Executive summary:

• The EPD FEEs have a DC offset that is tunable via a register whose value we call Vped.  We want to adjust each channel's Vped such that the ADC pedestal is the same with the FEEs on (but bias set to zero) as it is with the diff-Rx cards turned off.
• This year, there was some confusion about the PedAsPhys runs, so we simply used only pedestal runs taken at different Vped values, and a linear interpolation to find the "right" Vped for all FEE channels.

---

Procedure:

Rosi has done this in the past, using PedAsPhys runs, similarly to how we do the Timing and Bias scans.  Her instructions are here.  This year, we did not do PedAsPhys runs, for reasons not worth recording.  And in any event, simply comparing pedestal runs makes more sense.

Here are instructions, modified from Rosi's ones:

1) Log into evpops so you can change the EPD configuration:ssh -A -Y you@rssh.rhic.bnl.gov

• ssh -t -A -Y stargw.starp.bnl.gov
• ssh evpops@daqman
• cd /RTS/conf/epd
• cp EPD_TUFFmap.txt EPD_TUFFmap_todaydate.txt   <-- this is just to keep it for safety

2) To set all the Vpeds to -20:

• cp VpedPatterns2018/Epd_vped-20.txt EPD_TUFFmap.txt
• Then click Physics on the GUI at: https://dashboard1.star.bnl.gov/daq/EPD_UI/?EPD

3) Take a pedestal rhicclock clean for 1000 events only keeping trg and daq in as the systems.

• Keep track of which run number has which Vped.  You'll put these into the analysis macro (attached)

4) Repeat steps 2 and 3 for Vped=-20,-10,0,10,20,40,80

5) Turn the CAMAC crate holding the diff-Rx cards OFF, and take another pedestal run as in (3).  Keep track of this run number, too

6) get all the eq*.mean.* files from the daq computer, and take my macro below to analyze.

• You will see in the macro where to put the run numbers and Vped values

7) Return the EPD to its proper status:

• cp EPD_TUFFmap_todaydate.txt EPD_TUFFmap.txt
• click physics
• Take a pedestal rhicclock clean for 1000 events

8) Run the macro provided here (or at the bottom of the page).

• You'll have the edit the beginning of the macro, to use your run numbers and Vped values.  You might have to play a bit, but you'll figure it out.

---

Results:

With the FEEs and diff-Rx crate ON, we (well, Maria and Rosi) took 8 pedestal runs, at different values of Vped in the FEEs, as described above.  The resulting pedestal depends more or less linearly on Vped, except has a minimum value of 1.  We do a linear fit to the points with pedestal>1.  Then, we find the Vped value that would give the same pedestal as we find with a run we took with the CAMAC crate that holds the diff-Rx cards OFF.

As in all of these scans with 2x12x32 = 768 channels, a few of them have to be tweaked by hand.

The macro, new this year because we are using only pedestal runs instead of PedAsPhys runs as we've done in the past, is attached at the bottom of the page.  Also there are all data files needed to do the analysis.


Above: The right-most (eighth) QT board in EQ4.  (In the trigger numbering system, this is board 16.)  Black points show the pedestal as a function of Vped used in the run.  Red is a linear fit to points with pedestal>1.  Blue square (plotted arbitrarily at Vped=0 along the horizontal axis) is the pedestal measured with the diff-Rx cards OFF.  The pink line shows the linear extrapolation through which we find the Vped value which will give the same pedestal as the OFF runs.  This value is shown in the grey squares.  You can get the pdf with all of the boards here.



Above: Confirming the results, thanks to Maria.  Comparison of pedestals taken with the newly-adjusted Vped values, and pedestals taken with the diff-Rx crates OFF.  This is just for one board; you can get her full comparison here.