The Advanced Camera for Surveys
  instrument > calibration > results > by item > detector > hrc > build1 >




Detectors

WFC4

Bias, Overscan
Gain, Noise

HRC1

Bias, Overscan
Gain, Noise
CTE

SBC

Darks
Flats
Throughput

Shutter

Shading, accuracy

PSF

Encircled energy

Internal Count Rates

HRC1, WFC4
SBC

FLASH

HRC1, WFC4

Dispersers

Grism
Prism

Stray Light

Light leak
Filter dust


ACS Home

Last updated
18 December 2001 10:37:28

Maintained by
martel


Printable version

HRC#1 : Gain, Noise, Linearity, Saturation

AUTHORS : A.R. Martel, G. Hartig, and M. Sirianni

GOAL :

The main goal is to measure the primary system gains of 1, 2, 4 and 8 e-/DN with the photon-transfer method as well as the read noise for each amplifier of the HRC flight build 1 detector. The full well of the detector can also be estimated.

LOCATION AND DATE :

A thorough set of flat fields at gains=1, 2, and 4 were acquired on 3 Apr 2001 at GSFC/SSDIF with SMSs JGCH32A and JGCH32B for all four amplifiers. During the Thermal Balance/Vacuum Campaign 3, additional internal flat fields were acquired at gains=1, 2, and 4 with the CCD basic monitor SMSs JTVH01A, JTVH01B, and JTVH01C in Cold Soak#1 (Jul 13) and in the subsequent calibration phase, JTVH01B (gain=2) and JTVH01D (gain=8) were executed (Jul 17). All the thermal vaccuum flat fields were acquired with amp C only and a bias offset of 3.

Pairs of bias frames were acquired as part of SMSs JGCH32A and JGCH32B as well as in the calibration phase of TB/TV 3 campaign on Jul 20 (gain=1, 2) and Jul 21 (gain=4, 8) for all amplifiers.

INSTRUMENT CONFIGURATION :

Same as WFC#4 : Gain, Linearity, Saturation.

METHOD :

The flat fields are acquired with the SMSs JGCH32A (gain=2) and JGCH32B (gain=1,4). The illumination is provided by the internal tungsten 4 lamp through the F555W filter. A brief description of the photon-transfer method is given in the Appendix of WFC#4 : Gain, Linearity, Saturation. A sequence of flat field pairs is acquired at each gain setting, over as wide a range in count levels as possible, from the read-out noise at low levels (usually at the shortest integration time of 0.1 sec), through the linear part of the curve, and up to non-linearities and saturation at the highest levels. Full datasets are acquired consecutively with each amplifier. The bias level of each image was subtracted from the virtual overscan. We find that nearly identical results are obtained when a full median bias frame, constructed from the bias frames acquired with the SMSs, is subtracted from each flat field and the residuals removed with the virtual overscan. The statistics are measured on ten 20x20 boxes distributed over each quadrant. The gain is measured from a linear fit to the photon-transfer curve. Since the on-orbit gain will be set to 2 e-/DN, SMS JGCH32A has the most complete coverage of the photon-transfer curve while SMS JGCH32B only samples a few points at gains of 1 and 4 for comparison.

The gains were also measured with the flat fields acquired in TB/TV 3 with the basic CCD monitor SMSs (amp C only) to verify those measured from the above SSDIF flat fields and in particular, to accurately measure gain=8. The read-out noise is derived from the imaging area of pairs of bias frames acquired either in SSDIF or in the dark environment of TB/TV3.

