CMS 3D CMS Logo

PedsFullNoiseTask.cc
Go to the documentation of this file.
1 
3 
11 
12 #include "boost/lexical_cast.hpp"
13 
14 
15 // -----------------------------------------------------------------------------
16 //
18  CommissioningTask(dqm, conn, "PedsFullNoiseTask"),
19  nstrips_(256)
20 {
22  << "[PedsFullNoiseTask::" << __func__ << "]"
23  << " Constructing object...";
24  edm::ParameterSet params = pset.getParameter<edm::ParameterSet>("PedsFullNoiseParameters");
25  nskip_ = params.getParameter<int>("NrEvToSkipAtStart");
26  skipped_ = false;
27  nevpeds_ = params.getParameter<int>("NrEvForPeds");
28  pedsdone_ = false;
29  nadcnoise_ = params.getParameter<int>("NrPosBinsNoiseHist");
30  fillnoiseprofile_ = params.getParameter<bool>("FillNoiseProfile");
31  useavgcm_ = params.getParameter<bool>("UseAverageCommonMode");
32  usefloatpeds_ = params.getParameter<bool>("UseFloatPedestals");
33 }
34 
35 // -----------------------------------------------------------------------------
36 //
38 {
40  << "[PedsFullNoiseTask::" << __func__ << "]"
41  << " Destructing object...";
42 }
43 
44 // -----------------------------------------------------------------------------
45 //
47 {
48 
49  LogTrace(sistrip::mlDqmSource_) << "[PedsFullNoiseTask::" << __func__ << "]";
50 
51  // pedestal profile histo
52  pedhist_.isProfile_ = true;
53  pedhist_.explicitFill_ = false;
54  if (!pedhist_.explicitFill_) {
58  }
62  fedKey(),
64  connection().lldChannel(),
66  pedhist_.histo( dqm()->bookProfile( titleped, titleped,
67  nstrips_, -0.5, nstrips_*1.-0.5,
68  1025, 0., 1025. ) );
69 
70  // Noise profile
71  noiseprof_.isProfile_ = true;
72  noiseprof_.explicitFill_ = false;
77  }
81  fedKey(),
83  connection().lldChannel(),
85  noiseprof_.histo( dqm()->bookProfile( titlenoise, titlenoise,
86  nstrips_, -0.5, nstrips_*1.-0.5,
87  1025, 0., 1025. ) );
88 
89  // noise 2D compact histo
90  noisehist_.explicitFill_ = false;
92  noisehist_.vNumOfEntries_.resize((nstrips_+2)*2*(nadcnoise_+2), 0);
93  }
97  fedKey(),
99  connection().lldChannel(),
101  noisehist_.histo( dqm()->book2S( titlenoise2d, titlenoise2d,
103  nstrips_, -0.5, nstrips_*1.-0.5 ) );
104  hist2d_ = (TH2S *) noisehist_.histo()->getTH2S();
105 
106 }
107 
108 // -----------------------------------------------------------------------------
109 //
111  const edm::DetSet<SiStripRawDigi> & digis )
112 {
113 
114  // Check number of digis
115  uint16_t nbins = digis.data.size();
116  if (nbins != nstrips_) {
118  << "[PedsFullNoiseTask::" << __func__ << "]"
119  << " " << nstrips_ << " digis expected, but got " << nbins << ". Skipping.";
120  return;
121  }
122 
123  // get the event number of the first event, not necessarily 1 (parallel processing on FUs)
124  static int32_t firstev = summary.event();
125 
126  // skipping events
127  if (!skipped_) {
128  if (static_cast<int32_t>(summary.event()) - firstev < nskip_) {
129  return;
130  } else { // when all events are skipped
131  skipped_ = true;
132  if (nskip_ > 0) LogTrace(sistrip::mlDqmSource_) << "[PedsFullNoiseTask::" << __func__ << "]"
133  << " Done skipping events. Now starting pedestals.";
134  }
135  }
136 
137  // determine pedestals - decoupled from noise determination
138  if (!pedsdone_) {
139  if (static_cast<int32_t>(summary.event()) - firstev < nskip_ + nevpeds_) {
140  // estimate the pedestals
141  for ( uint16_t istrip = 0; istrip < nstrips_; ++istrip ) {
142  updateHistoSet( pedhist_, istrip, digis.data[istrip].adc() );
143  }
144  return;
145  } else { // when pedestals are done
146  pedsdone_ = true;
147  // cache the pedestal values for use in the 2D noise estimation
