Light Pollution Meter Information Page by Avery Davis

Updated April 11, 2003

Notes on the design and construction of a visual comparison photometer for measuring sky brightness and light pollution. This is a content-oriented page, so I hope you will forgive me if it lacks "style". This is simply a Microsoft WORD document with a "Save-As HTML".


Background and History
Printed Circuit Board and Optical Tube Assembly
Prototype Circuit
Post to Darksky list, May 9, 2001
Chronological Narrative. "What’s New" is at the end.

Background and History

I am a member of the International Dark-sky Association ( and the Southern-Arizona section of IDA, and at the March 2001 SA-IDA section meeting, which just happens to be held at the IDA HQ in Tucson, Liz Alvarez showed us a couple of interesting instruments which she had recently received. I immediately recognized them as being some kind of light pollution meters (since the Pitch-Black Meter described in an article in the February 2001 Sky&Telescope magazine was still fresh on my mind). Liz then asked me to take them home and make an evaluation of them for her, and start taking measurements with them, but she did not have any documentation to give me on them. One of them was from Dr. Steven Morris at L.A. Harbor College, and I have contacted him about it. Liz told me the other instrument was from Dr. Arthur Upgren of Wesleyan University, and I have been in contact with him, also. For background, I am not an astronomer, except in the amateur category, but I am an Electrical Engineer and a radio amateur, so I am more comfortable with electronics than I am with optics or mechanics.

I have built a "Pitch-Black Meter" circuit such as the one Gote Flodqvist described in Sky&Telescope, February 2001. Of course, being an electronics designer myself, I have designed what I think is a better circuit for controlling the LED brightness. Mr. Flodquist's concept of controlling the current of the LED is in my opinion a very good idea, I just think that his circuit is not the best for achieving that control. Additionally, Mr. Flodquist did not present a method of calibration for his PBM. I hope to design a light pollution meter that is simple and cheap to build, and is easy to calibrate.
I want to make clear to everyone what a "visual comparison photometer" is and is not. It doesn't really measure light. You use your eye to take in the actual light, and the instrument gives you a variable light source for comparison. Right now, the measurements are relative and uncalibrated (that is, no reference to any physical units). It is possible to calibrate the visual comparison photometer, and the people who used them in Canada in 1977 developed a calibration method, but I have not reproduced this. The article I reference by Pike and Berry (Sky&Telescope February 1978) describes it, and I highly recommend reading this article. If you could help me come up with an easy and inexpensive calibration method for this photometer, I would be most grateful.

Printed Circuit Board and Optical Tube Assembly

I have redesigned the circuit with a few improvements, but I am currently still working on an updated layout of a printed circuit board for it. The updated circuit includes two options for measuring the LED current. JP1 is where a microAmmeter should be inserted into the circuit. JP3 is where a milliVoltmeter would be added to the circuit. Only one meter is needed, but if a milliVoltmeter is used at JP3 and a milliAmmeter is not inserted at JP1, a shorting jumper should be installed at JP1. The circuit for the milliVoltmeter at JP3 includes R6, 1000 Ohms at 1%, as a meter shunt. This results in a current-to-voltage conversion factor of 1 milliVolt for 1 microAmp of current. Also, I added circuitry for a second battery to power a digital voltmeter, since a DVM cannot measure its own power circuit. This permits the option of installing a panel-mount DVM on the light-pollution meter assembly instead of using a separate digital multimeter so that the light-pollution meter will be completely self-contained. Recently, adjustable (trimmer) resistors have been added at R2 and R4 to allow tailoring the current range set by the control. R2 adjusts the minimum current, and R4 adjusts the maximum current. Minimizing the range of the LED control increases the resolution of the R3 brightness control.

To assemble the PCB, generally install the smallest components (resistors) first, then the transistors, LED, controls and the switch. The LED "pilot light", D2, is not needed if a DVM with an LED display is used.

