As it is with lasers, there are a large number of different photodetectors available. There are photodiodes, phototubes, photomultiplier tubes, phototransistors, photo field effect transistors, photovoltaic cells (solar cells), photoresistive cells, avalanche photo diodes, photodarlington transistors, even photo traveling wave tubes! We won't waste time exploring in detail some of the exotic kinds. The devices available with good performance at a reasonable cost are our main concern.
Next to cost, there are three other important parameters of concern. 1) Spectral response. Does the device function well at the wavelength of light we are dealing with? Presumably we will be mainly concerned with the red output of either a Diode or Helium Neon laser. 2) Sensitivity. Somewhat tied to spectral response at a given wavelength. The light sensitivities among the different kinds of sensors vary by at least a factor of 1000. 3) Speed. High speed is mainly a concern with short pulse width applications such as receiving light generated by a pulsed laser. Speed is not as much of a concern with audio or MCW modulated continuous output lasers.
Photovoltaic cells, or solar cells as they are commonly known, are not considered to be especially good for communications work. This is not what they are optimized for. They are fairly inexpensive but their speed is of question as their large area makes for large capacitance. Their "capture area" may very large, but this is not an advantage for situations where lenses or mirrors are used to gather the light. The spectral response is good for the portion of the spectrum that we are interested in, but this also means they are easily overloaded by sunlight.
Photoresistors have been around for a long time. Several kinds are available at very low cost. While the spectral response of the Cadmium Sulfide (CdS) photocell is very good at red light, they are considered far too slow for detecting modulated light.
In the category of phototransistors, there are bipolar phototransistors, photodarlington transistors, and photo Field Effect transistors (photo FET). Phototransistors are more sensitive than photodiodes as they have gain. This gain is achieved at the price of speed and dynamic range.
Bipolar phototransistors are inexpensive and have response times of 2 to 20 microseconds and good red sensitivity. They can be saturated by too much light and care should be taken to prevent this by optical shielding and filtering.
The photodarlington transistor contains a light sensitive transistor as in the bipolar phototransistor and an additional transistor that amplifies the output of the first one. These devices are very light sensitive. Minimum sensitivities are in the area of 6 to 24 mA/mW/cm2 for tungsten light. Speed is correspondingly slower, in the 10 to 100 microsecond range. As with phototransistors, photodarlingtons may have a base lead which is usually left unused. The Motorola MRD360 photodarlington is one of the best units with a minimum sensitivity of 24 mA/mW/cm2, a typical rise time of 15 microseconds, and a typical fall time of 65 microseconds. This device sells for under $3.
Common photodiodes are inexpensive, fast, and with good spectral response. They do not have the internal gain that photomultiplier tubes or avalanche photodiodes do, but much of that can be made up by use of a properly designed amplifier. Excellent photodiodes are available for $2 to $5 from sources like Newark and Digi-Key. Photodiodes generate a current proportional to the light striking them. Note that it is a current, not a voltage. The voltage generated by a photodiode is not linear with the light input. A current-to- voltage converter known as a transimpedance amplifier is used to convert the current output of a photodiode to a voltage. The current generated by a photodiode ranges from picoamps to milliamps. A transimpedance amplifier is as simple as one op amp with one feedback resistor. Not only does a current-to- voltage conversion take place in a transimpedance amplifier (e.g.: 1 milliamp in results in one millivolt out), but amplification occurs too. The amplification of the conversion is determined by the feedback resistor value: voltage out = current in times the feedback resistor value. Thus, an extremely tiny 1 nanoamp input (1 x 10-9 Amp) times a 100 megohm feedback resistance (1 x 108) results in 0.1 volt out - plenty to drive a speaker amplifier.
Until recently, I have done all of my Amateur Radio laser work, including record DX shots, with photomultiplier receivers. The Helium Cadmium laser record was also set with PMT receivers. I wish now to concentrate my efforts on photodiode receivers for several reasons: 1) I have learned how to design photodiode receivers that perform extremely well. 2) Photomultiplier tubes have drawbacks such as a high voltage power supply requirement, poor red sensitivity, and the cost of new tubes. 3) I feel that some Amateurs might be turned off by the drawbacks of PMTs, despite their superior performance. 4) I like to make equipment that is readily reproducible by others - and this is easier to do with simple devices like photodiodes than it is with PMTs. Nevertheless, because of their superiority in the area of low-noise high- gain, and their successful history of use in Amateur Radio laser work, photomultipliers deserve some coverage here.
Photomultiplier tubes are large and fragile, they require
high voltage power supplies, they cost more than most of the
common solid state detectors, they can be hard to find, and
their spectral response is usually not good at red. Why use
them? They have gain in the millions. Sensitivities
typically range from thousands to tens of thousands A/W. The
gain is developed inside of the tube and without excessive
noise. Rise times of 3 nanoseconds are typical. PMTs can
detect extremely low levels of light and are linear over wide
ranges of light intensity. The old and popular 931A PMT is
linear within 3% over a change in light intensity of 107.
The Avalanche diode is the closest solid state equivalent,
but it is even more costly, temperature sensitive, and
requires a highly regulated high voltage power supply. If
Avalanche photodiodes ever become low priced items then they
may have the advantage over the PMT. But for now, the
"hottest" optical front end is the PMT. Common PMTs look
similar to a glass envelope octal base radio tube. They
typically have eleven or more base pins depending on the
number of stages. There is no filament. Light enters a side
or end window and strikes a Photocathode. It is coated with
chemicals that determine the portion of the spectrum that the
tube will function best at. For every 10 to 10,000 photons
that strike the Photocathode, one electron is released from
it. This electron is attracted to a nearby element that is
energized about 200 volts higher than the Photocathode. This
element is called a Dynode. The striking electron dislodges
two or more electrons from the Dynode. This process
continues for several more Dynodes, one after another, each
Dynode biased about 100 volts above the last. This process
results in gains to 108. The electrons emitted from the last
Dynode are collected by an Anode that serves as the output
terminal. Most often a simple resistor voltage divider
circuit is used to power the tube. Negative potential is
applied to the Photocathode. A resistor connects the
Photocathode to the first Dynode. Each succeeding Dynode is
connected to the last by a resistor. The positive voltage is
applied to the end of the string. The resistors connecting
the Dynodes are typically 10K to 500K ohms. The resistor
between the Photocathode and the first Dynode is usually
twice that value. The voltage applied to the divider is on
the order of 750 to 1500 volts with current consumption
around 1 mA. The 931A and the 1P21 are two of the more
common PMTs. These are nine stage tubes as they contain nine
Dynodes. They require an eleven pin socket. Light enters
the side of the tube to strike a Photocathode with a surface
area of about 1/3 inch. The peak sensitivity is at 400 nm
and is unfortunately less that 10% at 633 nm (HeNe laser).
At 1000 volts supply voltage, the current amplification for
the 931A is 800,000. The Allied Electronics 1994 catalog
lists the 931A at $63. I have purchased many used tubes at
swap meets for $1 each.