Building a bolometer:

A bolometer detects the heat from absorbing light. This comes with limits sensitivity and speed, but works at wavelengths which are otherwise very difficult to detect, like thermal infrared.

At its simplest, a bolometer is simply a black object attached to a therometer. Typically a thermistor is used as the heat sensor, but I did not have any on hand, so I used a 1N4148 silicon diode. A diode's forward voltage falls with temperature, generally around 2 mV/Kelvin.

I covered the diode with soot (carbon) becuase it is a fairly good wideband light absorber, and built up a quick circuit test the principle:

After the diode cooled to ambient temperature, placing my finger on it dropped the output voltage by a good 30 mV. But, this simple circuit won't work as a bolometer: Even a tiny change in ambient temperature overpower the tiny effect from thermal radiation. To compensate, I added a second diode as a baseline temperatred reference:

There will be some difference even at the same temperature because the diodes will not be perfectly matched, but that can be zeroed out.

This is getting somewhere, but the diode bridge circuit still can't measure the actual amount of heat flowing into the sensor.

To do this, I added small heater to the sensor: a 1k resistor. The heater will be run in a feedback loop to maintain a constant temperature rise over ambient. This ensures that the total power flowing into the sensor is constant.

When illuminated, the amount the feedback loop reduces the heater power will be exactly the same as the optical input power.

With an opamp adjusting the heater power to keep a constant sensor temperature, the input power can be measured as the drop in applied heater power:

Just about any opamp will do, as long as it can suppy 5 mA of current. If yours has an offset-null adjustment, you could use it for the bias adjustment in place a potentiometer on the diode bridge.

I glued to the sensor diode (D1) to the heater, but taping them together works fine. Watch out for the photoelectric effect: glass diodes need to bes hielded from light. The thermal mass of the absorber/sensor/heater assembly should be as small as possible to keep the time constant reasonable.

The detector is quite sensitive to drafts, make sure to shield the diodes from any air currents. Commercial bolometers are often placed in a vacuum, but most common window materials are opaque to thermal infrared.

To make a measurement, first measure the baseline heater voltage with the sensor covered, then expose sensor to light or thermal radiation. After it reaches equilibrium (the output stops changing), measure the heater voltage again. Convert the 2 voltages to powers:

Power = Voltage2 / 1000 Ohms

... and subtract for the measured power.

Covering the sensor except for a small aperture it turns it into a terrible, 1 pixel thermal camera. Mine can detect the thermal radiation from a person from a few meters away, but it takes a good 20 seconds to respond.

It also works as a half decent laser power meter, but with a 1k heater resistor running off 5 volts, it will max out at just 25 mW. This could be improved using a smaller value resistor for the heater:

Using 100 ohms would let it measure up to 250 mW.

With a 50 ohm resistor attached to the sensor, it works as an RF power meter, but I can't get more then around 20 dB of dynamic range. On the plus side, it works from DC to daylight (literally).