I have finished measurements of 3 RS
PMTs. First of all I understand that I was
wrong with splitter coeff. assumptions. The splitter which I've used
in all previous
tests is not a RS, or any other standard type. This is just a simple
resistor network with
coeff. of exactly 2. In this tests data were taken without this splitter,
with direct connection
of PMT to ADC. Second thing which is also I understand now is that
the PMT (TEST#2)
I've used for all previous tests has a extremely high gain unusual
for average RS PMTs.
It is in 5 to 10 times more than an average. So Tadashi's measurements
in fact are consistent
with my.
Some explanations about this tests. The used
technique was the following. First I've made
a single photoelectron measurement with biggest HV with some very small
amount of light
from LED and efficiency not more 5-10 percent, so that this is a single
photoelectron peak
in the amplitude distribution. Then I raise amount of light with the
same PMT voltage. The
number of p.e. I have with this light is ratio between amplitudes.
Then I scale down the HV
with the same amount of light (p.e.) obtaining the coeffs Counts/p.e.
and p.e./counts as ratio
between amplitude and number of p.e. calculated previously. The plots
are: