G Ain; G Ain & D Ynamic R Ange; G Ain & N Oise - Oxford Instruments ANDOR Newton CCD Hardware Manual

Newton
output of the amplifier feeds the Analog-to-Digital (A/D) converter which in turn converts each charge packet into a 16-
bit binary number. The A/D converter on the Newton is located inside the camera head.
6.1.3 g
ain
6.1.3.1 g & d
ain
As the gain is increased it will at some point begin to cause a decrease in Dynamic Range. This occurs when the gain
equals the readout noise, in electrons. Therefore at 1 MHz readout rates the Dynamic Range will be maintained at over
20,000:1 for gains up to about 20 times, whilst at 62.5 kHz it will be maintained at over 43,000:1 for gains up to 6 times.
If higher sensitivities (and hence higher gains) are required then, there will have to be a trade off for Dynamic Range. To
maintain as much Dynamic Range as possible it is advisable not to use a higher gain than is necessary to measure a
signal.
6.1.3.2 g & n
ain
The output from the gain register is fed into a conventional EMCCD output amplifier. This amplifier, even in a scientific
EMCCD, will have a readout noise of a few electrons rms and around 10 or 20 electrons rms at MHz readout rates.
However this noise will effectively be reduced by the multiplication factor of the gain register which, when high enough,
will achieve noise levels below 1 electron rms. So by using the gain you can effectively reduce the noise to insignificant
levels at any readout speed. For example, the Newton EMCCD has a readout noise of a few to tens of electrons,
depending on read out speed.
Using gain will itself add some noise to a measured signal due to the statistical nature of the multiplication process. A
similar effect exists in EMCCDs and is referred to as the Noise Factor. The amount added is dependent on the signal
level and the gain. If there is no gain, then there is no extra noise. At high gain (tens of times higher) it is calculated as the
square root of N (where N is the signal in electrons). This will add to the shot noise of the signal to become the square
root of 2N. So if the signal is large enough to be above the readout noise then there is probably no need for gain and it
should be reduced or turned off.
Conversely, if the signal is being lost in the readout noise then increasing the gain is the only way to detect it. If the gain
is set high enough, then detection of single electron events will be possible. These events will appear on an image as a
spike several hundred counts high. In Andor EMCCD cameras the gain is limited to a maximum of 255 times at –50ºC for
standard systems. This is comparable to high end EMCCDs.
Version 1.4 rev 13 Oct 2017
R
ynaMic ange
oise
30