EELE 482 Lab #10

Lab #10

Detector Characteristics and Noise

Contents:

1. Summary of measurements 2

2. Introduction 2

3. Measurements 3

Detector Characteristics and Noise Page 1 last modified 11/26/2018 5:58 PM

EELE 482 Lab #10

1. Summary of measurements
Photodiode Characteristics
1. Measure the responsivity and quantum efficiency of a silicon photodiode at 633 nm
(using the HeNe laser) and at 468 nm (using an LED). We’ll use the New Focus
detector/amplifier for this measurement, and become familiar with its properties.
Noise Measurements
2. Measure the shot noise from a flashlight for several illumination levels using the New
Focus detector and the spectrum analyzer.
3. Measure the relative intensity noise of the 637 nm laser diode and the HeNe laser
using the New Focus detector and the spectrum analyzer.

2. Introduction

The purpose of these experiments is to familiarize you with the properties of

semiconductor photodiode radiation detectors, and the use of those detectors to
measure optical signals and noise. You will measure responsivity and quantum
efficiency for a silicon detector. Using the New Focus amplified detector and a
spectrum analyzer you will measure the intensity noise of a flashlight (shot noise) and of
a laser diode (which has additional noise such as partition noise and carrier
generation/recombination noise, which exist in excess of the shot noise). This noise is
often specified as relative intensity noise, or RIN, meaning the noise power relative to
the output DC power of the device.

We are using a silicon photodetector in the photoconductive mode. In this mode of

operation, there is a reverse bias across the junction of the diode. The output current is
proportional to input power over a large dynamic range (7 to 8 orders of magnitude),
limited at the low end by detector and amplifier noise, and at the high end by detector
saturation effects. A small area p-i-n photodiode may have a useful bandwidth in
excess of 10 GHz, making these detectors extremely versatile and widely used in
photonic systems.

The wavelength response of the photodetector is determined by the optical absorption

properties of silicon and the geometry of the device. A typical responsivity curve is
shown below for a silicon photodiode. Similar curves covering different frequency
ranges are obtained for devices made from other semiconductor materials. Important
materials for the near infrared, including the communication bands near 1.3 and 1.5 m,
are Germanium and Indium Gallium Arsenide. It is important to note the strong
dependence of the responsivity on wavelength. For absolute power measurements you
must calibrate your detector at the wavelength of interest.

Detector Characteristics and Noise Page 2 last modified 11/26/2018 5:58 PM

EELE 482 Lab #10

Typical responsivity curve for a Si photodetector (green), from

https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=841265.

3. Measurements

Photodiode measurements

You can measure the responsivity of a silicon photodiode by illuminating the device with
a known amount of laser power and then measuring the output voltage developed
across a bias resistor. You must first characterize the light sources you will use by
measuring the power in the beams.

1) Set up both a HeNe laser and a 468 nm (or other wavelength) LED at your station.
Use a 50mm doublet to collimate the LED output. Place an iris behind the lens, and
adjust to a diameter of between 5 and 10 mm. With your power meter, measure the
optical power P of the HeNe beam and the collimated LED beam. Remember to
make the measurements with the correct wavelength setting on the power meter.
2) Use the New Focus 2001 detector/amplifier for these measurements. Measure the
responsivity (in A/W) at 633nm, using the HeNe laser, and at 468 nm, using the
LED.
To do this, you will need a lens to focus the beam onto the small detector (use a 50
or 75 mm focal length lens – singlet or doublet is fine; think about the magnification
for the LED system – the LED chip is about 400 um across, and its image must fit
within the 1mm2 detector area. What is the expected waist size for the HeNe
beam?). Set the gain for the amplifier so that the output signal registers a few volts
(not saturated). Record this voltage V, and note the overall gain G (the value from
the rotary gain knob times the 1x-3x multiplier set with the rocker switch). From the
datasheet, the transresistance of the input stage of the amplifier is R1=567 ohms.

Detector Characteristics and Noise Page 3 last modified 11/26/2018 5:58 PM

EELE 482 Lab #10

Calculate the responsivity as 𝑉/𝐺𝑅1 𝑃 [A/W]. From the responsivity, calculate the
quantum efficiency at these wavelengths.
3) The New Focus 2001 detector/amplifier specifies a responsivity, in [V/A], for the
silicon detector (you’ll find the chart in the datasheet, on the class webpage).
Compare your calculated responsivity to the value on the chart, for both wavelengths
tested.

