A crew working on  4 Feb 2018:

Using Wanbing's monitoring/control software at: dashboard1.star.bnl.gov/daq/EPD_UI/

Until all databases are understood, Tonko gave us a quick & dirty way to turn all bias voltages to a fixed value.  So, we changed Vbias to 62.5 V.

Wanbing's display of Idark is then shown below.

Figure 1: EPD dark current distribution with every bias voltage set to 62.5.

Note: One FEE is not responding at all; more on this below.

Below is the result of opening up the NE FEE box and inserting two cellphones with their bright "flashlights" shining directly on the SiPMs:

Figure 2: EPD dark current distribution after including a large light in the NE box.

So we can clearly see an increase in dark current in many of the FEEs.  The red line in the bottom shows the current history for allegedly PP3TT4.  We can notice several things, one is that the FEEs are not all clustered in half-wheel as we expected (NE FEE box corresponds to PP07-12 East).  This indicates that the mapping is not correct in the database.   Another is that the monitoring system other than that seems to be working; we can change the bias voltages of the EPD as a whole and we see a reasonable response in the dark current.

We then turned off the FEEs in the NW box by turning off on LV PS (one that does not also power a TUFF box), which gave us this distribution:

Figure 3: NW Fee Box turned off

Turning off one FEE box turns off 1/4 of the FEEs.  Good.  However, the NW box should correspond to PP01-06 West.  Quite unclear where the above (incorrect) mapping is made.  (We can't fix it until we know where it is.)

Notice that the one sector (half-supersector) that was "dead" (Figure 1) corresponds to a FEE in the NW box.  This makes sense, because Tonko had reported that one FEE was not communicating; it was #6 in the NW box.  We later discovered that its address had been mis-set, and fixed it.  Likely this was never updated in the database, so appears as "dead."

Conclusions so far:

• The monitoring GUI seems to be working well!  Bias voltage, dark current, and temperature are reasonable, and respond to changes
• The geometrical mapping:  PP/TT <---> TUFF/Branch/Channel (or is it 1wireID??) is wrong.  This is not surprising, and we'll fix it, but we need to know how and where.
• Request by Mike: In many of the database tables, there is like Vset0, Vset1, fps_quad, fps_layer, which have to do with FPS.  When EPD tables were made, the headings remained the same but some (e.g. fps_quad) were simply unused and others (fps_layer) are re-purposed.  It's confusing.  Now that we are no longer a temporary detector, can we please make EPD-specific headings and remove dead wood?  Prashanth can specify details.