Specifications

Version 1.1 rev Jan 2013
Page 50
iXon Ultra
, Features and Functionality
3.3.1 - Sensor Pre-amp options
An EMCCD sensor can have a much larger dynamic range than can be faithfully reproduced with the Analogue/
Digital converters and signal processing circuitry currently available on the market. To overcome this shortcoming,
and to access the range of signals from the smallest to the largest, as well as to optimize the camera performance,
it is necessary to allow different pre-amplier gain settings. However, with regards to selecting something other than
the default highest pre-amp (most sensitive) setting for applications, it is recommended that this only ever be carried
out when faced with extremely challenging dynamic range concerns. It is very important, however, that for such high-
dynamic range applications, the user applies even more care to the amount of EM Gain applied (high EM Gain can
drastically reduce the true dynamic range of the camera). Ideally, for maximum dynamic range whilst maintaining
improved Signal to Noise (S/N), the EM Gain setting should be set equal to the read noise at the readout speed selected
(value obtainable from the performance sheet that comes with the delivered system).
Pre-amplier gain selection in CCDs is traditionally used to trade off S/N vs dynamic range. A higher pre-amp setting
means fewer electrons/count, resulting in a lower system noise oor, therefore a higher S/N. However, high pre-amp
settings may not match well to the pixel well depth of the sensor, therefore a lower setting can be selected to meet
the full well depth potential, e.g. a pre-amp setting yielding 1.5 e-/count may be selected to ensure that the 65536
digitization levels of a 16-bit A/D are closely matched to a 100,000 e- pixel well depth. A pre-amp setting of 1 e-/count,
while giving a lower noise oor, would not harness the full 100,000 e- well depth within the 16 bit A/D.
The situation is not nearly as straightforward for EMCCDs because:
1. EM Gain overcomes readout noise and amplies signals relative to the digitization noise (which is xed for a
given pre-amp setting).
2. Gain register pixels have a greater well depth than the imaging pixel well depth.
The latter point can be particularly confusing and has led to confusion in the eld. What this has meant, is that we have
set some of the lower pre-amp settings associated with the EM-output to match the extended well capacity of the gain
register pixels (as reported by the sensor manufacturer e2v). This means that these pre-amp settings are designed to be
used with EM Gain. Otherwise, the lower well capacity of the imaging pixels will saturate long before the A/D. This is why
some users have been confused at not being able to reach the full ~65k counts of the 16-bit A/D channel, when they
hadn’t applied EM Gain.