FORTE Photodiode Detector

Overview of the FORTE Optical Lightning System

The FORTE satellite carries two instruments that comprise the Optical Lightning System (OLS). One of these sensors, the Lightning Location Sensor (LLS), employs a CCD array having an 80° field of view (FOV), a 10 Å passband filter centered on 777.4 nm, and the electronics necessary to trigger on transient optical events having pulse widths shorter than a few milliseconds. The LLS was designed to determine the geographic location of lightning events with 10-km accuracy and is strongly based on the design employed on NASA's satellite-based lightning imaging sensors [e.g., Christian et al., 1989; Christian and Latham, 1998]. A description of the LLS, including a comparison with FORTE photodiode detector observations, has been given by Suszcynsky et al. [2001].

The other OLS instrument is the photodiode detector (PDD). The PDD consists of an unfiltered, single-element silicon photodiode with 1 cm ² of sensing area. Figure 1 shows the PDD response versus wavelength from 400 to 1100 nm and shows that the peak response is near 850 nm. The PDD calibration assumes a spectrally averaged responsivity of 0.3325 A/W. A 6-inch-long (5-cm-long) sunshade provides a circular 80° FOV to approximately match that of the LLS. Both the LLS and PDD measure the intensity of incoming light in the OLS's FOV and trigger on impulsive events. The PDD provides the means to record the light intensity time history of lightning flashes occurring within the FOV of the instrument. In this paper, we exclusively describe the PDD and present data obtained with the PDD.

Photodiode Detector Sensitivity, Sampling, and Background Signal Compensation

The PDD employs 12-bit sampling to provide an effective dynamic range spanning >5 orders of magnitude. The 12-bit sampling allows for a sign bit and 11 bits of resolution. The total dynamic range is piece wise linear, as shown in Figure 2. The finest granularity in the PDD sensitivity is better than 10 ^-5 W m ^-2.

The analog PDD electronics continuously compensate for the slowly varying background signal, allowing the full dynamic range to be used for transient signals. The background compensation rate can be set into one of eight rate modes. These eight compensation rate modes are derived from two compensation modes (night or day) and four subordinate compensation rates (fastest, fast, slow, slowest). In this logic scheme the slowest compensation occurs in the night/slowest compensation mode/rate setting, and the most aggressive compensation occurs in the day/fastest setting. The slower compensation rates are typically employed when the satellite views a dark Earth. The faster compensation rates are intended for day/night terminator crossings, when the light level from the background scene changes rapidly. In all eight mode/rate settings, the compensation acts to return the signal level to zero but at different rates and only when the signal level exceeds 4.45 10 ^-4 W m ^-2. The compensation rate mode setting is programmed as a time-dependent parameter via uplink when FORTE is in view of the ground station (Albuquerque, New Mexico).

The compensation rate mode setting marginally affects the recorded optical lightning signal data. In the case of a Gaussian pulse having an amplitude greater than 4.45 10 ^-4 W m ^-2, the most aggressive compensation (day/fastest) will reduce the peak amplitude by 8% and the magnitude of the integrated signal by 10%, with no time shifting of the peak. In on-orbit tests of the most aggressive day/fastest compensation mode, an on-board light source was turned on in the PDD field of view with an amplitude of 1.4 10 ^-3 W m ^-2 (essentially a step function); the signal was fully compensated (to zero) within 3-ms. The least aggressive compensation mode is >2 orders of magnitude slower. For relatively fast rising (order 10 s), short-duration (order 100 s) optical lightning pulses we expect the most aggressive compensation to have only a small effect on the magnitude and shape of the signal. We expect the least aggressive compensation mode to have a negligible impact on the recorded signal and to provide optimal results when the lighting background is slowly varying.

A review of the nearly 2 million events detected and recorded by the PDD during the first 30 months of operation reveals that most of the events (62%) were obtained under the night/slow configuration, with only a mild variation in usage with respect to local time. In terms of percentage of detected events this was followed by the day/fastest configuration (23%) and then by the night/fastest mode (14%). While we do not expect the events collected in the latter two modes to have significantly distorted waveforms, we can nevertheless afford to restrict our consideration of the PDD data to only those events collected in the night/slow mode for the purpose of introducing the instrument and associated data.

Photodiode Detector Trigger Modes

The PDD employs four trigger modes: internal trigger mode, external trigger mode, slave mode, and test mode. The internal trigger mode allows for the autonomous detection of optical events occurring within the PDD field of view. In this paper, we present and discuss only those data generated by internal triggers. However, we note that the external mode allows the PDD to be triggered by either the radio frequency (RF) subsystem on board FORTE or by the LLS but is essentially like the internal trigger mode (described in the next paragraph). The slave mode allows the LLS to trigger the PDD, and a longer PDD time history is retained. The test mode allows satellite operators to diagnose the health of the PDD sensor.

In the case of the internal trigger mode the compensated photodiode signal is digitized every 15 s and inserted into a FIFO (first in, first out) storage buffer. When 32 samples have been written into the FIFO buffer, subsequent samples are then compared with the trigger level. The trigger level is a noise-riding threshold, where the trigger level is based upon a contemporary noise level calculation performed by the OLS processor once per second. The trigger level is derived by simply taking the product of an adjustable multiplier (typically 5) and the calculated PDD background RMS noise level provided by the OLS processor. If the digitized signal level exceeds the trigger level for a predetermined number (selectable between 0 and 31 and typically set to 5) of samples, then a trigger is generated. This duration test typically rejects triggers due to the passage of energetic particles through the photodiode that typically have lifetimes of <75 ms (five samples). Once a trigger is generated, another 96 samples are written into the FIFO buffer. An event message containing the 128 samples is then formatted, stored in the OLS memory and the PDD electronics are reset.

The time tag for the trigger is assigned based on a 2.098-MHz OLS clock, which is conditioned by a 1-Hz GPS-derived signal. We expect thermal variations and uncertainties in the OLS circuitry to introduce <25 s of variability into the assigned trigger time. The minimum intertrigger delay is dictated by the requirement to acquire the 96 posttrigger samples, to format the event message, to reset the electronics, and to acquire 32 new samples. In practice, we have found this intertrigger delay to be 4.4 ms (Figure 3). This delay is sufficiently small enough to allow the PDD to capture sequences of cloud to ground lightning strokes within a flash, provided the trigger criteria are satisfied (see, e.g., section 3.4).

One caveat to the capture of stroke sequences within a flash is that the PDD will temporarily disable itself until the next GPS-derived 1-Hz time hack if more than a specified number (usually 10) events are recorded within a specified time window (usually 40 ms). This rate limit is imposed to avoid filling the OLS memory with a rapid succession of transient events that occur due to glint off of spacecraft or ocean surfaces or from anomalous signals that may occur during day/night terminator crossings when the background scene is changing rapidly.

Next: Data Reduction Up: Optical observations Previous: Introduction