I did two different cuts and compared the tracks we got per uC after correcting deadtime. Deadtime was calculated by looking at the number of T3 triggers in the End of Run entry in the HALOG and the number of T3s ((D.evtypebits&8)==8) in the data and taking the ratio of the two.
My guess that this is so large is the tracking code skips over events that it deems "too busy". While we may be able to increase the number of tracks that we see at a higher current by playing with parameters a bit, after all things considered I don't think we're going to do much better than break even.
I think we should stay at 8uA.
A copy of this log entry has been emailed to: bogdanw@jlab.org, nl8@virginia.edu, holmstro@jlab.org
Figure 1