148  peds_.clear(); pedsfl_.clear();
149  for ( uint16_t iapv = 0; iapv < 2; ++iapv ) {
150  for ( uint16_t ibin = 0; ibin < 128; ++ibin ) {
151  uint16_t istrip = (iapv*128)+ibin;
152  if (usefloatpeds_) {
153  pedsfl_.push_back(1.*pedhist_.vSumOfContents_.at(istrip)/pedhist_.vNumOfEntries_.at(istrip));
154  } else {
155  peds_.push_back(static_cast<int16_t>(1.*pedhist_.vSumOfContents_.at(istrip)/pedhist_.vNumOfEntries_.at(istrip)));
156  }
157  }
158  }
159  LogTrace(sistrip::mlDqmSource_) << "[PedsFullNoiseTask::" << __func__ << "]"
160  << " Rough pedestals done. Now starting noise measurements.";
161  }
162  }
163 
164  // fill (or not) the old-style niose profile
165  if (fillnoiseprofile_) {
166  // Calc common mode for both APVs
167  std::vector<int32_t> cm; cm.resize(2, 0);
168  std::vector<uint16_t> adc;
169  for ( uint16_t iapv = 0; iapv < 2; iapv++ ) {
170  adc.clear(); adc.reserve(128);
171  for ( uint16_t ibin = 0; ibin < 128; ibin++ ) {
172  if ( (iapv*128)+ibin < nbins ) {
173  adc.push_back( digis.data.at((iapv*128)+ibin).adc() );
174  }
175  }
176  sort( adc.begin(), adc.end() );
177  // take median as common mode
178  uint16_t index = adc.size()%2 ? adc.size()/2 : adc.size()/2-1;
179  cm[iapv] = static_cast<int16_t>(adc[index]);
180  }
181  // 1D noise profile - see also further processing in the update() method
182  for ( uint16_t istrip = 0; istrip < nstrips_; ++istrip ) {
183  // calculate the noise in the old way, by subtracting the common mode, but without pedestal subtraction
184  int16_t noiseval = static_cast<int16_t>(digis.data.at(istrip).adc()) - cm[istrip/128];
185  updateHistoSet( noiseprof_, istrip, noiseval );
186  }
187  }
188 
189  // 2D noise histogram
190  std::vector<int16_t> noisevals, noisevalssorted;
191  std::vector<float> noisevalsfl, noisevalssortedfl;
192  for ( uint16_t iapv = 0; iapv < 2; ++iapv ) {
193  float totadc = 0;
194  noisevals.clear(); noisevalsfl.clear();
195  noisevalssorted.clear(); noisevalssortedfl.clear();
196  for ( uint16_t ibin = 0; ibin < 128; ++ibin ) {
197  uint16_t istrip = (iapv*128)+ibin;
198  // calculate the noise after subtracting the pedestal
199  if (usefloatpeds_) { // if float pedestals -> before FED processing
200  noisevalsfl.push_back( static_cast<float>(digis.data.at(istrip).adc()) - pedsfl_.at(istrip) );
201  // now we still have a possible constant shift of the adc values with respect to 0, so we prepare to calculate the median of this shift
202  if (useavgcm_) { // if average CM -> before FED processing
203  totadc += noisevalsfl[ibin];
204  } else { // if median CM -> after FED processing
205  noisevalssortedfl.push_back( noisevalsfl[ibin] );
206  }
207  } else { // if integer pedestals -> after FED processing
208  noisevals.push_back( static_cast<int16_t>(digis.data.at(istrip).adc()) - peds_.at(istrip) );
209  // now we still have a possible constant shift of the adc values with respect to 0, so we prepare to calculate the median of this shift
210  if (useavgcm_) { // if average CM -> before FED processing
211  totadc += noisevals[ibin];
212  } else { // if median CM -> after FED processing
213  noisevalssorted.push_back( noisevals[ibin] );
214  }
215  }
216  }
217  // calculate the common mode shift to apply
218  float cmshift = 0;
219  if (useavgcm_) { // if average CM -> before FED processing
220  if (usefloatpeds_) { // if float pedestals -> before FED processing
221  cmshift = totadc/128;
222  } else { // if integer pedestals -> after FED processing
223  cmshift = static_cast<int16_t>(totadc/128);
224  }
225  } else { // if median CM -> after FED processing
