J Kissel, reporting for S. Dwyer, J. Hanks, G. Vajente

While making a routine parameter change in FM1 of the SRCL1 filter bank (a cdsFiltCtrl that can receive front-end input from a PD trigger; the change was to change the filter ramp time from 0.3 to 0.1 to 0.25, loaded by hitting the "individual" bank coefficient load), we found the filter started to do bonkers things. This filter, in FM1, is intended to be a +106dB (x200000) gain-only filter. However, with the input disabled, and an input offset of 0.1 and having only this filter bank on (under both front-end control and manual control) would result in an output of millions of counts -- and would still be varying as though there was an input. Also, the status lights underneath the filter request button were behaving confusingly as well -- the "requested" light could be manipulated to be on, but the "it's on" light would never flip green, OR (after trying to load several over different configurations) we could get it to hang with the "it's on" light to stay on, and the "requested" light was off / red. And those statuses did not correlate with the output. 

We found no other choice but to restart the h1lsc front-end process (with the help of Jonathan), which appears to have fixed the problem.

Note that this a new computer (recently upgraded two weeks ago), and we're running a new RCG (two weeks ago).