Raster scope monitor and controls - See Also Hall A Wiki Page
- 1 Where the raster scope is
- 2 What a raster is
- 3 How the raster should be set
- 4 Beam/60Hz Line Synch
- 5 What the Raster Means in Counting Mode DAQ with HRS channels
- 6 Raster Calibration 2019-06-28 11:30 PM
Where the raster scope is
The raster scope is located in the 2nd from left rack of electronics in the electronics room and is about 3 feet off the ground. It is an X Y scope and should have some pattern displayed on it whenever the raster is on. It is displaying the X and Y currents going through the raster magnet coils. The raster amplitude set values should be displayed in the Hall A General Tools GUI and also the Target GUI.
What a raster is
Raster is the German word for "screen", which was incorporated into English with the advent of scanning CRT TVs.
The electron beam in the CRT (CEBAF) is pushed by electrodes (magnets) so that it goes from side to side. For the parity quality experiments we want to prevent the beam from blasting a hole in one single spot in the target, and so we have the beam scan from side to side rapidly to distribute heat across it.
The Hall A Raster is two sets of two sets of coils, of which only one is in use for the low energy Prex/Crex experiments: one in X and the other in Y. Their frequencies are ~25kHz, meaning that they trace a continuous diagonal (nearly 45 degree) Lissajous figure.
How the raster should be set
Their frequencies should be mutually harmonic, meaning that the difference in frequencies between them is some multiple of the helicity window flip rate. The reason for this is so that the pattern traced in X and Y will repeat itself an integer number of times in each of the helicity windows of one Quartet (or other multiplet) pattern of windows used to calculate the asymmetries.
This is important to avoid having differences in the integrated electron beam position between helicity flip windows (that could either broaden our asymmetries or introduce false asymmetries from pick-up noise in the raster electronics itself).
Updating Raster Sync with Function Generator
The raster was set up on 5/30/2019 (halog) - Reference photos: Function Generator Locking on Scope, Raster Current Readbacks on Scope, Raster X vs Y on a 2D scope trace, What the Agilent Controls look like, The Raster Current set values controlled by knobs (experts/MCC only),
To sync the MPS signal to the raster we use an Agilent 33500 signal generator. It has a 10MHz input to sync the outputs with the signal coming in. The raster will utilize these 10MHz clock based output frequencies to determine its X and Y ramp frequencies. The output of the signal generator is synced to the MPS 120Hz signal from the injector manually by putting the MPS signal into a scope and tuning the Agilent function generator until both output frequencies (X and Y are different by an integer multiple of MPS) are locked w.r.t. each other and the 120Hz signal.
This setup is stored in the memory of the Agilent Function Generator and can be recalled by doing the following:
- Press the system button.
- Under the screen press the save/recall button.
- Press the recall state button.
- Look down the list and select the TREX-30Hz.sta button and press select and the state will automatically return.
Beam/60Hz Line Synch
The helicity multiplet start is synched with the 60Hz line power phase - this means that the last window in a multiplet will be shorter or longer than the first windows (randomly, on the order of 35us jumps between multiplets), and the Lissajous figure will start at a slightly different phase on each new multiplet (which for Prex is every 30Hz) which will make the raster pattern not stand still.
Also the "60Hz" timing can get generally shorter or longer over the course of a day, meaning that the "good" mutual harmonicity of the X and Y coils can "go bad" over the course of a few hours. This is potentially very bad, especially when the target densities are degraded by melting in the center (the fate rastering is intended to avoid).
Possible tuning to an independent HelBoard's 120Hz
It is possible to set up a circuit with a pair of 794 gate generators to hand-produce a stable and ~120Hz or ~60Hz fake "beam synch" signal and convert it into the optical signal we need. Thankfully we commissioned a send/recieve fiber/TTL converter box and can get more if needed. This hand-made "beam synch" can be plugged into a spare Helicity Control Board that we have in the counting house VME crate and use it to generate a much more stable MPS signal with no back and forth jitter, and nearly perfectly synched to 120 Hz, within a few microseconds truncation/extension on the Nth multiplet window length.
