Post date: Jun 28, 2015 10:51:38 PM
Two main items today: 1) testing Shinji's RF files and looking at the pattern of beam loss, and 2) starting tune measurements
1)
Setup: injected bunch length - 50 microsec
Corrector 900A
F: 814/ D: 825
Initially Shinji's files used -0.1msec before the nominal 11 MeV energy whereas the KURRI team used -0.5ms. We observed that there was a large difference in the amount of bunch charge captured and accelerated between the two cases. We made two attempts to rectify this.
a) increase the LINAC tank voltage to increase the energy of the incoming beam a bit, to see if capture increases in case the injection energy was a little low. (Adjusted 5.1->5.6 main voltage, but this made no difference, reset to nominal)
b) add the extra 0.4ms in Shinji's file (Kuriyama-san did this by just changing the timings from 0.1ms to 0.5ms in the files). This time we recovered the same transmission as with the KURRI file for the same setting. This was successful, so we adjusted all 4 files to be tested to have the same -0.5msec offset.
We tested all 4 files (constant/variable k with 20/30 degrees phi_s). The difference between constant k and variable k in each case was negligible, but there was a big difference when we changed phi_s.
The images of the oscilloscope for each case are here: https://drive.google.com/open?id=0B0TZJx0b9o5zfkNqQk1uLXpsYlhfa2RXanV0WGR1Qm5IeFZFaVVsdUZoWEhnQ09rQjIxczA&authuser=0
The .csv files were also saved, and can also be linked later.
2) In the afternoon we started looking at betatron tune studies. We initially attempted this with the horizontal perturbator and we saw a lot of beam loss on the device itself (ie. the beam was large). We attempted to use the radially movable BPM to see the sidebands on the RSA but without success. Instead the triangle plate array was used for horizontal oscillations and S12 for vertical.
Injected bunch length: 3 microsec
Perturbator: 1900 sinusoidal cycle 'burst' at a frequency adjusted to see sidebands.
Initially attempts were made to find the tune with an accelerated beam, but eventually the normal flat-top scheme was used as this was easier to find.
We only had time to take a couple of different energy points in the afternoon. The initial vertical tunes came out much lower than before at about 1.25 while the horizontal were similar (nb. we are using a different working point 814/825 so we expect this, but they were perhaps lower than anticipated).