Fixed y2017 tests.

The 5 ms timestep -> 5.05 ms timestep change didn't propegate fully
and the tests caught a bug.  I need to go re-test this on a robot
before submitting.

The gear ratio also changed so disabled zeroing was moving too much.

Change-Id: I8add457559cc3b46a6664d4339717bfcb0271787
diff --git a/y2017/control_loops/superstructure/shooter/shooter.cc b/y2017/control_loops/superstructure/shooter/shooter.cc
index 795a7d4..03abd99 100644
--- a/y2017/control_loops/superstructure/shooter/shooter.cc
+++ b/y2017/control_loops/superstructure/shooter/shooter.cc
@@ -67,10 +67,7 @@
   // Compute the distance moved over that time period.
   average_angular_velocity_ =
       (history_[oldest_history_position] - history_[history_position_]) /
-      (chrono::duration_cast<chrono::duration<double>>(
-           ::aos::controls::kLoopFrequency)
-           .count() *
-       static_cast<double>(kHistoryLength - 1));
+      (0.00505 * static_cast<double>(kHistoryLength - 1));
 
   // Ready if average angular velocity is close to the goal.
   error_ = average_angular_velocity_ - loop_->next_R(2, 0);