Physically, I designed the board at 2.0" x 3.8" so it would fit inside a 2" I.D. PVC pipe, with holes for the switch and brightness control. I based the "optical tube" on a 2" T-union at the objective end. To make the unit compact, the main tube is a 1" dia. Schedule 40 PVC pipe section, and the 2" dia. Opening of the T-union functions as a glare shield. Given that the opening is 1" I.D., and the desired field of view is 7 degrees, and assuming that the distance from the retinal plane to the eyepiece is about 1.5 inches, then the overall length of the tube assembly computes to be about 6.67". A 5.25" section of 1" I.D. PVC pipe is attached to one of the through-ports of the T-union through a Schedule 40 reducer/adapter: 2"x1" (alternately two adapters, 2"x1-1/2" and 1-1/2"x1", could be substituted). At the other end of the 5.25" pipe, a pipe union is used to attach a 1" schedule 40 pipe union. For the eyepiece (peep-hole), a 1" pipe to ½" NPT adapter is used. A filter or gel is installed over the ½" hole in the adapter (this filter should match the color of the LED). The interior of the tube should be painted flat or matte black (I used Krylon "Ultra-Flat").

A 4.75" section (origninally 3") of 2" I.D. PVC pipe is installed at the side-port of the T-union to contain the circuit board. The board is installed into the pipe so that when the assembly is installed in the T-union the LED-end of the board will intrude into the optical path so that the LED may be seen in the eyepiece. Holes are cut in the sides of the 4.75" pipe section to allow access to the board mounted controls for brightness and ON/OFF. IMPORTANT NOTE: The width of the circuit board is critical for a good fit into the PVC pipe housing, but the tolerance of the board fabrication is very loose. Typically, the sides of the board must be filed down a little (about 1mm) to make a good fit so the switch and control match the holes cut for them. The board is then glued in place.

An end-cap installed on the bottom of the 4.75" pipe section will contain and support the batteries. If a flat-end cap can be found, a jack-screw with ¼"x20 thread could be installed to provide a tripod mount.

Recently, Cliff Haas and I have decided that yellow or amber would be a more appropriate color for the LED and the filter.

Here is the schematic of the LED brightness control circuit board:

Here is the parts list for the LED brightness control circuit board:










BS6I-ND DigiKey, 9V Battery Snap Cable, 6"




YEL, 10mm, Diff

276-216 RadioShack: 585nm, 250mcd @ 40mA




RED, T1-3/4, Diff

LN21RPHL Panasonic, P300-ND DigiKey





Cut from PZC36SFAN Sullins





PN2222, NPN SSGP, PN2222-ND DigiKey





470,1/4W,5%,CFR, 470QBK-ND DigiKey; YL-VI-BN-GD





20K-Ohm 0.75W 10% 3/4" Rectangular Multi-Turn PC-Mount Cermet Trimmer w/End Adjust, Clarostat 343P20K, 900-4663 Radio Shack





EVL-HFAA01B503, Panasonic, P4B0503-ND DigiKey





100K,1/4W,5%,CFR, 100KQBK-ND DigiKey





1.00K,1/4W,1%,MFR, 1.00KXBK-ND DigiKey





10K,1/4W,5%,CFR, 10KQBK-ND DigiKey





PN22SJNA03QE C&K, CKN1191-ND DigiKey





Switch Cap, 5.86mm dia., 981A02000 C&K, CKN1197-ND DigiKey





Red LED Panel Meter, 7516-ME MPJ





4.41"x2.44"x1.06", Hammond 1591BSBK, HM103-ND DigiKey, OR: 5x2.5x2, 270-1803 Radio Shack. Enclosure for panel meter.

Parts for Tube Assembly:





End cap, 2" dia., Sched. 40 PVC, used for battery holder





Pipe, 2" dia., Sched. 40 PVC, cut to 4.75" length, used to house PC Board





Standard Tee fitting, 2" dia., Sched. 40 PVC, used for objective housing and glare shield





Adapter (Reducer Bushing), 2" to 1" Schedule 40, used for barrel connector





Pipe, 1" dia., Sched. 40 PVC, cut to 5.25" length, used for barrel (optical tube)





Pipe union, 1" dia., Sched. 40 PVC, used for eyepiece coupler (mount)





Reducer Bushing, 1" Sch.40 to ½" NPT, used for eyepiece and filter mount (get the kind that is flat on the inside to make it easy to glue on the filter gel)





Yellow Filter: Gel, Roscolux #11 "Light Straw", cut out about 1" diameter disc.




Flat Black

Krylon "Ultra-Flat". The interior of the tube should be painted flat black.

The total cost for the circuit board components is about $33, and the total price for the tube assembly components (less the filter gel and the paint) is about $12. Thus, the total cost for the components for a complete unit is about $45.

Vendor sources:
DigiKey, 1-800-344-4539,
MPJ = Marlin P. Jones & Assoc., Inc., 1-800-652-6733,

Link to housing assembly image
Link to PVC Cutout Details image
Link to detail assembly image
Link to fully assembled Davis-Photometer image
(NOTE: This prototype was assembled for left-hand operation just for my personal preference.)

Links to guides on soldering:

My preference is to always use SN 63% / PB 37% (63/37) solder (this is also known as "eutectic" solder). Also, I always use RMA (Rosin - Mildly Activated) flux. Other fluxes that I would recommend are "Rosin - No Clean" and "Water Soluble", but I have no actual experience with them. Stay away from acid fluxes commonly used for plumbing. These are unsuited for electronics as the acid will lead to eventual corrosion and circuit failure.

Here is the assembly drawing for the LED brightness control circuit board:


Here are references to articles describing visual photometers and their use for measuring light pollution:

Gote Flodqvist, "A Simple Dark-Sky Meter", Sky and Telescope, February 2001, pp. 138-140.
Robert Pike and Richard Berry, "A Bright Future for the Night Sky", Sky and Telescope, February, 1978, pp. 126-129.
Richard L. Berry, "Light pollution: A losing battle?", Astronomy, Vol.5, No.9, Sept. 1977, pp.26-31.
Richard L. Berry, "Light Pollution in Southern Ontario", RASC Journal, Vol. 70, No. 3, Whole No. 540, June 1976, pp.97-115.
Upgren, A.R., "The measurement of night-sky brightness", Observatory, vol. 117, no. 1136, Feb. 1997, p.19-24.
Upgren, A.R., "Night Sky Brightness from Visual Observations II. A Visual Photometer", JAAVSO, 20, 244, 1991
J. Stock & A.D. Williams, Stars and Stellar Systems, Vol. II, ed. W. A. Hiltner (University of Chicago Press, Chicago), 1962, p.374


Report on use of light pollution meter
International Dark-sky Association -
Radiometry vs. photometry FAQ =
The LED Museum
Marktech Optoelectronics
EE Expert Anton Kruger archive
Anton Kruger: Using an LED as a Photodiode
Anton Kruger: How do LEDs Work?
Anton Kruger: When Can LEDs Replace Incandescent Lamps?
Anton Kruger: LED Efficacies
Anton Kruger: LED Resources on the Web

Prototype Circuit

When I performed a computer simulation of Mr. Flodquist's circuit, I found that the full range of LED currents that he showed in his table of measurements was controlled using merely 4% of the range of his brightness control, which is why he used a multi-turn potentiometer. My circuit adds a transistor, to add feedback, and two resistors, to tailor the range of the control, with the result that a single-turn potentiometer gives a control range of about 2 to 30 microamps. I installed my circuit in a small plastic box, and mounted banana plugs on the bottom of the box spaced so that they would plug directly into the DMM. This combines the controller with the DMM, so there is one less thing to hold. The red LED pilot lamp was mounted so that it provides illumination of the DMM readout.

My prototype circuit is shown below:

LED1 = LED, T-13/4, Diffused, Green, 565 nm, 15 mcd output at 30mA (e.g., Panasonic LN31GPHL)
LED2 = LED, T-13/4, Diffused, Red, 700 nm, 3 mcd output at 25mA (e.g., Panasonic LN21RPHL)
Q1, Q2 = NPN small-signal general purpose transistor (e.g., PN2222)
DMM = Digital Multimeter, 200 microamp scale
J1 = Panel jack, two contacts (e.g., RCA jack)
P1 = Cable plug, two contacts (e.g., RCA plug)
P2 = Banana plug, red
P3 = Banana plug, black
R1 = 470 Ohm, 1/4W, 10%
R2 = 33k Ohm, 1/4W, 10%
R3 = 50k Ohm audio taper potentiometer (e.g., Philmore No. PC85)
R4 = 22k Ohm, 1/4W, 10%
R5, R7 = 100k Ohm, 1/4W, 10%
R6 = 1k Ohm, 1/4W, 10%
S1 = SPDT toggle switch (e.g., RS 275-603A)
Battery = 9 volt rectangular (e.g., MN1604)

