AdV-DET (Commissioning)mwas - 12:31 Wednesday 09 November 2016 (35413)
Print this reportB1p quadrants galvo loopsWorked a bit on the B1p galvo automation to try to make it smart.
A week ago (and for other galvo loops), if the correction saturates the loop opens and require to open and close the loop manually to reactivate it.
Now the loop in some conditions will close itself after the actuation saturates using the GUARD_SOFT switch. This guardian is automatically reset if the beam disappears or moves from far away to with reach of the galvo loop, but it should keep the loop open if it is just on the edge of the galvo range, to prevent the loop to be oscillating from open/closed state all the time.
The loop also checks that the beam is not too misaligned, if the beam is very misaligned (above +/-2 on the normalized DC signals like SDB1_B1p_QD1_H_norm) there is no point in trying to close the loop, the galvo will saturate before getting to the center of the quadrants. This automated loop opening/closing is done separately for vertical and horizontal.
I am not sure if that more complicated logic is better. It means it isn't required to reset the loops each time the beam is moved away. But it may oscillate in some scenarios and be more confusing to diagnose.
Had a quick look at the galvo loop bandwidth, it seems to be 2-3 Hz, but it is a bit confusing as the beam moves a lot with SDB1 under air-flux.
Comments to this report:
mwas - 14:23 Wednesday 09 November 2016 (35415)
Print this reportThis new automation setup has worked quite well for the past 2 hours, with many changes in alignment due to other work. Each time the loops opening and closing back when the beam was available.
Increased the galvo loop gain, from 0.2 to 1.5 for B1p_QD1, and from 0.2 to 1.0 for B1p_QD2. These higher gains are a factor 4 below oscillation at ~20Hz, where the open loop phase transfer function goes through the zero.