Re: [ATM] Dark Sky Meter

[Dominic-Luc Webb <dlwebb@canit.se>]

Interesting you want to use a PIN (I have a few words about that). I
am sorry I entered into this a bit late (many things going on over here).

What wavelength(s) did you have in mind?

What sensitivity did you want (why not a CCD or PMT)?

Did you want spatial information or are you confined to a fixed position?

Are you wanting to correlate with other types of data, like humidity,
altitude, atmospheric scintillation, etc?


I have tinkered with PINs and while I can get them to work, they
are not usually my first choice because the opamp stage is so
difficult to build. As Bob May pointed out, AC circuitry could
be helpful. There are some alternatives as I have mentioned, like
using a working sensor from a TV. You can always replace the
photodiode with another one if IR or some other property doesn't
suit your needs. Also, consider that there are inexpensive PINs
that come pre-packaged with an integrated opamp, although, there
is not much flexibility regarding wavelength. There are also diode
arrays of various sorts that come with integrated opamps that can
even provide limited spatial info. Ditto for avalanche photodiodes.
If you decide photodiodes are the correct route, you should give
APDs a thought, but consider that by all accounts, a very good
quality power supply will be needed. Then a warning....

The stated sensitivity for the SFH2030F, a common cheap PIN is
0.62 A/W at 830 nm (peak is actually 950 nm). The stated sensitivity
for an S2381, a very expensive avalanche PIN that claims to be many
times more sensitive than a common PIN is only .49 A/W at 800 nm (and
the peak really is 800 nm for this PIN). Conclusion, for 800 nm light,
the dirt cheap SFH2030F is actually more sensitive than the APD that
costs about 60 times more. For considerably less money than the APD,
you can get similar sensitivity from a diode array as might be used
in a spectrophotometer. In fact, your best bet might be to actually
find a surplus spectrophotometer. In this case, you can do a spectral
analysis. I do not believe a homemade detector would meet or exceed
what could be had from such an instrument. Even with a good detector,
you would need to build a very high quality signal processing circuit
that would include at least a well constructed opamp and power supply
and means of readout. I have an ancient spectrophotometer from a college
auction (Perkin Elmer, 5 bucks) that uses a prism and photocell
arrangement and this works better than any PIN detector I have built
myself. The readout could use some modernization, like computer interface
for A/D recording, but this is trivial circuitry compared to designing
and building a decent opamp circuit.

For the more serious folks that are able and willing to spend a
little more money, there are B&W CCD and PMT alternatives. Consider
that B&W CCDs have been used as diode arrays and that PMTs are very
fast. These both have extremely high sensitivity compared to even
the very best PINs and APDs. You can sometimes find a surplus
spectrophotometer that uses a PMT. This is what is found in the
top-of-the-line spectrophotometers. The intermediate price range
ones use diode arrays. Other items that could have these are
HPLC and protein purification detectors.

Dominic-Luc Webb



_______________________________________________
ATM mailing list http://www.atmlist.net/