The PECA files contain the set of titles banks PDST where the bank number specifies to which 1/2 crate the bank applies. For each 1/2 crate there is a line for each cell/channel/card which contains 5 numbers: A bit-encoded status word specifying is any of the following values should be considered ``bad'' (so a zero here mean everything's great), the Q_HS pedestal, the Q_HL pedestal, the Q_LX pedestal and the time pedestal.
The QECA files contain the set of titles banks QSLP where the bank number specifies to which 1/2 crate the bank applies. For each 1/2 crate there is a line for each cell/channel/card which contains 6 numbers: the linearly fit slope*100 for Q_HS, the intercept*10 for Q_HS, the same for Q_HL and the same for Q_LX. Fits which are considered ``bad'' are flagged by the slope being set negative (or to zero in some cases).
The TECA files contain the set of titles banks TSLP where the bank number specifies to which 1/2 crate the bank applies. Rather than containing fit variables for the TSlope shape, this file now contains the actual median digitized values obtained for each injected delay so that one can do a direct interpolation between these points rather than trying to roughly parameterize a tricky form. The first number in each titles bank gives the number of time samplings used. Next is given the number of packed words per channel (the digitized values are packed to save space!). Then there is a list of the injected time values (expressed in hundredths of a ns). Now, for each cell/channel/card there is given a bit-encoded status word specifying is any of the following values should be considered "bad", and then a series of packed digitized values corresponding to the injected times.