Precision of Sky Quality Meters

[deepskytraveler]

During a recent observing run near Ft. Davis, TX Dragan Nikin and I had the opportunity to compare frequent readings between our Unihedron SQM-L meters over multiple nights. We had both purchased these instruments with the intent to acquire quantifiable measurements of sky brightness and then able to use those measurements as a relative comparison of sky brightness between observing locations. We both understood the multitude of factors that affect the sky brightness measurements. However we were not prepared, nor are we able to explain, the wide variances we consistently encountered between our readings – acquired at the same times, same observing location, and literally with the instruments physically adjacent to each other and resting on a flat surface, i.e., pointing to the same angular segment of sky.

On every night except for one my meter's readings averaged 0.1 to 0.2 magnitudes darker than Dragan’s. On that one night Dragan's meter was consistently reading 0.1 to 0.2 magnitudes darker than mine. Throughout the week it was not uncommon to see the readings between these meters vary by 0.2 to 0.5 magnitudes. Even on a single meter acquiring a series of back-to-back measurements over a period of 60 to 120 seconds there were sometimes variances of up to 0.5 magnitudes. According to the Unihedron “Each SQM‑L is calibrated using a NIST-traceable light meter. The absolute precision of each meter is believed to be +/-10% (+/-0.10 magnitudes per arcsec squared).” My unscientific sampling does not seem to be in alignment with Unihedron's stated precision.

I’m interested in hearing from other SQM owners about their experiences. What degree of variance do you get in multiple back-to-back readings (after throwing out the first due to the internal circuitry changing temperature on initial power-up)? Have you gathered measurements using multiple SQM meters under the same conditions - what were your findings? Let us know.

[Marko]

Interesting read and I am very curious. Several of us here in central CA have compared meters in much the same way over the last few years and I for a while calibrated what I heard other observers say the reading was by an offset if I did not have my meter handy at the time. I think I ALWAYS get a slightly darker sky than Steve G here on DSF even though I am often only15 feet away. (Guess he does not know how to pick the right place to put his scope ... LOL).

You have covered the variables that you only compare SQM-L to SQM-L and same spot in sky and throw out the first 1 or two readings, they vary and so on. I have also noticed you have to keep any light (especially red) from hitting the red plastic front cover so I try to be aware of that as I measure. (no headlamps, nearby PCs and so on). I try also to find an area of the sky that is fairly dark and star free but still close to zenith in order to not be influenced by say pointing directly at M45 because the L meter has a narrow cone of highest sensitivity. We all have seen variance in pointing at milky way and in fact when I was down in Australia the variance of pointing right at our galaxy's core which is way closer to zenith down there was quite alarming.

I have questioned the following two variables that may be involved in your differences:
1) Both meters should be equalized to the same temperature and you can read that with the meter as well by the way if you keep the key down I think does that reading. If they are very different such as one was sitting out and the other was in a pocket close to the body that from my electrical engineering background I suspect may play in the delta SQM readings but I have not proved it to myself, just seems reasonable.
2) A factor that I am SURE makes a difference is both meters must have a very clean, dry, and unscratched surface in front of the sensor. The guy with mud or dust or dew on the front of the meter will experience a much 'darker' sky than the guy with the clean and dry plastic front. I jest here but you see my point? If one guy wipes the front of the meter with his thumb and another with a lens cleaning cloth there is something in front of the sensors that varies.

As a side point I had several years back a wide-field SQM (the non-L version) and one summer at CalStar I began to notice it was really 'dark' and for month after that the skies around me got 'darker and darker' in fact I had a WONDERFUL set of SQM readings till the next time I synced back up with a known other SQM and decided my meter was WAY out of cal. I have it to this day and don't know what to do about that issue. So my point here is if two meters start to deviate in settings over weeks of time and gradually diverge I would expect whatever it was I had happen is happening to your own meter (most likely the one with the far 'darker' skies).

[FaintFuzzies]

Yeah, I was one of the original purchasers when someone in TAC offered to get a bunch for us quite a few years back - at least 10 years ago.

I've found the same thing as the OP found. There were at least 3 or 4 of us with the SQM...and all three are DIFFERENT...pointing at the same part of the sky standing next to each other holding this thing high. Our differences was no more than 0.3...usually at 0.2.

I just use my naked eye and determine the NELM and be done with it. I more often than not, quickly re-determine the NELM as the night progresses or notice if the skies looks "better" or "worse".

So I sold mine years ago...as I'm a bit old school when it comes to determining sky quality. Been doing visual astronomy for 40 years. No doubt that this is a great instrument, but it wasn't for me.

[Dragan]

I purchased my SQM last year thinking its the way to go for a non-subjective quantitative measure of the quality of the night sky. Since purchasing the meter, I continually find differing readings (not only with my meter but with friends meters as well) and this past week in Ft Davis was just another case of meters varying too much. I'm finding there are just too large a discrepancy when it comes to the readings any two meter will give. And if two side by side meters can't give a comparable reading, how do you say which is the right one?