This significantly more stable and hopefully accurate copy (assuming all helicity boards have the same 120Hz definition too) of the intra-Multiplet helicity window timing can now be used to actually tune the Agilent Raster X-Y frequency generator to a safe helicity window-harmonic Lissajous configuration. Though it will still be the case that the phase of this Lissajous pattern w.r.t. subsequent multiplets of the injector HelBoard's actual beam synch will definitely precess around as its Nth multiplet helicity window changes its time.
HelBoard software update
See also - Helicity Control
The HelBoard software has been updated (with advice from Gary Croke) - The Counting House VXWorks Bootscript's HelBoard.o executable is now in a new ~/devices/helbrd-sync/ copy folder. This update allows for VXWorks command line user editing of the "Clock Mode" register - now we have a 5th register to control the beam synch mode (beam synch vs. free clock) and its fixed output frequency (30, 120, 240):
- WriteHelBoard(4,0) -- 30 Hz
- WriteHelBoard(4,1) -- 120 Hz
- WriteHelBoard(4,2) -- 240 Hz
- WriteHelBoard(4,3) -- Free Clock
All require some clock signal to be plugged into the front, but these are the exact frequencies that it will generate in the first N-1 multiplet windows before resetting the QRT trigger (I think... maybe it resets QRT/relative MPS trigger on each external "beam synch"?)
Note that in Free Clock mode the user has control over Tsettle and Tstable, while in Beam Synch mode the user only has control over the relative amount of the helicity window that is occupied by Tsettle. In both modes the user can control delay and multiplet pattern type.
What the Raster Means in Counting Mode DAQ with HRS channels
Raster Calibration 2019-06-28 11:30 PM
The Raster Frequencies before tuning the the Helicity Flip rate had been set to approximately the correct settings, but this was done during the beam sync mode of the control board. Beam sync makes it hard to see the relative phase drift of the raster frequencies w.r.t. the flip rate as the Nth (4th) helicity flip period gets a different integration time in order to allow the following multiplet to latch on the start of beam synch.
The raster settings before changing them were :
- Channel 1 (yellow on the Agilent) - 25.07478380 kHz
- Channel 2 (green on the Agilent) - 24.95511650 kHz
I think that channel 1 corresponds to the Y and 2 to the X raster currents (as these are the labels of their cables through patch panel land), but what X and Y mean in terms of the effect we see on the spread of the raster in the Counting Mode DAQs is still a matter of opinion on coordinate frame nomenclature choice (see raster current in Counting Mode DAQ calibration studies from Catherine:https://logbooks.jlab.org/entry/3695377)
I updated the raster synch during a free clock 120 Hz test. Set helbrd top 120 Hz (with 100us Tsettle, Tstable to 8233.35us)
Raster Frequency Synching
We want to set both frequencies to an integer multiple of 120 Hz and have their difference be 8x120 = 960 (previous experience says this may be optimal) 25kHz/120 = 208.33, so 24,960 is the sweet spot multiple of 120 I pick 25440 and 24480
New Settings taken during 100 us Tsettle and 8233.35 us Tstable
- Channel 1 (yellow) 25.439949 kHz
- Channel 2 (green) 24.479949 kHz
- The important observation (by eye) is that after performing this sync I see that both x and y raster signals are static (no phase drift) w.r.t. the 120 Hz helicity flip rate and w.r.t. each other. - When I first tuned them to those nice integer numbers above, I saw both were synched to each other, but that they drifted w.r.t. the 120 Hz helicity frequency. - In order to correct for this drift I tuned both frequencies the exact same amount downwards (same amount in order to preserve the difference of 8x120 Hz, and hopefully being a bit low on both will not mess up anything. This was a choice made by Cameron and Bob together)
Data taking on June 29 OWL: - After we tuned the raster we left the helicity control board in free clock mode (under clearance from Kent to do what we like) - After tuning the raster we chose to leave the trigger timings all at the prior 90us tuned parameters (I think only flexio latch was based on Tstable, so this has minimal effect) - If we still see residual correlations in the Main detectors we should try:
- Going back to beam synch with this tuned raster
- Tuning the raster to those nice integer numbers from before instead in free clock
- Same, but with beam synch
Pictures of DAQ/Apparatus after test: media:XYfrequencyScope-29062019.jpeg, media:XYcurrentScope-29062019.jpeg, media:XYvoltages-29062019.jpeg, media:AgilentChannel1-29062019.jpeg, media:AgilentChannel2-29062019.jpeg
Saved settings into "120_Free_Tset" save point