I use the same method as James, you can see the pictures in the attachment. first picture the slow raster is 1.1V, the size of the hole in the middle is 1cm, so the size of slow raster = (4751-1079)/(4102-1833)*1cm=1.6cm second picture the slow raster is 1.5V, the size of the bottom hole is 1cm, so the size of slow raster = (5428-471)/(4032-1746)*1cm=2.18cm I did the same calculation on yesterday morning, the result is matched James' result. I also checked the slow raster power supply's magnet current for 1.5V today and yesterday, it is same, both is 5.57373A for x and 5.40225A for y. you can check it in the start of run in halog. So the only reason that slow raster size change is the chicane and target field changed. yesterday is at 2.5T configuration and today is at 5T configuration, seems target field will affect beam size.