Is an SQM meter DIYable?

[kbrown]

I read the wikipedia page about SQM meters but it's unclear to me what it actually measures? Is it just a camera that sees in certain bandwidth and produces a number based on what hits the sensor? Does anyone know more about the subject?

Would be great to have a device piggybacking the telescope that logged some information about the "seeing".

As a related experiment I've been pointing my infra-red thermometer at the sky in varying conditions after reading a few articles on the subject. Looks like it would be a feasible way to detect clouds in the dark. Clear skies gives much colder readings than even the thinnest clouds, which I find interesting.

[vlaiv]

You can do your own SQM with a simple guide camera, wide lens - like one used for meteors - all sky variety, or more narrow if you want to measure part of the sky. You just need to calibrate your output.

SQM reading is given in magnitude per arc second squared (or arc minute) and it is based on number of photons captured in interval of time.

You can try "absolute" calibration, but it involves knowing a lot of things, like exact quantum efficiency of your sensor, aperture of lens, any light losses in lens, etc ... Or you can simply do relative calibration. That one is much easier.

Take a frame, do regular frame calibration (subtract dark frame of matching exposure) - calculate average pixel value over whole image and compare to value from another calibrated SQM meter. If that other meter is giving you magnitude value - make sure you do log conversion before - to get photon ratio. Find a coefficient that multiplies your average pixel value to get photon ratio from other SQM device. Do this a couple of times under different LP. Take average of coefficients.

When you want to calculate your on SQM reading - do the same, take exposure, calibrate with darks, get average pixel value and then multiply with above coefficient. After just convert to magnitudes.

[JamesF]

This might also be interesting?  https://roamingdrone.wordpress.com/2008/11/13/arduino-and-the-taos-tsl230r-light-sensor-getting-started/

James

[kbrown]

16 minutes ago, vlaiv said:

You can do your own SQM with a simple guide camera, wide lens - like one used for meteors - all sky variety, or more narrow if you want to measure part of the sky. You just need to calibrate your output.

SQM reading is given in magnitude per arc second squared (or arc minute) and it is based on number of photons captured in interval of time.

You can try "absolute" calibration, but it involves knowing a lot of things, like exact quantum efficiency of your sensor, aperture of lens, any light losses in lens, etc ... Or you can simply do relative calibration. That one is much easier.

Take a frame, do regular frame calibration (subtract dark frame of matching exposure) - calculate average pixel value over whole image and compare to value from another calibrated SQM meter. If that other meter is giving you magnitude value - make sure you do log conversion before - to get photon ratio. Find a coefficient that multiplies your average pixel value to get photon ratio from other SQM device. Do this a couple of times under different LP. Take average of coefficients.

When you want to calculate your on SQM reading - do the same, take exposure, calibrate with darks, get average pixel value and then multiply with above coefficient. After just convert to magnitudes.

That's awesome, thanks! So I basically already have what I need for this in terms of hardware (mono guide camera on a finderscope piggybacking my telescope). I guess you could do this with your primary imaging CCD too if you're only interested in that tiny section of the sky... Not sure if I can borrow an SQM meter anywhere but I shall investigate if I can do at least an approximate absolute calib. Would be good to be able to compare against a real meter though at some point.

[kbrown]

11 minutes ago, JamesF said:

This might also be interesting?  https://roamingdrone.wordpress.com/2008/11/13/arduino-and-the-taos-tsl230r-light-sensor-getting-started/ 

James

Interesting indeed. Now my head is full of all sorts of ideas which probably ends up adding another 3kg of junk on my already heavy rig

[vlaiv]

11 hours ago, kbrown said:

That's awesome, thanks! So I basically already have what I need for this in terms of hardware (mono guide camera on a finderscope piggybacking my telescope). I guess you could do this with your primary imaging CCD too if you're only interested in that tiny section of the sky... Not sure if I can borrow an SQM meter anywhere but I shall investigate if I can do at least an approximate absolute calib. Would be good to be able to compare against a real meter though at some point.

Yes, easiest thing to do would be to take any of your images and try to work out sky brightness from that.

I'm not sure that sky brightness is clearly defined though. There should be some sort of standard for SQM reading. Maybe there is and it's just me not being aware of it.

What I mean by this is that I'm not sure what magnitude is used as base line for SQM reading. It is probably V band from UBVRI - or what we usually call visual magnitude. In V standard, 0 mag star produces about 880,000 photons per cm squared per second. That should be your starting point.

V band is suitable for measuring visual SQM value, but for AP it should be sensible to define SQM readings in bands of interest - like L (400-700) or for each color (roughly 400-500, 500-600, 600-700nm).

At some point you need to switch from average to median - because the more "zoom" you have - you will be able to resolve more stars, and you want sky brightness. So either median, or average with some sigma rejection for bright pixels (or stars).

[drjolo]

I created some time ago simple calculator to estimate sky background brightness on existing subframes https://stargazerslounge.com/topic/310533-sky-background-brightness-calculator/ . 

[cjdawson]

I've got an app for my iPhone called Dark Sky Meter that is supposed to produce an SQM value.  Would be shocked if there wasn't also an android version.

[Macavity]

There is a plot from the "Knoll Shaefer"(!) model (qv) relating 
Sky Brightness & Visual Limiting Magnitude! Not easy to find! ?

As with most empirical laws you can probably drive a small
coach and horses through it. But gives a fair idea of things?  
For my DIY (Arduino) solution, I just added a "trim-pot" so I
can adjust my SQM based on perceived limiting magnitude! ?

[AngryDonkey]

17 hours ago, vlaiv said:

it is based on number of photons captured in interval of time.

I might be dense here but what is the actual figure for interval of time in relation to magnitude per arc second squared? Is it 1s? Is it the length of time the eye can 'integrate' light?

Thanks!

Mike

[vlaiv]

9 minutes ago, AngryDonkey said:

I might be dense here but what is the actual figure for interval of time in relation to magnitude per arc second squared? Is it 1s? Is it the length of time the eye can 'integrate' light?

Thanks!

Mike

Well, magnitude system is just logarithmic scale of ratio of two values. Mag 10 star is 10000 times less bright than mag 0 star.

So in reality there is no direct relation to time. Time is there so you can scale your integration time to obtain photon count that you can compare to photon count of mag 0 source.

You want to compare two sources at same integration time to determine their relative brightness. So if you know that Mag 0 star emits ~880,000 photons per second per cm squared, then you can work out how to compare that value to photon count per pixel on your image. Take aperture of your scope, losses in light train, sensitivity of your camera, and integration time and work out how much photons did you actually collect from 1cm squared of aperture in one second and compare that to 880,000.

If figure for mag 0 star was given in photons per minute per foot squared, then you would convert to those units prior to comparing and finding scale factor.

[AngryDonkey]

Thank you, that makes perfect sense!

[Macavity]

Unihedron (SQM people), rather generously, provide an interactive page
and references for various formulae they use in their calculations. ?
(I knew I got it from *somewhere* and wanted to find the LINK again!)

http://unihedron.com/projects/darksky/NELM2BCalc.html

I have come across some interesting stuff re. CLOUD detection via IR.
For one's "general education" probably worth while checking out...