Saturation study of ROMEO


Index

  1. Short summary
  2. Saturation Function
  3. Scripts, data, and Keywords
  4. Charge saturation vs waveform saturation with 10ns single pulse beam
  5. Charge saturation vs waveform saturation with 50ns single pulse beam
  6. Single pulse beam and continuous beam
    1. A) 10ns waveform (1 shot of 10ns beam, identical to the above example)
    2. B) 50ns waveform with 10ns bin ( continuous 5 shot of 10ns beam)
  7. Weighted MCHits, semi-current saturation vs waveform saturation with 50ns beam
    1. A) 5ns bin, 50ns beam
    2. B) 10ns bin, 50ns beam
    3. C) 25ns bin, 50ns beam
    4. D) 50ns bin, 50ns beam


Short summary

This is studies about saturations in ROMEO, with two different saturation mode "waveform saturation" and "semi-current saturation".
The semi-current saturation used to be default mode of ROMEO, while waveform saturation is used in pmt-simulator.
With 10ns square beam, my result shows considerable underestimate of output NPE in semi-current saturation, over 500 input NPE / 10ns.
This discrepancy will be smaller if we assume wider waveform over 50ns (which is more realistic).

On the other hand, the weighted input hits with wide bin-width over 10ns gave opposite results.
Since a weighted hit represents numbers of unweighted-hits within given time bin-width, the hit has much higher amplitude than the unweighted case.
Therefore the waveform saturation works much stronger.

With the default setting of hit-constructor, we have 3ns bin weighted hits in the first ~400ns and then 25ns bin weighted hits for FADC.
The latter wide weighted hits may result in an underestimate of total NPEs if we apply waveform saturation.
(EHE group is not using waveform saturation currently)

contact person:  hoshina(at)icecube.wisc.edu


Saturation Function

The saturation function is obtained by Chris Wendt.

x    : input current [mA]
y    : output current [mA]
P   : 1.0709e-3  (constant)
X0 : 48.965 (constant)
X1 : 129.74 (constant)

1/y = 1/x + P * (ln(1+(x/X0)^3) / ln(1+(x/X1)^0.5))

This is the saturation curve of y / x as a function of input current [A] (coded in RMOPMTCurrentBasedSaturation.h)




Scripts, data, and Keywords

Keywords :


Charge saturation vs waveform saturation with 10ns single pulse beam

[script For the case of 4 and 5, input PEs are merged into one MCHit with weight of  x-axis NPE]

Beam property:
The input photons distributes from t = 100 ns to 110 ns (10ns beam).
Number of photons per beam is shown in x-axis.
For the case of 4 and 5, input PEs are merged into one MCHit with weight of  x-axis NPE.
  1. No saturation. [data]
  2. Waveform saturation only, no binning. [data]
  3. semi-current saturation with charge-bin width 2ns. [data]
  4. Romeo internal charge binning + waveform saturation. [data]
    Waveform saturation is applied after 2ns charge binning.
  5. Weighted MCHit(EHE MCHit) + semi-current saturation with 10ns charge bin width. [data]
    The input is only one MCHit, but is weighted by a factor of 100 ~1000.
    No charge binning (since the input is only 1 MCHit) but bin width 10ns is used to calculate average_current.
  6. Weighted MCHit + waveform saturation. [data]
    No charge binning. Merged bin-width 10ns is not used to calcurate saturation.


[top]

Charge saturation vs waveform saturation with 50ns single pulse beam

Same figure but beam width is extended 50ns.
  1. No saturation. [data]
  2. Waveform saturation only, no binning. [data]
  3. semi-current saturation with charge-bin width 2ns. [data]
  4. Romeo internal charge binning + waveform saturation. [data]
  5. 5 Weighted MCHits(EHE MCHits) + semi-current saturation with 10ns charge bin width. [data]
  6. 5 Weighted MCHits + waveform saturation. [data]


[top]


Single pulse beam and continuous beam

Alex's study suggests that the shape of waveform (gaussian or rectangular) affects saturation.
In the practical cases, however, waveforms are wider than 10ns especially at high-PE density.
The semi-current saturation with weighted hits at ~50ns waveform (case B)  shows much better agreement to non-weighted hits with waveform saturation, compared with the single 10ns waveform pulse (case A).

A) 10ns waveform (1 shot of 10ns beam, identical to the above example)



Next is the ratio plot of purple / red and blue / red.
With the blue line I tried to reproduce Alex's fig , however, my result shows smaller deviation from 1.







[top]

B) 50ns waveform with 10ns bin ( continuous 5 shot of 10ns beam)



Next is the ratio plot of purple / red and blue / red.
With the blue line I tried to reproduce Alex's fig (you need to multiply x-axis of next fig by a factor of 5), however, my result shows much smaller deviation from 1.  Presumably we are using different setup??









[top]


Weighted MCHits, semi-current saturation vs waveform saturation with 50ns beam

[script]

A) 5ns bin, 50ns beam




[top]

B) 10ns bin, 50ns beam

see also B) 50ns waveform with 10ns bin ( continuous 5 shot of 10ns beam)

[top]

C) 25ns bin, 50ns beam





[top]

D) 50ns bin, 50ns beam



[top]