Dear Mark/Filipe
We run the card with the 20MHz internal clock - our TTL pulses are clean
(when we look at them on a scope, and are all terminated correctly) and
have a duration of ca. 90ns. Our program, based upon the exmaple records
the number of clock 'ticks' between the rising edges of the TTL pulses.
When we put a 10Hz pulse train in the interpulse duration comes out at
100ms - spot on. When we start to try and time our random pulses, along
with the 10Hz pulses we find that summing all of the time intervals
between the 10Hz pulses gives random values - ie they fall between ca. 95
and 99ms, and there is a general trend that the larger the number of
counts recorded in the 100ms period the larger the discrepancy, although I
haven't found an exact correl
ation. ( I refer to the mean number of counts
per interval./laser shot as the 'count rate'.).
I should add that we know which pulses are which - ie which ones come from
the laser, becuase we have a gate on the pulse train that only allows
pulses that arrive at the detector for the period 0-10ms after the laser
shot, giving a quiet 90ms interval befroe the next pulse, which is the
laser TTL pulse. All of the pulses - ie from the laser and the PMT are
generated by the same chip.
Thanks for your interest.
Andy