Instead of building a tube assembly to mount the green LED, I attached an LED to Art Upgren’s light pollution meter so that I can easily compare the two techniques. The main thing I lack is a method of calibrating the meter (this could also be considered as training the operator). I hope to enlist Dr. Morris and Dr. Upgren to help me with this. Then, I will ask the IDA to promote the widespread use of the meter. The reception of the combination "pitch-black meter" and light pollution meter was very encouraging at the April 2001 meeting of the SA-IDA.

Post to Darksky list, May 9, 2001:

Whither sky brightness measurement?

The IDA website,, has a page describing a forum on measuring sky brightness that took place in early 2000, but there is no link to proceedings or results from this conference. Please forgive me if this is an old subject for this list, but I have only recently joined. The reason for my inquiry is to gauge the interest in the use of a visual photometer for light pollution measurement, such as the one described in the February 2001 issue of Sky and Telescope magazine.

I am a member of the International Dark-sky Association ( and the Southern-Arizona section of IDA, and at the March SA-IDA section meeting, which just happens to be held at the IDA HQ in Tucson, Liz Alvarez showed us a couple of interesting instruments which she had recently received. I immediately recognized them as being some kind if light pollution meters (since the Pitch-Black Meter described in an article in February Sky&Telescope magazine was still fresh on my mind). Liz then asked me to take them home and make an evaluation of them for her, and start taking measurements with them, but she did not have any documentation to give me on them. One of them was from Dr. Steven Morris at L.A. Harbor College, and I have contacted him about it. Liz told me the other instrument was from Dr. Arthur Upgren of Wesleyan University, and I have been in contact with him, also. For background, I am not an astronomer, except in the amateur category, but I am an Electrical Engineer and a radio amateur, so I am more comfortable with electronics than I am with optics or mechanics.

I have built a "Pitch-Black Meter" circuit such as the one Gote Flodqvist described in Sky&Telescope, February 2001. Of course, being an electronics designer myself, I have designed a better circuit for controlling the LED brightness. Mr. Flodquist's concept of controlling the current of the LED is in my opinion a very good idea, I just think that his circuit is not the best for achieving that control. Additionally, Mr. Flodquist did not present a method of calibration for his PBM. I hope to design a light pollution meter that is simple and cheap to build, and is easy to calibrate.

Chronological Narrative:

Dec. 28, 2001: My printed circuit board layout is complete, and the artwork was mailed out to the fabrication shop.

Jan. 7, 2002: Debate on the color of the comparison source and eyepiece filter. Cliff Haas said:

I do have some reservations about using a green light source for this task, however. It is similar to using an apple to measure or describe an orange. Because the frequencies of the comparison presented (green LED) are not matched to the source it seems unlikely that a significant correlation of brightness can be established. I would think a white LED passed through a filter would provide a more reasonable comparison source.
Better yet might be to use a red, green, and blue LED array that is reflected onto a planar surface inside of the observation tube with each LED attenuated to match the frequency of light and somehow measure that composite result.

My reply to Cliff:

For the initial units, the readout will be in microamperes of current through the LED. Once we have a reliable calibration of the LED, then we can convert milliamperes to something more meaningful. The LED wavelength does not need to be matched to the source if it is matched to the peephole filter. The filter gel I used for the peephole is matched to the wavelength of the LED emission. This is similar to the old unit, where the yellow emission of the incandescent bulb was matched to a yellow filter at the peephole. I very carefully chose a gel that had a nice peak at 565nm, which is the emission wavelength of the LED I chose.