RESULTS :

The system gains and read-out noise of each amplifier for gain=1, 2, 4, and 8 e-/DN are listed in Tables 1 to 4 below. For comparison, the measurements for both the SSDIF and TB/TV 3 environments are tabulated. Plots of the photon-transfer curves are shown in Figs 1-3 (SSDIF) and 4-7 (TB/TV 3). Some points of interest :

Table 1 : Gain in e-/DN (SSDIF)

Amplifier Gain="1" Gain="2" Gain="4"
A 1.06 +/- 0.01 2.05 +/- 0.01 4.05 +/- 0.04
B 1.09 +/- 0.01 2.12 +/- 0.01 4.20 +/- 0.03
C 1.18 +/- 0.01 2.22 +/- 0.01 4.29 +/- 0.04
D 1.11 +/- 0.01 2.09 +/- 0.01 4.14 +/- 0.04
Average 1.105 +/- 0.005 2.118 +/- 0.007 4.17 +/- 0.02

Table 2 : Gain in e-/DN (TB/TV 3)

Amplifier Gain="1" Gain="2" Gain="4" Gain="8"
Jul 13 Jul 17
C 1.15 +/- 0.01 2.26 +/- 0.03 2.25 +/- 0.04 4.36 +/- 0.07 8.68 +/- 0.17


Table 3 : Read-Out Noise in e- (SSDIF)

Amplifier Gain="1" Gain="2" Gain="4"
A 5.33 +/- 0.26 4.93 +/- 0.30 6.18 +/- 0.28
B 4.54 +/- 0.22 5.00 +/- 0.66 5.85 +/- 0.18
C 4.52 +/- 0.18 4.68 +/- 0.17 5.71 +/- 0.25
D 4.87 +/- 0.24 4.69 +/- 0.21 6.17 +/- 0.28
Average 4.75 +/- 0.11 4.73 +/- 0.12 5.93 +/- 0.12

Note : The read noise is measured in DN and is converted to e- using the gains measured in SSDIF.

Table 4 : Read-Out Noise in e- (TB/TV 3)

Amplifier Gain="1" Gain="2" Gain="4" Gain="8"
A 4.41 +/- 0.16 4.60 +/- 0.19 5.65 +/- 0.25 -
B 4.17 +/- 0.13 4.40 +/- 0.13 4.40 +/- 0.11 -
C 4.51 +/- 0.16 4.74 +/- 0.18 4.60 +/- 0.07 12.01 +/- 0.57
D 4.70 +/- 0.15 4.98 +/- 0.19 5.78 +/- 0.18 -
Average 4.42 +/- 0.07 4.62 +/- 0.08 4.71 +/- 0.06 -

Note : The read noise is measured in DN and is converted to e- using the gains
measured in SSDIF (gains=1, 2, and 4) and in TB/TV 3 (gain=8 for amp C only).


CEI SPECIFICATIONS :

In STE-50, "Information End Item Specification (Part II)", Table 4-6 states that the requirement for the RMS noise per pixel, which includes both the read noise and noise from the dark current, is <4.5 e- and the goal is <3.5 e- for a reference integration time of 13 min (780 sec). The read noise was measured from the imaging area of the bias frames and is listed in Tables 3 and 4 above for different amplifier and gain combinations. The dark count rate measured in TB/TV 3 is approximately 8 e-/pix/hour, which amounts to ~1.7 e-/pix for a 780 sec exposure. The noise on the dark rate is taken as shot noise i.e. the square root of the dark current (as in the ACS Exposure Time Calculator, see p. 103 of the ACS user manual), so the total noise is the square root of the sum of the read noise squared and the dark rate (1.7 e-/pix). From Table 4, we find that only amp B at gain=1 meets the CEI specification (total noise of 4.37 e-). None of the other amplifier/gain combinations meet the noise requirements, including the default on-orbit configuration, amp C at gain=2. In the same CEI table, the full well is specifed as >99900 e-. Our measured full well of ~135000 e- meets this requirement.

CONCLUSION :

The system gain (e-/DN) and total noise (e-) of the primary gain settings of 1, 2, 4, and 8 for the HRC build 1 detector were measured. The full well is estimated at ~135000 e-. The gains from SSDIF (Table 1) and the noise measurements from TB/TV 3 (Table 4) are the primary reference.