There are two tramps in the LONG_SEARCH_IR state in ALS_DIFF for when we are trying to find IR, I've changed these to 1 second when stepping around a known value and to 2 seconds when transitioning to a new known value. Previously these were 4 and 10 respectively. 

Years ago I tested this and found that I couldn't rush this pzt, but I didn't document why or to what extend. I've been frustrated with how long it has been taking, so I wanted to try it out again and see if I could speed it up and/or find out why I couldn't before. I couldn't really create any issues, but perhaps they only rear their head up when we aren't the best aligned. I ended up bringing them down to the values listed above with repeated success so I'll leave it there and continue locking.