Noise Measurements

See the Hints for the noise measurements note at the end of the section.

4) Use the New Focus detector and the RF spectrum analyzer to measure the shot
noise from a flashlight beam. Connect the output of the detector/amp to the input of
the spectrum analyzer, and also to a voltmeter in order to measure the DC level of
the detected signal. Note that your conversion factor for the detector is now
different, since the input impedance of the spectrum analyzer is 50 ohms (AC
coupled, meaning that at DC it is very high impedance), and the model 2001 has a
series output impedance of 16 ohms. You also are making a broadband (optical)
measurement rather than a measurement at a discrete wavelength, so that the
absolute calibration is no longer known. Make the measurement for several different
illumination levels, and plot the output noise density vs. DC voltage level, both
normalized for the gain settings used to make the measurements, using a log-log
axis.

Comment on the spectral properties (RF spectrum) of the noise you observe.

5) Measure the intensity noise of the laser diode using the same method as above.
You will be able to make relative intensity noise measurements if you use the
appropriate gain values to reflect the responsivity of the detector at your laser diode
wavelength. If you want to make an absolute noise measurement, you will need to
be careful to focus your calibrated beam so that all of the beam falls on the detector
active area. Then knowing the responsivity measured previously, and remembering
the 16 ohm series output impedance on the New Focus detector, you can convert
your noise voltage measurement to optical noise power. How many “times shot
noise” is the RIN you measured, for this particular operating point for the laser
diode?

Comment on the spectral nature (RF spectrum) of the diode noise.

6) Repeat the measurements to find the intensity noise of the HeNe laser. How does it
compare to the diode laser, and to the flashlight?

Hints for the noise measurements: The bandwidth of the New Focus detector is
highly dependent on the gain settings. The noise floor also is dependent on the gain
settings. In order to make meaningful measurements, you must select a high enough
gain in order for the shot noise to be dominant over the detector and amplifier noise,
Detector Characteristics and Noise Page 4 last modified 11/26/2018 5:58 PM
EELE 482 Lab #10

typically 102 or higher. If you make measurements over a reasonably wide dynamic
range, say two orders of magnitude for the optical signal, you will need to change the
gain setting in the middle of your measurements. Therefore, you should make your
measurements at sufficiently low frequency such that you remain in the pass-band of
the amplifier in all cases.

I recommend the following initial settings, and you can experiment further if you like.
Use gain settings on the New Focus detector of either 3x102 or 3x103 (or both in
sequence, with appropriately adjusted illumination levels). Watch the voltmeter, and
adjust the light intensity until you are not saturating the amp.

Set the spectrum analyzer to sweep from 0 to 100 kHz, with 1 kHz resolution bandwidth
and 1 kHz video bandwidth, and video averaging on and set to 10. Turn on the marker,
and enable ‘marker noise’ using the softkeys. The readout will be in dBm(Hz), meaning
decibels relative to 1 mW rms, in a 1 Hz bandwidth. To convert from dBm to volts, it is
necessary to assume a 50 ohm load, so that dBm = 10 log10 ((v2/50)/.001). Move the
marker to somewhere between 20 and 40 kHz, and make your reading of noise power.
Leave the marker at the same frequency for all comparative measurements.

Because shot noise varies as the square root of the signal, if you plot DC voltage vs.
noise voltage (corrected for the 50/(50+16) gain factor) on a log-log plot, you should see
a linear relationship with a slope of ½. If you want to verify absolute noise power, you
will need to make some assumptions about the average detector responsivity and
amplifier gain over the wavelength range of interest. Reasonable results obtain with
the following assumptions:

Amplifier transimpedance: (1.1x107 V/W)peak/(0.6 A/W)peak = 18x107 V/A (in highest

setting) (we assume response dominated by most favorable wavelengths)
DC gain = 18x107 V/A (in highest setting)
AC gain = 18x107 x 50/(50+16) (in highest setting)

Then you can calculate photocurrent = dc voltage / DC gain. Then calculate shot noise
current, multiply by AC gain and convert voltage to dBm(Hz) assuming a 50 ohm load.
See whether your measurements line up with the predicted shot noise values.

Detector Characteristics and Noise Page 5 last modified 11/26/2018 5:58 PM