226  if (usefloatpeds_) { // if float pedestals -> before FED processing
227  // get the median common mode
228  sort( noisevalssortedfl.begin(), noisevalssortedfl.end() );
229  uint16_t index = noisevalssortedfl.size()%2 ? noisevalssortedfl.size()/2 : noisevalssortedfl.size()/2-1;
230  cmshift = noisevalssortedfl[index];
231  } else { // if integer pedestals -> after FED processing
232  // get the median common mode
233  sort( noisevalssorted.begin(), noisevalssorted.end() );
234  uint16_t index = noisevalssorted.size()%2 ? noisevalssorted.size()/2 : noisevalssorted.size()/2-1;
235  cmshift = noisevalssorted[index];
236  }
237  }
238  // now loop again to calculate the CM+pedestal subtracted noise values
239  for ( uint16_t ibin = 0; ibin < 128; ++ibin ) {
240  uint16_t istrip = (iapv*128)+ibin;
241  // subtract the remaining common mode after subtraction of the rough pedestals
242  float noiseval = (usefloatpeds_ ? noisevalsfl[ibin] : noisevals[ibin]) - cmshift;
243  // retrieve the linear binnr through the histogram
244  uint32_t binnr = hist2d_->GetBin(static_cast<int>(noiseval+nadcnoise_), istrip+1);
245  // store the noise value in the 2D histo
246  updateHistoSet( noisehist_, binnr ); // no value, so weight 1
247  }
248  }
249 
250 }
251 
252 // -----------------------------------------------------------------------------
253 //
255 {
256 
257  // pedestals
259 
260  if (fillnoiseprofile_) {
261  // noise profile (does not use HistoSet directly, as want to plot noise as "contents", not "error")
263  for ( uint16_t ii = 0; ii < noiseprof_.vNumOfEntries_.size(); ++ii ) {
264  float mean = 0.;
265  float spread = 0.;
266  float entries = noiseprof_.vNumOfEntries_[ii];
267  if ( entries > 0. ) {
268  mean = noiseprof_.vSumOfContents_[ii] / entries;
269  spread = sqrt( noiseprof_.vSumOfSquares_[ii] / entries - mean * mean ); // nice way to calculate std dev: Sum (x-<x>)^2 / N
270  }
271  float error = spread / sqrt(entries); // uncertainty on std.dev. when no uncertainty on mean
272  UpdateTProfile::setBinContent( histo, ii+1, entries, spread, error );
273  }
274  }
275 
276  // noise 2D histo
278 
279 }
280 // -----------------------------------------------------------------------------
int adc(sample_type sample)
get the ADC sample (12 bits)
T getParameter(std::string const &) const
virtual ~PedsFullNoiseTask()
std::vector< float > vNumOfEntries_
virtual void book()
Utility class that holds histogram title.
const std::string & title() const
static const char mlDqmSource_[]
std::vector< float > vSumOfContents_
void updateHistoSet(HistoSet &, const uint32_t &bin, const float &value)
static const char noise2D_[]
virtual void update()
Class containning control, module, detector and connection information, at the level of a FED channel...
const uint32_t & event() const
T sqrt(T t)
Definition: SSEVec.h:18
static void setBinContent(TProfile *const profile, const uint32_t &bin, const double &entries, const double &mean, const double &spread)
static const char noiseProfile_[]
#define LogTrace(id)
ii
Definition: cuy.py:588
int extract(std::vector< int > *output, const std::string &dati)
static const char pedestals_[]
DQMStore *const dqm() const
CompactHistoSet noisehist_
std::vector< int16_t > peds_
collection_type data
Definition: DetSet.h:78
PedsFullNoiseTask(DQMStore *dqm, const FedChannelConnection &conn, const edm::ParameterSet &pset)
void histo(MonitorElement *)
const uint32_t & fedKey() const
std::vector< double > vSumOfSquares_
const FedChannelConnection & connection() const
std::vector< float > pedsfl_
virtual void fill(const SiStripEventSummary &, const edm::DetSet< SiStripRawDigi > &)