Using a white LED with a peephole filter would be a waste, since the white LED is not a broadband source, such as an incandescent bulb, but is instead three discreet sources: red, green and blue. A peephole filter would just filter out two of them, unless the wrong filter (say, yellow) was chosen which would filter out all three. Now, if you think a different wavelength would be more appropriate, we have several to chose from with LEDs, along with my suggestion for matching filter gels:
Red, 700nm: Roscolux #19, "Fire"
Orange, 630nm: Roscolux #23, "Orange"; or Roscolux #20, "Medium Amber"
Amber, 590nm: Roscolux #11, "Light Straw"; or Roscolux #20, "Medium Amber"
Green, 565nm: Roscolux #388 "Gaslight Green"
Blue, 470nm: Roscolux #363, "Aquamarine"

(I have since discovered (2/28/02) that the most popular type of white LED uses a blue LED which excites a yellow phosphor that is on top of it. This produces a spectrum that has many similarities to the spectrum of a typical incandescent lamp, the main exception being the sharp spectral line in the blue, about 470nm.)

My impression from reading Richard Berry's articles was that calibration was achieved by training the user using a device that produced a known illuminance. The operator trained himself with the visual photometer by trying to achieve repeatable measurements using different settings of the calibration device. He then recorded his photometer settings versus the calibration device settings, and thus the readings on the visual photometer are completely arbitrary. Although with additional complexity (and cost) this calibration could be incorporated into the readout of the visual photometer. The main thing I am hoping you and Art can provide me is the optical design of this calibration device. Next, we need a measurement program, a database to collect the data, algorithms to analyze the data, and reports (papers) to communicate the results.

Jan. 26, 2002: Cliff Haas objected that the 5mm LED originally used was too small. When placed at the objective end of the tube, it was practically a point source, and that trying to compare a point source with a diffuse source (i.e., the sky glow) was an apples-oranges comparison, and not scientifically valid. I suggested that we use a larger LED ("jumbo" LEDs are 10mm in diameter, and still inexpensive). Cliff had another solution, in the form of a more sophisticated optical tube design. My recommendation is to use a 10mm diameter green LED, such as the Radio Shack 279-215, or the Lumex SSL-LX100133XGC (DigiKey catalog no. 67-1400-ND).

Jan. 30, 2002: Cliff Haas sends information about his new optical tube design. I will add his diagram after I secure his permission. Here are his comments on the design:

Thanks for your note and comments. I got the feeling you may not have understood the new design I developed. It does not allow any view of the LEDs in the radial array. They are completely out of sight. Only results of their luminance is visible.
The new comparative photometer design allows a full unobstructed view that reflects diffuse luminance back to the eye without obscuring the view or distracting the eye from its path of vision. The luminance is viewed instead of with foveal photoreceptors (cones) with the parafoveal photoreceptors (cones and rods just outside of the fovea).
Included is an image showing the new design. Three baffles have distinct functions. The first baffle has a large aperture and holds the radial array of 6 LEDs. The second baffle has a smaller aperture tuned to the maximum angular path of vision from the eye through the end of the tube. The first baffle is the illuminating baffle and the second is the illuminated baffle.
The area formed between the first two baffles is the luminance chamber which is painted matte white to maximize reflected luminance within the chamber. These inter-refections will cause a very even diffuse luminance to result on the rear face of the illuminated baffle without having any direct light source to step down the pupil size. This is why more luminance may be required and the reason I asked about a trim pot on the circuit board to increase the range.
The reason for an unobstructed optical chamber is to maximize the amount of light entering the eye. It should allow a more accurate assessment with a lighter package than previous designs.
Please send me your mailing address so I can send you a sample of the illuminating baffle for the 6 LED array so you can wire it up for testing. My hope is the original circuitry can facilitate control in the fashion it does for the unit you sent. I'm curious what the candela output was for the LED your design employed?
Cliff Haas

Chair IDA Traveling Exhibit Committee

Feb. 1, 2002: Excerpt from my reply to Cliff Haas:

To answer your question about the LED, the specification sheet lists the optical output, Io = 15.0 mcd, when full rated current is applied (If = 30 milliamperes). Different colors have different ratings, and also whether the lens is clear or diffused.
For example, for T-1-3/4, 5mm dia.:
Red diffused at 25mA, Io = 3.0mcd, peak wavelength = 700nm
Red clear at 25mA, Io = 5.0mcd, peak wavelength = 700nm
Amber diffused at 30mA, Io = 8.0mcd, peak wavelength = 590nm
Amber clear at 30mA, Io = 20.0mcd, peak wavelength = 590nm
Green diffused at 30mA, Io = 15.0mcd, peak wavelength = 565nm
Green clear at 30mA, Io = 20.0mcd, peak wavelength = 565nm
My circuit applies a maximum output of about 1/1000 this level, or 30 microamperes. The light output seems to drop below my threshold of seeing at about 3 microamperes for the green LED. The output versus current is not specified, but it should be close to linearly proportional for current in versus output power (watts). Maybe you could estimate the optical output range required from each LED in your photometer design, and from that I could estimate the current drive needed and make the necessary modifications to my control circuit.

Feb. 6, 2002: I have received the first batch of ten printed circuit boards, and I find I goofed on the footprint for the brightness control potentiometer. Two of the pads are too small, and it was placed 180 degrees from what I intended. It could still be used, but the pot will have to be mounted manually, involving drilling new holes and soldering wires to the pins.

Feb. 10, 2002: Cliff Haas sends the following email:

I completed assembling a working comparative photometer that measures night sky brightness. This is the new design that I developed last month. It uses Avery's controller to drive 6 white LEDs and operates as well as expected. It has not made any comparison readings yet. I am still seeking a piece of photographic gel to filter the frequency of light approaching the eye. (See the attached image)
The luminance chamber works very well and provides a soft and even diffuse comparison source that does not require the eye to change focus at all. A clear unobstructed 7 degree view of field is offered to the observer.
Maybe somehow we can establish a method to accurately associate milliamp readings of current drain on a digital voltmeter to a 7 degree section of the sky? The unobstructed seven degree view was created to allow the fraction of a steradian used in calculating standard lighting metrics to be used for these purposes also.
Clear skies,
Cliff Haas

Feb. 11, 2002: Part of my reply to Cliff Haas:
I have thought up a calibration method for your photometer design. What you need is a plate or disc of matte white to cover the hole in the illuminated baffle. Then, you can place a calibrated light meter at the eyepiece and directly calibrate the illumination versus LED current. Ideally, each and every photometer would be individually calibrated, both at 25 deg. C and at 0 deg. C, but I think that we will find a high degree of consistency for a given make and model of LED.
In this application, I think white LEDs might make sense, although, if you do add a gel or filter at the eye piece, you should then use a monochromatic LED with a matching filter or gel. Look at my previous email where I list the wavelengths available in LEDs, and pick the one you think is most appropriate. I have already selected a gel appropriate for the green LED at 565nm, and I can do the same at the other wavelengths.
Looking at your concept for using indirect illumination from the LED, instead of looking directly at it, I felt that greater current should be necessary to drive higher levels of light output from the LEDs than I allowed for in the circuit I built that you have now. I have done some computer simulation, and increasing the available current is a simple matter of changing two or three resistor values. The main thing I need to know is the maximum current level desired.

Feb. 15, 2002: My additional comments to Cliff Haas:
After additional computer simulation of my LED controller circuit, I have convinced myself that it should work just fine with a series string of six LEDs in place of the original single LED. [In practice, three LEDs in series is the limit of usefulness] But I see the maximum current is now just 25 microamps. Increasing the maximum LED current is a simple matter of changing a single resistor value, R4. For example, changing the present value of 22K to 5.6K will increase the maximum current from 25 microamps to 100 microamps. Please let me know if you find this change useful.
I am working on a new PC board layout to better accommodate the new optical tube design. I envision a rectangular box with a digital readout on one face, and an on/off switch and brightness control on another face. Do you think this should be mounted on one side of the tube, or set up like a pistol-grip on the bottom? Please send me your recommendation.
Other suggestions for ergonomic improvements I have come up with include:
A diagonal mirror, either at the objective end or the eyepiece end, so that the sky near the zenith may be observed more comfortably.
Various schemes for handles, pistol grips, or even a rifle stock, to increase comfort of use.
Provisions for mounting to a camera tripod (1/4-20 socket on the bottom).
Make the eyepiece opening 2" or 1.25" diameter to permit use of standard telescope accessories.
An elevation angle indicator (e.g., a protractor and a plum bob).

