Difference between revisions of "BCal Beam Test Plots, January 3, 2007"
From GlueXWiki
| Line 1: | Line 1: | ||
| − | From suggestions from Alex and others I wrote a piece of code that plots z finds sigma and then adjusts the PMT gain correction factors until the sigma of the gaussian fit is at it's minimum. Z is defined as | + | From suggestions from Alex and others I wrote a piece of code that runs in ROOT that plots z finds sigma and then adjusts the PMT gain correction factors until the sigma of the gaussian fit is at it's minimum. Z is defined as |
: <math> z = \frac{E_{cal} - E_{beam}}{E_{beam}} </math> | : <math> z = \frac{E_{cal} - E_{beam}}{E_{beam}} </math> | ||
| − | You can find the [[peakfit.C]] code here. | + | You can find the [[peakfit.C]] code here. ( in ROOT > .x peakfit.C ) |
The code goes through and adjusts each parameter by an amount epsilon until sigma is minimal. Of course the ADC conversion factor (E_beam/E_cal) has to be calculated each time too. Doing this shows a good improvement in the "floor term" and the energy dependent is still approximately 5%. | The code goes through and adjusts each parameter by an amount epsilon until sigma is minimal. Of course the ADC conversion factor (E_beam/E_cal) has to be calculated each time too. Doing this shows a good improvement in the "floor term" and the energy dependent is still approximately 5%. | ||
| + | |||
| + | The fit is using <math> \sigma_{z} = \frac{a}{\sqrt{E(GeV)}} + b </math> in the package xmgrace. | ||
[[Image:Bcalres.gif]] | [[Image:Bcalres.gif]] | ||
Revision as of 16:29, 3 January 2007
From suggestions from Alex and others I wrote a piece of code that runs in ROOT that plots z finds sigma and then adjusts the PMT gain correction factors until the sigma of the gaussian fit is at it's minimum. Z is defined as
You can find the peakfit.C code here. ( in ROOT > .x peakfit.C )
The code goes through and adjusts each parameter by an amount epsilon until sigma is minimal. Of course the ADC conversion factor (E_beam/E_cal) has to be calculated each time too. Doing this shows a good improvement in the "floor term" and the energy dependent is still approximately 5%.
The fit is using 
in the package xmgrace.
