Figure 1

A schematic of a sensor is shown in fig1 (the ccd contains an array of these sensors (659x494 in the case of the vesta's)). When a photon falls on the sensor an electron is freed in the semiconductor. This is able to travel for a while before becoming captured again.

Figure 2

Figure 2 shows a voltage applied across the surface of the ccd. Now when the electron is liberated by the photon it is drawn into the 'well' underneath. CCD's are characterised by the depth of these wells (how may electrons the well can hold). The deeper the wells the greater the dynamic range and the more expensive the ccd. While the electron is in the well and as long as the voltage across the chip is maintained the electron is held. So we now have a system there the collected electrons is proportional to the light hitting the sensor. Next we need a way of reading this information out.

Figure 3

Also on the surface on the chip are a series of vertical registers (one for each column of sensors) and a horizontal register fig 3. Now we come to the charge transfer pluses.

Figure 4

When these are applied the charge from the sensors is moved in to the corresponding vertical register. Fig 4.

Figure 5

These registers are now clocked down the chip using the vertical clock inputs to the chip. Each clock cycle moves the pixels down by one, the bottom pixel ending up in the horizontal register. (fig5).

Figure 6

Finally the pixels are clocked along this register though an on chip amplifier and out of the ccd by the horizontal clocks. This signal then enters the camera's a/d converted for conversion into a digital signal.

With this much-simplified model we can get an idea of what the mod is doing.

Firstly the mod disconnects the shutter pulse (pin 10 on the vertical driver). To understand the shutter consider the ccd running at 30 frames per second. The total time the sensors are exposed would be 1/30th of a second. However if there is no voltage applied across the chip (fig1) then the freed electrons can't enter the well. If the voltage is only applied or 1/2 the duration of each frame (1/60th of a second) then the effective exposure will be 1/60th of a second. What the shutter pulse is doing is controlling the voltage across the chip. The long exposure mod requires that the voltage is maintained during the duration of the long exposure and so disables the shutter control.

Next the chip read out. Using the standard divers the slowest the rate the camera can be read out is 1/5th of a second. However all the mod has to do to increase the duration of the exposure is stop the accumulated charge in the well from being moved into the vertical registers. ie to block the charge transfer pluses. The most convenient way of achieving this is at the vertical driver chip. This chips takes the standard logic level outputs of the cameras timing generator and converts the signals the voltage levels required by the ccd. By blocking this signal from pins 8 and 13 of this chip the ccd no longer gets charge transfer pluses and the electrons are allow to accumulate in the ccd's wells. Now we have a long exposure but we still need a way of re-enabling the charge transfer pluses to allow the image to be read out. This is controlled by a signal from the pc's printer port via the logic chip. When the printer port signal is high then charge transfers are enabled and vice versa. So all the mod is doing is disabling the shutter and controlling the charge transfer pluses. This means the vertical and horizontal clocks are left running. This turns out to be a good thing. If these clocks were stopped as well 'thermal noise' would build up in these registers and lead to a degradation of the image. Also the on chip amp is left powered. This is not as good! On long exposure you will be able to see some electro luminescents from this the top left hand part of the image. This has to be removed together with 'hot' pixels by dark frame subtraction.