class HMdcOffset: public HReconstructor

 Destructor.

 pointer to iterators and categories

 Sets the the default values for the fits.

 Sets ascii output of HMdcOffset for debugging output.

 Sets rootfile output of HMdcOffset where all created histograms were written.

 Retrieves TF1's for min tof substraction from root file
 and switches the usesage of tof substraction kTRUE.

 Retrieves TF1's for min tof substraction from root file
 and switches the usesage of tof substraction kTRUE.

  Inits HMdcOffset and the needed HMdcCalParRaw, if this container does not exists

 Histograms for inverted Time1 and integrated Time1 per Tdc-Channel are created
 in a subdirectory structure.

 Histograms for inverted Time1 and integrated Time1 per Tdc-Channel are created
 in a subdirectory structure.

 Created histograms are deleted

 Created histograms are deleted

 Histograms for inverted Time1 and integrated Time1 per Tdc-Channel are filled

 Histograms for Time1 vers Tdc-Channel and cell are filled

 The offset is calculated from two linear fits. Some pictures of the fitted histograms
 can be found on 

 .
 The first linear fit (red) is done to find the rising edge of the integrated spectra.
 The second linear fit (blue) is done to substract the noise.
 The ranges for both fits can be set through the functions setNoise()(x-range) for the
 second fit (default values: 100,50) and setThreshold()(y-range) for the first fit
 (default values :0.15,0.50). The y-range is calculated by the percentage of the maximum
 height of the spectra.
 The offset ist calculated from the intersection point of both linear fits.
 The sigma of the offset is calculated from the sigmas of the two linear fits.

 If the analyzed file is a pulser file for external calibration a gaussian fit is done.
 To set this fit the function setPulserFile() has to be used in the macro. The range of
 the fit is set through the function setRangeGauss()(default value: 50) around the maximum.
 The mean and sigma of the fit is written to the ascii debugging output.
 The hist is checked for multiple peaks.The peaks which have been found are written to an array.

 The adresses of the channel(sector,module,mbo,tdc),offset, two fitparameters for the
 two linear fits, the sigma of the offset ,offset and the integral(number of entries for this channel)
 are written to the ascii debugging output.In the case of an pulser file (external calibration)
 the mean and the sigma of the gaussian fit are also written to the output.

 All created histograms are written to a rootfile.
 The file is structured for sector,module,mbo.

 Opens the ascii debugging output and writes the headerline.

 fill NTuples for offsets and multiple peaks

 fills CalParRaw with offsets, offseterrors and Methods
 Checks if offset is in valid range arround mean value of offsets.
 If it is out of range the value is replaced by the mean value.
 Not connected channels are marked with -7.

 This function is called after the execute function is finished. At this point
 the arrays for the drift-time are filled. Froms this arrays the histograms for
 the drift-time and the integrated drift-time are filled. The fits for the calcutation
 of the offsets are done and the offsets and the sigma of the offsets are
 calculated .All histograms are written to a rootfile output and the information
 of the fits to an ascii debugging output. The offset and the sigma of the offset
 are filled into the container (HMdcCalParRaw) for the calibration parameters.
 To characterize the method of calculation of the offsets the function setOffsetMethod()
 of HMdcCalParRaw is used. As default value for the method 2 is used for  automatic
 calibration. If no calibration is done this value is set to 0.

 Function to create subdirectories

 Need some work for multiple hists in start detector
 Better select multiplicity 1 in start.

 Fired cells are take from the list of Cells inside one HMdcClus.
 Raw data of Time1  multiplied by the slope of the channel taken from the
 container of the calibration parameters are filled into the array for the
 drift-time . This array contains 2048 (corresponding to the resulution of
 the tdc chip) entries for each tdc channel.

 Raw data of Time1  multiplied by the slope of the channel taken from the
 container of the calibration parameters are filled into the array for the
 drift-time . This array contains 2048 (corresponding to the resulution of
 the tdc chip) entries for each tdc channel.