Towards a Cheap, Effective Light-Pollution Meter (warning long)

Sky & Telescope had an article a couple of years or so ago about this
very thing. It was a home-built device featuring basic optics and a
relatively simple electronic design. We constructed one and it worked
well. I saved a (scanned) copy of the article, but, alas, it doesn't
show the issue and date! I can e-mail the scans to you if you'd like,
as they total only about 300 KB.


Gary
Benmore Peak Observatory


On Sun, 7 Sep 2003 04:37:30 -0600, "PrisNo6"
wrote:

Dan McKenna wrote in message ...
About 20 years ago I built a sky meter that worked well for me ant it
was simple.
snip

Would you be willing to share any plans?

- Kurt

PrisNo6 wrote:

Dan McKenna wrote in message ...
About 20 years ago I built a sky meter that worked well for me ant it
was simple.
snip

Would you be willing to share any plans?

- Kurt

Kurt,

I would need to re do it with the parts that would be used for the final
device.
At the moment I am on another project that has dragons breathing down my
back side as motivation and so no time.

Dan

Dan McKenna wrote in message ...

Thanks. I you ever change your mind, I'm sure Tony or myself would be
interested. - Kurt

PrisNo6 wrote:
snip
Would you be willing to share any plans?
Kurt,
I would need to re do it with the parts that would be used for the final
device.
At the moment I am on another project that has dragons breathing down my
back side as motivation and so no time.

Dan

PrisNo6 wrote:

Dan McKenna wrote in message ...

Thanks. I you ever change your mind, I'm sure Tony or myself would be
interested. - Kurt


Hi Kurt,

I just tried to find the lowest cost 1 cm^2 PIN photo diode for the sky meter.
We can get a UDT plastic case device for about $40 per diode in quant of 10 pcs.

No free samples and the minimum order is $200

My guess is that the Op Amp would be in the $5.00 range.
The meter might be built around a small micro processor
with on board AD.

The Analog devices ADu812 at $8 might do the job
One would want to add a LCD display
and the option of some kind of port like serial or usb.

If you use a micro then the temperature calibration could happen
and the optical filters could also be used with calibrated results.

The other way would be to make a unit with a 0 to 5 volt output
and let the user measure the output. One might be able to use a LCD
volt meter $25 to $30 and do the rest analog, just a few dollars of parts.

If enough people are interested we could make a few devices to see how they work.

I guess we need at least to make 5 units and do to my limited time would need to
form a working group that would buy in and contribute to the production.

Here is a model:

We buy the photo diodes.
I design, test and calibrate the detector.
At that point we might find that the results are not good enough.
So every one gets there photo diode and that it. Not good

If it is good enough then some one else takes the design a produces a circuit board
and if we have a code head in the group we can add a micro.

Does this interest you ?

Dan

Steve wrote:

Dan McKenna wrote:

Hi Kurt,

I just tried to find the lowest cost 1 cm^2 PIN photo diode for the sky meter.
We can get a UDT plastic case device for about $40 per diode in quant of 10 pcs.

What device was it Dan ? Why a PIN type ? Speed isn't an issue. Are they
much lower noise than standard photodiodes ?


You are correct speed is not an issue
The PIN diode has lower leakage and so it presents a higher impedance necessary
for the detector in voltage mode if you want a large dynamic range.



No free samples and the minimum order is $200

My guess is that the Op Amp would be in the $5.00 range.


What spec ? A Fet input opamp is going to be less than 5 USD.


We need an electrometer grade opamp i.e. the lowest input current, bandwidth is not
important
in the voltage mode. The configuration is a voltage follower with gain.


The Analog devices ADu812 at $8 might do the job
One would want to add a LCD display
and the option of some kind of port like serial or usb.

If you use a micro then the temperature calibration could happen
With the ADUC812 or the newer Cygnal parts, there is an on-chip
temperature sensor too.


If it is good enough then some one else takes the design a produces a circuit board
and if we have a code head in the group we can add a micro.

I'll code it for you, on an 8052 core.


8052 sounds good.

Another approach would be not to use an A/D. instead the photo diode is used in current
mode
and the opamp is configured as a charge amplifier with pulsed reset.

This configuration would result in a charge to frequency converter.
For low light levels that would result from a dark sky blue filter and a small field of
view we
could switch from frequency to period measurement.
Now that we will use an 8052 you could do the period to log 2.5 conversion.
This might be the cheapest solution.

I also would try to incorporate a small led that would be used to check the gain.

If we use the charge amp we would want a lower capacitance photo diode in a reversed bias
mode.
The Op amp in this case would need to have a low input capacitance as well as current and
current noise.
Because of the need to have a short reset pulse for dynamic range considerations we then
need the at least a 10 Mhz bandwidth for the opamp.

This would mean that the price of the photo diode drops to the few dollar range as we
could
use a light meter photo diode. I think Hamamatsu makes the best one. They have a diode
with
a built in filter to reject IR

The diode is cheap enough to have a two channel job that would measure the sky for
unfiltered ccd exposures. i.e. measure OH sky glow with the non filtered pin diode

Doesn't TI make a light to frequency detector ?
Has any one tried this on the night sky ?
I will look it up.

Dan


Steve

--
Steve Taylor
Technical Director
Astronomy Centre
http://www.astronomycentre.org.uk

Dan McKenna wrote:

What device was it Dan ? Why a PIN type ? Speed isn't an issue. Are they
much lower noise than standard photodiodes ?



You are correct speed is not an issue
The PIN diode has lower leakage and so it presents a higher impedance necessary
for the detector in voltage mode if you want a large dynamic range.


Hi Dan,

Current mode would eliminate the prohlem wouldn't it ? so perhaps we can
throw more money on the op-amp ? I take it you're thinking of a
Burr-Brown (TI) device ?

This configuration would result in a charge to frequency converter.
For low light levels that would result from a dark sky blue filter and a small field of
view we
could switch from frequency to period measurement.
Now that we will use an 8052 you could do the period to log 2.5 conversion.
This might be the cheapest solution.
Very cunning. The 8052 derivatives with the newer timer units work very
well.


Doesn't TI make a light to frequency detector ?
Has any one tried this on the night sky ?
I will look it up.

They do (though they aren't called TI in THAT division anymore AFAIK),
but I never tried to push the sensitivity too high - I was fighting it
DOWN at the time !
Best regards

Steve
--
Steve Taylor
Technical Director
Astronomy Centre
http://www.astronomycentre.org.uk