Without being able to test it, here's two possibilities, whose implementation would be up to you:
1) By multiplying clamp(...) with timedelta (or a multiple of timedelta). The problem with it is, that while the angle to the mouse might change in big steps, the reaction from the turret would be slow, maybe too slow.
2) By using lerp(). lerp basically sets some intermediate value between 2 known values, based on a t-value (the third parameter). lerp(1, 4, 0.25) would result in 1.75 (25% of the way from 1 to 4). If you use it like
set angle to lerp(.angle, clamp(...) + Green_1.Angle, n * TimeDelta)
it will also turn over time (with n being a number to test, the higher, the faster the turning)
I have never worked with it, but there's a turret behavior. Is it possible that it solves some of your issues?