November 2020: Dry Run

Table of Contents

Introduction

In November 2020 the B2B system was tested during a dry run. The following main features were deployed
  • SIS18 and ESR kickers
    • trigger kicker control electronics ('fire the kickers')
    • basic data acquisition
      • time of output signal of kicker control electronics
      • time of kick by signal from kicker magnet probes
      • distribution of acquired data in hard-real time via the timing network; delivery to the customer at 1ms after the kick
  • integration into real accelerator schedule by listening from 'events' from the real Data Master
  • modes
    • trigger extraction kicker upon EVT_KICK_START
    • fast extraction from a ring in sync with the RRF H=1 group DDS signal
    • fast extraction of bunch from a ring and transfer to another ring ('bunch to coasting beam')
    • no bunch to bucket
  • simultaneous operation of CBUs at SIS18 and ESR

Setup

See here.

Log

date time what remark
2020-11-19   SIS18 kicker, trigger at EVT_KICK_START, acquire signals DONE
    increase energy of SIS18 kicker to a high value to check we have no issues with electronic noise DONE
2020-11-23   ESR kicker, trigger generated by B2B, acquire signals DONE
2020-11-26 14:00 Main Control Room sets up a coupled pattern for testing repeated every ~15s
    UNILAC -> SIS18 -> ESR -> Extraction 'to CRYRING' this allows to test individual features as well as the full chain from UNILAC to ESR extration
    SIS18: pattern "DRYRUN_SIS18_FAST_TE_ESR_20201123", SID = 1  
    ESR: "SL_ESR_DRYRUN20_2CRYRING_no_line", SID = 11  
    (lot's of activities by other people in parallel)  
2020-11-26 14:00 test both SIS18 and ESR kickers using the old 'timing generator' system DONE (check required to make sure the 'old' mode will work in the beam time 2021)
  14:30 switched over to B2B system  
  15:00 extraction from SIS18 upon EVT_KICK_START_1 (click)
  16:00 extraction from ESR upon EVT_KICK_START_2 (click)
  16:15 no H=1 signal from SIS18 group DDS; (click)
    group DDS is happy with another pattern (SID4=4), but dead with our pattern try to solve this with the help experts and control room, reason unclear
  18:00 give up on getting a signal form SIS18 group DDS, try ESR  
  18:15 ESR 'fast extraction' phase locked to H=1 group DDS with H=1 @ 871776 Hz (click)
  19:30 (leave things running til next morning to check the system runs reliably)  
2020-11-27 08:45 ESR 'fast extraction' till now, SIS18 group DDS remains dead for SID = 1 frown, sad smile (click)
  09:20 SIS18 group DDS is back! Resupplying ALL set-values solved the problem  
  09:25 SIS18 'fast extraction' phase locked to H=1 group DDS with H=1 @ 785294.9 Hz see 'bunch 2 coasting beam' SIS->ESR
    (ESR still runs with 'fast extraction' since yesterday evening)  
  09:27 'bunch to coasting beam' SIS18 to ESR (click)
    (ESR still runs with 'fast extraction' since yesterday evening) dry beam through the full chain from UNILAC to ESR extraction!
    (ESR kicker now 'kicks' at injection and extraction)  
  11:00 increase energy of SIS18 kicker to a high value to check we have no issues with electronic noise DONE
  12:30 collect statistics till now, end of dry run  

Extraction upon Event 'Kick Start'

drNov2020_SIS18EKS.JPG
Figure: Trigger SIS18 Kicker by EVT_KICK_START_1. Shown are LVTTL signals created upon EVT_KICK_START_1 (played by the Data Master) and CMD_B2B_TRIGGEREXT (played by the B2B System). Both events are played at the same time.

drNov2020_ESREKS.JPG
Figure: Trigger ESR Kicker by EVT_KICK_START_2. Shown are LVTTL signals created upon EVT_KICK_START_2 (played by the Data Master) and CMD_B2B_TRIGGEREXT (played by the B2B System). Both events are played at the same time.

No H=1 Signal from SIS18 Group DDS

drNov2020_noRF.JPG
Figure: No H=1 RF Signal from SIS18 Group DDS. Shown are H=1 group DDS signals (yellow), signal from comparator (blue) and LVTTL signals created upon events EVT_KICK_START_1 and CMD_B2B_TRIGGEREXT (green, pink). Note that the RF signal is off for about 1 second during the entire SIS18 cycles. This issue was finally solved by a new data supply ('Vollversorgung').

Fast Extraction ESR (B2E)

drNov2020_ESR18B2E.JPG
Figure: 'Bunch 2 Extraction' @ ESR (fast phase locked extraction). Signals shown are group DDS (yellow) and the trigger event for the ESR kicker (blue). Note that even after ~3800 extractions the maximum difference between the two signals has an upper bound of ~2ns, which demonstrates the reliability of the system. The standard deviation of the skew is ~350ps.