Feb. 16, 2002: I have refined my design for an optical tube assembly using my PC board and 2" diameter schedule 40 PVC pipe and fittings. Details are in the text of my description of the PC board-based unit.

Mar. 16, 2002: Here are the topics I hope to discuss at the IDA Annual Meeting in Tucson:

Light Pollution Meter, or Pitch-Black Meter, or Visual Comparison Photometer for measuring sky brightness.

Outstanding Issues:

Calibration (Method, Standards, Units, Fixtures); Photometer Type (Direct View of Source versus Indirect View of Source); Comparison Source at Center Versus Rim; Source Type and Color (LED: Red, Yellow, Green, Blue or White; Incandescent); Filter: Needed? If needed, what type and color?

Mar. 19, 2002: The IDA Annual Meeting in Tucson went very well. Art Upgren, Cliff Haas and I manned a table in the Exhibitors area to show off our visual photometer prototypes, and were given an easel to show our "poster paper" literature. We also had a number of useful conversations with each other and with many other attendees. Here are some highlights:

1) An eyepiece filter is sometimes necessary, but not always. At very low light levels, such as measuring the brightness of a dark-sky site, color becomes imperceptible. My own personal experience is that I couldn’t perceive any difference between using white versus green LEDs. Cliff, on the other hand, was testing his prototype at a very polluted source, and the color contrast he perceived between the white LEDs he was using and the HPS-orange glow he could see in his sky made it difficult to perform a brightness comparison. Because of this, we decided to try out using Amber/Yellow for the comparison source, and for the eyepiece we will use Roscolux #11, "Light Straw", gel.

2) Bob Clear came up with an alternative design which might settle the rim versus center comparison source debate. He has seen some visual comparison instrument designs that used a left-right separation, that is, the field of view is split into to half discs separated (bisected) along the vertical diameter. This would work like Cliff’s indirect source photometer, except that instead of illuminating a visual rim, a half-disc on either the left or right side of the FOV would be illuminated. In order to maintain the same angular area of sky equivalent to a 7 degree field of view, the diameter of the apparent disc would be increased by the square root of two. This is so that the apparent FOV area, PI times the radius squared, will be twice that of the original 7 degree disc FOV, and thus the actual sky viewed will be the same for both instruments. I will work on creating a prototype of this instrument. I will likely use three amber LEDs in series to provide the illumination.

3) I built a "generic" LED controller in time for the IDA Annual Meeting. This one uses a red LED display, so it is self-illuminated, and has adjustments for minimum and maximum current. I found that there is almost no interaction between the minimum and maximum adjustments. That is, the value of R2 sets the minimum current, and R4 sets the maximum current. I used a 20K multiturn rectangular trimmer at both R2 and R4, but added a 10K fixed resistor in series with R2 so that the range would be 10K to 30K. Also, I installed a linear potentiometer for R3 (brightness control), as opposed to the audio-logarithmic taper pot that I used before. In actual use, I found that the linear taper was more "pleasing". Cliff liked it so much that he wanted to take it with him for use with his prototype, so I exchanged it for my original prototype, and he took the new one back too CT with him.

4) I found that my circuit will not work well with more than 3 LEDs in series. My version of the Haas photometer used six LEDs configured as two parallel sets of 3 LEDs in series.

October 8, 2002: I am still working on updating the printed circuit layout. Joe Frannea has made some wonderful images of the prototype Davis-photometer, and I have added links to the web page.

December 20, 2002: The Southern Arizona Section of the IDA has decided to pay for the materials to build five units. I have received 10 of the new, updated PC boards, and I have ordered and received five panel meters. I have given Byron Skinner three of the new boards with component kits and three of the panel meters, which he will assemble. I will assemble two of the boards by the end of the year. I would like to find somebody to build up the optical tube assemblies and meter housings for five units. With a little effort, we can have five units ready by the next IDA annual meeting.
Byron and I have discovered a few minor errors on the new PCB. These will be easy for me to fix before we try another batch.

