The AMI waveforms is expected by the Local Coincidence circuit

AMI Waveforms
Wisdom from  Thorsten, the author of DOM MB FPGA firmware:
  1. The transition from positive to negative should be fast. In the firmware I ignore slow transitions. They are probably noise. The transition should be within about 25ns. The width of the pulse itself is not critical.
  2. If you don't get coincidences try to reverse the polarity of the pulse you inject. The firmware ignores the negative to positive transition.

Building an AMI pulse with LeCroy 222 gate and delay generators

With difficulty, the product information can be found at the LeCroy web site manual link.

AMI pulse from NIM Modules
The module; the page

Building an AMI pulse with TTL Logic


See the data sheet for the family of device used... for instance for LS-TTL
AMI pulse from TTL devices

g. przybylski, lbnl, for IceCube