Test of 2019 batch (23 PMTs)
Specification from Hamamatsu (old GlueX u/p):
PDF or
HTML.
Preliminary test
All 23 PMTs were put in a light-tight black box. Their signals were connected to
a scope. They were checked at 1200V, 1500V and 1750V for
- Base divider current (about 340uA at 1750V);
- Absence of DC offset on the last dynode;
- Shape and magnitude of their pulses on the scope;
- Counting rate (i.e., noise).
Everything looked normal.
Gain measurement: single photoelectron method
We have used NIM 10x amplifier in the past for this test.
Unfortunately, they were all broken this time around and no
replacement could be found. Nevertheless, an unamplified SPHE peak
could be clearly separated from a pedestal and fitted. Two plots
below show the QDC spectrum for the lowest and the highest
measured PMT gains:
Measured PMT gains and their ratio to specified gains can be found
here.
New PMTs are at the end of the table (S/N starts with FA0xxx).
Good news is that measured gains more or less follow the specified gains.
However, they are about 20% lower on average. In the past, they were much
closer. I can think of 2 things which have changed recently and may explain
that: (a) A new CAEN HV unit has replaced an old LeCroy HV unit, and it may
be supplying lower HV than nominal 1750 V; (b) A new Caen v792 QDC
has replaced an old Caen v256 QDC. Its actual "charge/per QDC count" may be less
than 100e-15 C as its specification claims.
Now bad news: there is 1 problematic PMT FA0378. Its measured gain is only ~45% of
the specified one, and its spectrum looks strange with very broad pedestal and
long tail (see the plot). It may require further investigation.