Bad pads

Before doing hitfinding or tracking or anything, we have to know which of our roughly 14000 pads we can trust. Once we determine that a pad is trustworthy, we must determine the calibration constants for that pad (gain, t0 offset, etc.). These calibration constants come mostly from pulser runs.

Here, we just look at pads that show quite weird behavior, like ADC values that are always negative (!) or overflow. Looking at these data, we see a couple things:

The ADC values often take on values that cannot be given by our 12-bit ADC. Therefore, one suspects the digital signal processor (DSP) on the reciever card.
In the pulser runs, it looks like the first few pads on a stick are the most likely to have problems. This may again suggest a DSP problem, if the DSP is not "ready" in time for the first data to hit it.

Given this, it may well be that pads that look hopeless in the pulser data are ok in the physics data, since these two types of data have different formats and are handled differently by the DSP.

This is important since it looks like 5% of the pads are bad (by very simple criteria -- see the plots below) in the pulser data. It may be that less are bad in the physics data. That would mean that we cannot extract exact calibration data for these pads from the pulser data, but we may not want to throw them away.

See here for some bad pads from a pulser data run from File 5 of Tape 35.

Back to The E895 hitfinder home page.

Back to E895 Main Page