary)
Total reflectance/transmittance: 95%
Camera: Any modern CCD that meets the criteria
Pixel Size: 80 microns (ideal for 1-2" FWHM seeing at that focal length)
Dark current: 1 e-/s/px
Peak Quantum Efficiency: 50%
Model of QE Curve: R:1.4, G:1.0, B:1.6
Sky Brightness: 21.81 (Bortle 3)
Target: Magnitude 20 Star (minimum light of VAR!)

Drum roll, please:

Using a 45-minute exposure, single frame, no stacking:
Target electron count: 114,000 (saturated)
Signal-to-Noise Ratio: 290

For a comparison with what I accomplished the other night with a 12.5" f/9 Ritchey-Cretien:
It took a stack of two 5-minute subs to achieve SNR 14.
For the Hooker Telescope, as configured above, it would take 13 seconds.

Aperture is King!
(plus a modern camera helps too)

Brian

[admin]

Hi,

that suggests that the modern cameras collect image information at a rate that is about 200 times higher than the photographic films used a hundred years ago!

Cheers, Robin

[oopfan]

Robin,

It would appear so.

And to all you young whippersnappers out there with your fancy electronic cameras, you will never know the acrid smell of "Stop Bath" in the morning!

Brian

[admin]

Not that young thank you very much

I never did film astrophotography, but I certainly did film processing as part of the optics practicals in my physics degree way back in the midst of time. I'm sure they're all done with digital cameras now and probably take about half as much time to complete because of it. Mind you, making a hologram probably still requires real film and I'm fairly sure we did that.

Cheers, Robin