The diagrams of Figure 1 give the working range of Hall B orbit lock in terms of orbit errors at IPM2C22A and IPM2C24A. The 2 plots on the left show the "correctable" orbit errors as the area inside the boxes for x & y. The coordinates are, respectively, the orbit error at IPM2C22A, and the differential orbit error between IPM2C22A & IPM2C24A. The latter is usually a small number due to the short distance between these 2 BPMs, and imposes the tightest constraint on the effective lock range for the same reason. If the lock is started with both correctors at 0 strength and orbit errors on both BPMs being 0, at least one lock corrector will max out when the orbit errors wander out of the box.

The 2 plots on the right show a magnified view of the above, with the correctable orbit errors located between the dashed lines. It can be seen that the lock can handle differential orbit error between the 2 BPMs of about ± 2 mm in x and ± 1.5 mm in y, when the orbit at the first BPM is at GOLD position. This range shifts and changes size as the orbit moves away from GOLD. One can nonetheless expect a correctable range of ± 1 mm in the differential orbit error in both planes when the orbit at IPM2C22A wanders by no more than ± 5 mm from GOLD. This is a pretty comfortable range given the distance between the 2 BPMs and the nA BPM resolution. Also, if the orbit error is transient and does not persist over 25 seconds, the lock cycle time and fudge factor will mitigate the effect on the correctors.

Notice in the above we assumed the lock correctors start at 0. Figure 2 shows how this range can change if the lock correctors are not 0 when the BPM readings are loaded into GOLD. In this light there is strong interest to set lock correctors as close to 0 as possible when "load absolute to GOLD" happens. Using 2C22H, 2C23V and correctors upstream of the lock to relieve the lock correctors would help, provided there is no other adverse orbit consequences.

For the same reason, setting the GOLD orbit to arbitrary values and expecting the lock to automatically fix the orbit is a dubious practice for this particular lock. The safe way to do it is to use the other correctors mentioned above to arrive at a desired orbit (while keeping the lock correctors as close to 0 as possible), load absolute orbit to GOLD, then run the lock.