I'm with Alvin though. In my trailer I have a full set of limiting magnitude charts. I use to use them religiously. I think it'll be time to dust them off again.

[Marko]

These days I use the meter as one of a few indications and agree with the earlier comments that it is not absolute reading. I think there is one point that should not be overlooked and that is the value of a 'delta-darkness' meter on any given night that has two uses. You can watch the progression of darkness as the night goes on and you can judge the impact of large light domes nicely with this meter.

For most serious sessions I get a long ways from the city but when I am only 10-20 miles from a big city due to casual observing or in my backyard (sodium light district) it's frequently nice to judge the variation in sky brightness with the SQM-L narrow cone of detection to see just how much brighter one quadrant is than another. You cannot measure too low due to readings get real good pointing at the ground but if you stay up 40 degrees it is a fair delta-darkness estimator. The wide version of the meter is not good for the delta-darkness reading as it has a huge cone of detection.

[Howard B]

As a side point I had several years back a wide-field SQM (the non-L version) and one summer at CalStar I began to notice it was really 'dark' and for month after that the skies around me got 'darker and darker' in fact I had a WONDERFUL set of SQM readings till the next time I synced back up with a known other SQM and decided my meter was WAY out of cal. I have it to this day and don't know what to do about that issue. So my point here is if two meters start to deviate in settings over weeks of time and gradually diverge I would expect whatever it was I had happen is happening to your own meter (most likely the one with the far 'darker' skies).

The first two SQM's I had (the original wide angle version) worked fine for a couple months after getting them but then started to display progressively darker readings during the night - like in the 25 to 30 range! - and became worthless. I sent them both back to Unihedron and was graciously and quickly sent a replacement SQM. The unit I have now, the third one, has been stable for a couple years now and I use it to take a reading before and after each of my observations as a relative check on sky darkness. I've seen the same level of variation from unit to unit as noted by everyone else and have chalked it up to the production variance of the sensors.

Several years ago Chuck Dethloff organized a group buy of 10 SQM's for members of the Portland club (RCA) and he tested them all side by side to make sure they worked before handing them out. If I remember correctly most of the SQM's consistently gave readings within 0.01 to 0.02 but several were brighter or darker up to 0.3. So the typical SQM can't be regarded as giving an absolutely accurate reading, but if used in the same way over time it can be used to accurately judge changes in sky darkness.

For me this is a good thing because I suck at NELM estimates!

[Don Pensack]

We did a test of 15 observers at Mt. Pinos a few years ago, using the AAVSO charts, to try to estimate the NELM at the site. We got a range of 2 magnitudes!. To me, that doesn't invalidate MELM measurements, but they are only, IMO, relevant to one observer and the figures are meaningless to others. The reason is that there are dramatic (and I mean DRAMATIC) differences in visual acuity and sensitivity and experience.
In contrast, comparing SQMs, we typically find a difference of 0.1 magnitude or less if the units are pointed at the zenith from the center of the parking lot.
Even if the difference between units is 0.3 magnitudes (we've never seen that difference unless one is not pointed at the same point in the sky), that is miniscule compared to the differences in NELM.
If someone says his sky is "mag.6.5" that implies a dark site, but I've been to some of those sites and they weren't dark at all by my reckoning. On the other hand, if someone says the sky is "mag.21.5", I know exactly what to expect, and the reality is always nearly exactly what was expected.

I've been tracking sky brightness at Mt. Pinos since 2005, using a wide-angle SQM, and can now say the site has barely grown brighter in that span of years (less than 0.1 mag.) because I always take hourly readings over the new moon night. The overhead Milky Way adds about 0.15 magnitude compared to being outside the nominal field of view.
I very much appreciate getting rid of the subjective ZLM or NELM because the time needed to make a good estimate is time I'd rather be observing.

After just short of 50 years of observing, having an objective measurement device to make comparisons of sites and variable conditions is worth its weight in gold. If I'm going to drive 2 to 4 hours to observe at a dark site, I need to know the site is truly dark. NELM just doesn't cut it.

[Marko]

Don has a very valid point above. The point is so what if it varies 0.2 to 0.3 because NELM is all over the map from different observers. Again, what I do is note NELM AND SQM. In this way I offer my own NELM (which will only get worse as my eyes age) along with SQM. Also for me the relative readings I get from outing to outing tell me a lot about the site in a delta-darkness sort of way even though the absolute accuracy may be only +- 0.2 or 0.3. It is fast, I already own the darn thing so why not note it.

More recently I am trying to be disciplined and also judge seeing on close, well known, doubles rather than a general star-bloating feeling because even when your eyes age you can still with proper optics tell if you split the pair and how well I should think.

To demonstrate to ones self how SQM can look very much better than things really are I have by mistake in very dark skies when not paying much attention used the meter backward (towards the ground) and when you do that you get a REALLY dark sky reading. Also at last GSSP on the first day I got a great reading as a 'I wonder what it is now' when it clouded over 100% after midnight. SQM alone is only one number, darkness. One needs more info to qualify a sky but at least it is a more or less absolute number even though not greatly accurate.