March 20, 2003: The five units are finished! Joe Frannea acquired the parts for the optical tubes, and did a wonderful job of making the barrel, the circuit housing, and even figured out how to fabricate a tripod mount in the battery holder cap. Byron Skinner assembled boards, mounted meters in the boxes, and even purchased boxes and LEDs on a trip to California when local outlets didn't have stock. I completed assembly of the five units in time for the IDA Annual Meeting, and put together a poster paper/presentation.

April 11, 2003: Email reply to Don Higgens concerning use of darksky meter:
Sorry it has taken so long to reply. Your questions indicate to me that I need to write up a "Users Guide" for the darksky meter. Right now, the best I can do is to direct you to the to Sky&Telescope articles I reference on my LPMeter web page. I will try to answer your specific questions, below.

At 01:27 PM 4/7/03 -0600, don higgins wrote:
>Hi Avery: Don is getting practice with the darksky meter in rural Cochise
>County, in Tucson, and at Vega-Bray Observatory.
>I am slowly learning the "difficulties" of having one person do everything;
>however the comparitive results are giving numbers to
>what the eye already knows.
>For now: 1. do low temperatures influence the readings = cold of night?

There will be some temperature effect, but it should not be a big deal except at extreme temperatures. The tightest temperature specification is for the digital meter:
Accuracy: +/-5% over 18C to 28C with RH<80%
The LED itself will change brightness with temperature at a given current, but over this temperature range the efect should be insignificant.
Now, operation should be possible over a wider temperature range: 32F to 122F (0C to 50C). Over this temperature range, there might be a noticable effect, but I doubt it will be significant in this application. However, I should perform a test to prove this.
If you anticipate the instrument will experience temperatures outside this range, you chould remove the batteries as they can make an awful mess...
>2. Can the meter tolerate having bright light sources in the field
>when we try to document the worst light pollution? That is should
>the field be "just sky", so as not to bias the overall map of an area's
>readings by having bright, direct light on the recorder?

Since this is a visual comparison photometer, the question is not what the meter can tolerate, but what YOU can tolerate. The instrument is intended to let you compare a diffuse light source to the LED area brightness, and so any point sources will be a distraction. It is up to you to decide if you can make the comparision between the diffuse sky versus the LED area without being distracted by bright points of light in the field of view.
>3. I would like to get better measurements at Vega-Bray, so if you two visit
>there this spring perhaps I could meet you there for a tuturial
>on use of your meter? That is we could learn together at VBO.
>Measurements of light pollution at VBO are part of a long campaign to
>protect the viewshed of the observatory from bad local lights.
>Just now, the baaad lights are winning!

Sure! We have not set a date, but I will let you know when we do.
>4. Andrew Cooper of TAAA has a CCD with which he says we can calibrate
>all 5 meters at a single session in Tucson requiring only a darkened room.
>Is that calibration necessary or already been done for the 5 units? Andrew
>is willing to help with the calibration, if you want to do it?

No, I haven't already done anything like this, and it sounds like a great idea, but the ultimate "calibration" is also training for the operator, that is, each user will have his unique personal "calibration". But, this will help with establishing a calibration method. I will try to contact Andy in the near future.
>5. Last question: when we do the initial "dark read" before beginning
>recording, how do we use that #? Is it simply written down as part of the
>other numbers, or do we do some "subtraction" immediately? What does the
>final data set for a location look like?

While we still lack a good calibration method, the best we can do is to record all data that might have some significance, and hope that a later calibration can be retroactively applied. I suspect that the calibration will require a more sophisticated algorithm than simply subtracting the dark current.
>6. Thanks for your continuing interest in this neat project!
>Perhaps I (or Byron) will take one to RTMC in California for Memorial
>Day weekend star party. Best Don

Yes, we should try to get one of us to make a presentation and demonstration there.
>7 PS: The measurement at the zenith is important to a lot of amateurs
>for comparing light polution. However, we rarely observe straight overhead,
>so I am favoring the "multi-pointed" survey of a particular
>locality to show the range across the sky, around the horizon, always
>without moon light.

This is exactly what was presented in the IDA Newsletter article I referenced:
Report on use of light pollution meter
This kind of data presentation would be immediately useful, even without calibration.