Here I compare the tower pedestals measured:

1. Without beam, through slow control;

2. With beam, with L2ped algorithm in run 10135070.

In the following plots X axis is the DaqID = crate*160+channel.

In the first plot I show the difference between pedestals measured with and without beam.

The second plot shows the ratio of pedestal widths with and without beam.

Conclusion:

Pedestals with the beam do shift by up to 3 ADC counts and broaden by a factor of up to 3.

To reproduce these calculations, download the attached raw data.

The effect of these new pedestals is shown on the following plot. It seems like a minor improvement only.