In default operation, the image intensifier is assumed to have uniform efficiency across its surface. This can be wrong if the MCP has been damaged by being exposed to too much light at high gain. This causes a deadening of the MCP in specific areas and can affect the extracted beam position and beam sigmas. One can either correct this by calibrating the efficiency of the image intensifier across its surface, or move to a known good area of the image intensifier surface, or do both.

We look for dead or low efficiency regions on the MCP by using a UV lamp positioned near one side of the image intensifier. The image intensifier will be fairly uniformly lit and any deviations from a fairly uniform intensity image will indicate poor efficiency regions.

The calibration should only be done when no beam is present in the Tevatron. The calibration sequence can be started by clicking the CalNow button ( ) in the main panel. After the next cycle through protons and pbars the calibration panel will open and the calibration will start, first for the proton MCP, then the pbar.

Note that if the light beam is directed on an inefficiency part of the image intensifier, e.g. the red-orange areas, the best thing is to move the light image via the flip-in mirror to a good region of the MCP. Then the calibration data does not actually have to be used to correct the collected image data. Note that the calibration is approximate since it assumes the UV lamp illuminates the MCP uniformly. This is thought to be good enough as one just requires the illumination if not completely uniform to be smoothly and only slowly varying across the region of interest. It can be seen in the two calibration images above that the UV illumination is from the right in the proton MCP and from the lower right for the pbar MCP.

In default operation the beam image is directed to a good region of the MCP and the collected data is not corrected by the calibration image. If calibration correction is desired, one needs to select the MCP Norm button ( ). The calibration data from the last calibration is used. There is currently no feature to select a particular calibration image. (Also if the image frame grabber window is changed between calibration and image collection, I don't yet know if the calibration correction is done properly. Also I do not know of a feature to view the previous calibration that will be used, so if a calibration was done previously and failed (UV lamp has problems) one cannot be sure the calibration correction is done correctly.) It is better at this point to just move the beam image to a good region of the MCP and not turn on the MCP calibrationcorrection.

In order to get the 2D profiles for individual bunches the image intensifier has to be gated to only capture light for the particular bunch of interest. The actual gating mechanism hardware is described in the Image Intensifier page. The gate timing for each bunch is set in the settings panel using a coase timing in units of BS (beam sync) ticks and a vernier scale in units of ns (specified as seconds.)

The gate that is sent to the MCP is a square pulse of width 200 ns and -200V. However due to the attenuation and dispersion caused by the long RG58 cable run, the pulse reaching the MCP is not square but fairly triangular in shape with a width of about 150 ns. The timing was set with this wide gate and no attempt has yet been made to try to use a narrower gate to make sure only light from a single bucket is captured.

The timing is based on a BSCLK (beam sync clock) signal input into a BD Camac 279 module. This module outputs a pulse at the BS tick rate and timed relative to one of the A,B, or C markers (the A marker is used I think.) To achieve timing diferences that are smaller than one BS tick (7 buckets) a delay module is used with a timing that can be set in ns.

We know from BD that the bunches within a train are separated by 3 BS ticks and that each train is separated from the next by an integral number (20) of BS ticks. This means that once the vernier scale in ns is set so that we can find the first bunch of a train, the rest of the bunch timings are easy to obtain. To find one bunch we run in ungated mode with a lower MCP high voltage to make sure that the light is imaged on a good part of the image intensifier. Then in gated mode we use the TV monitor to look for a signal as we change the timing of the gate. Since the gate is wide enough we can usually see some light from a bunch just by changing the number in the bs timing box (see the setting panel web page) by one until an image is seen. Then the ver timing box can be changed until the gate is centered on a bunch, e.g. by changing +/-50 ns and +/- 100 ns. Once centered find the first bunch by looking +/-3 BS ticks from the current bunch. The first bunch of the next train is at a BS tick number of 20 larger than the last bunch of the previous train. Note that the bs timing box number goes from 1 to 159 BS ticks, so if bunch 1 is at 157 BS ticks, then bunch 2 is at 1 BS ticks.

The current timing for Run 2a is given in the table below, where it should be noted that the proton is still using the old copper based BSCLK signal in the proton 279 module while the pbar uses the new fiber based BSCLK signal. Both sets of cables are at the CAMAC crate but the pbar copper based BSCLK signal is no longer working. Greg Vogel tells me it is lost somewhere between Echo0 and C0 but wants to move everything over to the fibered based BSCLK anyway. When the proton 279 is moved to using the fiber based BSCLK signal the timing for the protons will have to be redone.

Note that the proton and pbar timings are different, and the pbar uses its own pbar A marker.