Results for the C-SPAD for the NERC Space Geodesy Facility

Results for the Herstmonceux C-Spad detector
Authors Dr. Roger Wood, Dr. Graham Appleby and Philip Gibbs

Introduction

Initial test results

Temperature effects

Possible error sources when using C_SPAD

Picture gallery

Introduction (Top)

We have used Single-Photon Avalanche Diode (SPAD) detectors at the Herstmonceux SLR station since 1992. Their principal advantages over the traditional photo-multiplier tubes (PMTs) are the fast rise time of the avalanche to give good epoch timing, compactness, stability and robustness.
The disadvantages are that they are intrinsically noisier and the tiny chip makes alignment difficult. They also have time-walk effects (ie the response time is dependent on both the light pulse energy and the temperature of the device).
However, the effects of these drawbacks can be minimized at the the telescope by using neutral density filters to attenuate the energy of the return pulses at the level of single photons, and by frequent calibration. In routine operation the advantages far outweigh the drawbacks

The improvements of the C-SPAD over the previous SPAD are as follows

A larger chip for easier alignment.
The detector chip is cooled to -60C and is much less noisy.
The electronics are temperature stabilized to minimize response time variations.
A second imput channel has been added for which the emitted electrical pulse is delayed by an amount equal to the time-walk induced by the particular energy of the incoming photon pulse.

Initial tests of the C-SPAD performance (Top)

An extensive series of tests were carried out to investigate how the range to a local calibration target varied with the energy of the returning pulse. The return energy was estimated by monitoring the rate of true returns detected by the C-SPAD. (Return Energy)
The experiments were carried out on days when the outside temperature varied very little thus minimizing any temperature effects. Each experiment consisted of ranging to a ground target while controlling the return energy using neutral density filters and irises. At high pulse energies photon numbers were estimated by extrapolation from lower energies using the known ratios between the attenuations for the different neutral density filters and irises.

Figure 1a
Time walk for compensated and uncompensated channels of the C-SPAD before tuning

Figure 1b
Time walk after the C-SPAD was tuned for correct laser pulse length.

The results from these initial tests are shown above. Time-walk (cm) is plotted against estimated photons. The shape of the uncompensated channel follows theoretical expectation. Theoretical model It was quickly realised that the compensated channel had been set up using a short pulse. The electronics were retuned to match our longer pulse length and the behaviour for the compensated channel was then as expected.

Temperature effects (Top)

We already knew from our experience with previous SPADs that the calibration range showed clear correlation with temperature. Figure 2 shows an 8-day time series of calibration range against temperature for the uncompensated channel and a clear correlation can again be seen.

	Figure 2a Daily variation in the calibration range in mm.
	Figure 2b Daily variation in the outside temperature

The range variations for both channels on the C-SPAD were plotted against temperature as shown in figures 3a and 3b. As can be seen from the plots the slope for the two channels have the opposite sign. This tells us that the slope is not due to expansion of the telescope or target, but is in fact a function of the detector. As the slopes are relatively small the effects can be easily eliminated by frequent calibration,especially when temperatures are changing fast.

Figure 3a
Range variation plotted against outside temperature for uncompensated channel

Figure 3b
Range variation plotted against outside temperature for compensated channel

Possible error sources when using C_SPAD (Top)

Arming C_SPAD too close to track

Picture Gallery (Top)

CSPAD at Cassegrain focus of Herstmonceux system

Return Energy

For each firing of the laser only a single event can be recorded by the C-SPAD. This can be a true return, detector noise or daylight. We must distinguish between these different returns when calculating the true data return rate.
We compute the return rate from ranging sessions by counting the number of laser shots in a given time interval, normally 10 seconds. For each of these shots we check whether a noise event is detected before the true event could occur, which reduces by one the effective number of laser shots. From the number of true returns from the target we can compute the true return rate as a percentage of the number of effective laser shots.
Under normal working conditions the software works out the percentage return rate and maintains the system at single photon level by the use of neutral density filters.

For a detector with quantun efficiency qe, we can relate the return rate to the number n of photons reaching the detector from

rate = 100x(1-(1-qe)ⁿ).

For the C-SPAD we have qe=0.2. For standard calibration ranging this rate is maintained at about 10-15% by attenuating the outgoing beam, and by selecting the appropriate ND filter and iris so that n<=1.
We have determined the relationship between all of the NDs and irises based on return rates up to 100% return rate. These relationships enable us to estimate the return energy when the return rate is greater than 100%.

Back to Initial tests

Time walk model

We have estimates based on our own experiments and the manufacture details for each component of our system. (C-SPAD,Laser,Start-diode,SR620 timer) Quadratic combination of these values gives us an expected single-shot single-photon rms range precision very close to the experimental mean. Using these parameters, we have set up a model of the expected change of calibration value as a function of number of returning photons as described in detail in
Satellite Signatures in SLR Observations, NASA Conf. Proc 3214
Eighth International Workshop on Laser Ranging Instrumentation
G.M.Appleby
Back to time walk results

(Top)

Go back to SGF main page

	Figure 1a Time walk for compensated and uncompensated channels of the C-SPAD before tuning
	Figure 1b Time walk after the C-SPAD was tuned for correct laser pulse length.