Ramp Monitors

1
Ramp Monitors

• Tom Russo

2
Objectives for Run 7

• We are proposing two systems for improved
  monitoring during the ramp.
• The first is an addition to the existing BBQ
  hardware which will improve the existing
  turn-by-turn ramp monitor. Key point is it will
  require no expert intervention for operation.
• The second system will create a bunch-by-bunch
  turn-by-turn ramp monitor.
• typical data from SPS
• The system for Run 7

3
Motivations

• Changing from an expert only system
• Advances in commercial digitizer technology
• 8 channels x 24 bits at revolution frequency - 78
  KS/s
• 2 channels x 14 bits at bunching frequency
  9MS/s
• Advances in memory and data bus capabilities
• terabyte hard drives
• 40MS/sec data streaming to hard drive
• Advances in pickup sensitivity
• 3D AFE (Direct Diode Detection)
• 1m stripline with motion control
• Take advantage of these advances
• full ramp TbT data with high sensitivity time
  and frequency domains
• full ramp BbB TbT data with medium sensitivity
  t and f domains
• Complement to existing diagnostics MBPM etc
• We have prior experience with NI DSAs and the
  transfer of FFTs to the control system
• Simplified timing (no swept LO, no line hopping
  etc)

4
The turn by turn system as implemented for Run 6

• 3D AFE our most sensitive pickup
• at baseband
• no bunch-by-bunch information
• no information within the bunch (however,
  spectral signature of mode coupling/fast
  head-tail instability may be visible)
• peak detects sees hottest bunch
• HP 89410 DSA
• capture mode cant FFT fast enough to do this
  real time
• no interface to Control System, no stones, no
  archiving,
• display was screen dumps of post-processed FFTs
• time domain data not conveniently available
• limited record length cant capture full ramp
• Specialist operation only

5
Transition Instability (weak) as seen by 3D
AFEduring Run 6- role of mains harmonics in
'seeding' instability?- sidebands characteristic
of coupled mode/fast head-tail?

• gt

.0025
½ synchrotron period
6
New and improved turn-by-turn system for Run 7

• 3D AFE our most sensitive pickup
• PC with 8 channel NI 4472 DSA
• trigger on accramp
• use 4 channels for the 4 measurement planes
• use 3 channels for gammat, lumi, beam abort
• time capture with FFT post-processing with Lab
  View
• SNAP interface to Controls
• Deliver 3 (x4) data sets
• time domain around transition
  (100KS/s x 5s x 4 channels x 3
  bytes/sample 6MB)
• high time resolution FFT around transition (3MB)
• high frequency resolution FFT up the ramp (??MB)
• Archive all raw data on local 1TB hard drive

7
(No Transcript)
8
Turn-by-turn system for Run 7
1m Stripline BH BV YH YV
24 bit A/D ch1 ch2 ch3 ch4 ch5 ch6 ch7 ch8 trig
ger
3D AFE BH BV YH YV
Control System GPMs, archiving,
SNAP
V108 gammat lumi abort spare accramp
local 1TB hard drive (archiving)
Windows PC running LabVIEW
9
BbB implementation for Run 7

• Goal is to provide bunch by bunch, position vs
  time information, for all the bunches in the
  machine.
• Utilize an existing 1 meter strip line and a
  commercial fast 2 channel digitizer.
• Similar to a system presently used at the SPS

10
(No Transcript)
11
(No Transcript)
12
x3 decimation
28MHz RF
NI 5122 Block Diagram
13
Bunch-by-bunch turn-by-turn system for Run 7
1m Stripline BH in BH out BV up BV down YH in YH
out YV up YV down
D hybrids
HP VME Mux
14 bit A/D ch1 ch2 clock/3 trigger
Control System GPMs, archiving,
SNAP
28MHz RF
local 1TB hard drive (archiving)
optional ext trigger (Artus,)
Windows PC running LabVIEW
14
BbB TbT summary what we get from this

• Robust time and frequency domain transition data
  for every ramp, 2 planes (take your pick)
• 1m pickup for sensitivity, motion control for rev
  line suppression
• 14 bits for dynamic range
• beam synchronous digitizer clock (x3 decimated
  28MHz RF)
• no need for swept LO, line hopping, (unlike
  Schottkys)
• continuous streaming of data to local disc at
  20MS/sec
• Additional possibilities ??