Bunch to Coasting Beam SIS18 -> ESR

drNov2020_SIS18ESRB2C.JPG
Figure: 'Bunch 2 Coasting Beam' from SIS18 to ESR. Signals shown are group DDS (yellow) and the trigger events for the SIS18 kicker (blue) and the ESR 'injection kicker' (pink). After 600 transfers the upper bound difference between the H=1 and trigger events is about 2.5ns, the standard deviation of the skew is about 500ps. This is a bit more noisy than the measurement of the night, but there is lot's of unrelated activity around (even drilling of holes)!

Kicker Signals and Diagnostics

Setup

The following figure shows the relevant signals for triggering the kicker and signals used for diagnosics.

kickerSignalsDiagnostic.png
Figure: Kicker signals and diagnostics. The figure on the oscilloscope screen show the following signals: Trigger from B2B (yellow), output from ACO electronics (blue), kicker magnet probe (magenta) and kicker magnet probe obtained via a capacitive voltage divider (green). Details see text.

The Bunch-2-Bucket system sends a timing message with CMD_B2B_TRIGGEREXT (not shown) to the Timing Receiver (TR) mounted in a 1U server. On-time, the TR creates a LVTTL signal (yellow) that is fed into a first set of electronics modules (this signal replaces the one from the famous 'Timing Generator'). The electronics performs certain tasks including the generation of a TTL signal (blue) that is fed into the main kicker electronics. When this happens, the kicker ignites and the kicker magnet is energized. A pickup probe allows to monitor the magnetic field at the kicker magnet. The probe signal is split and feeds various customers (magenta). The special feature of importance here is a capacitive voltage divider. A fast comparator is used to convert the output of the voltage divider signal to LVTTL (green). The bunch-2-bucket system uses the rising edges of signals for various purposes.
  • electronics out
    • signal is active: confirmation the electronics is supplied with settings data and the kicker is active
    • monitoring of the set-value of the delay that has been set for the kicker
  • probe signal
    • signal is active: indication, that the full kicker scenario is working
    • monitoring of the real kick time

Data, Analysis and Discussion

The B2B system acquires the timestamp of the rising edges of the 'electronic out' and 'magnet probe' signals. The intention is fast data acquisition, analysis and delivery of the data to the customer within 1ms after the kick. The purpose of the data is twofold: Firstly, they serve to monitor and diagnose the B2B and kicker systems. Second, the data is required, as an example, by our colleagues from beam instrumentation. However, the data acquired and distributed by the B2B are very basic and do not intend to compete with proper digitization of the analogue signals by other technical system at GSI/FAIR.

kickerSISMonitorSignals.png
Figure: Measured delays of the rising edges of the electronics out (blue) and kicker magnet (red) signals with respect to the trigger signal CMD_B2B_TRIGGEREXT. Please note, that the x-axis is not continuous and has nanosecond scale. For details see text.

Analysis
The figure above shows the observed timestamps [ns] of the electronics output signal and the kicker magnet probe signal. The data has been acquired in 520 extractions from SIS18 within ~2.5 hours of operation. In the analysis, the due-time of CMD_B2B_TRIGGEREXT has been subtracted from the data. Thus, each extraction provides two data points; the delay of the electronics output signals and the and delay of the beginning of the kick at the kicker magnet. Delays in (short) cables are not considered here. The data is then represented as histograms.
what kicker electronics magnet probeSorted ascending
corr std dev [ns] 1.02 1.82
std dev [ns] 1.08 1.86
# of transfers 520 520
minimum [ns] 2427 3675
average [ns] 2429.2 3681.2
maximum [ns] 2431 3685
Table: Statistics of kicker electronics and magnet probe delay measured using the b2b system. The corrected standard deviation has been obtained by quadratically subtracting a value of 0.35 ns for the standard deviation of the b2b system.

Discussion
The 1ns granularity of the timestamps acquired at the timing receiver is sufficient to measure the time distribution of the signals. Each of the two data sets is well confined in a small window of 6ns (11ns) for the electronics out (magnet probe) signal. The measured magnet probe signal is smeared out by a factor of ~2 compared the electronics out. This can easily be explained by the following.
  • the probe signal is an analogue signal
  • the amplitude changes if the kick angle in the ring is changed
  • the comparator is operated at a fixed setting
  • other customers are tapping the same probe signal
    • they connect or disconnect at arbitrary times
    • cables to the customers may be long; additional reflections do occur
    • the signals from the capacitive divider are not a 50 Ohm system which may lead to strong additional reflections
  • especially the so-called 'dumps' of the kicker system might cause additional electronic noise that is picked up. The dumps are 'fired' in less than 1 us before 'kicker ignition'. Especially when the kicker systems runs at high load, the pickup of electronic noise from the dumps can clearly be seen as a contribution to signals on an oscilloscope.

Altough the acquired timestamp data from SIS18 seem to be surprisingly precise, they should be used with care as can be seen in the following figure taken in the ESR kicker room. Here, two sets of signals are available, one for each kicker 'bank'. One of these is used for injection, the other for extraction. For one of these banks the signal looks clearn. But for the other bank, the signal from the capacitive resistor is very much distorted due to reflections of the electronics signal to various customers.

kistTestNov2020_ESR.JPG
Figure: Signals in the ESR kicker room. The probe signal itself is clean (yellow). The signal from capacitive resistor (blue) is perturbed by a multitude of reflections. Signal at the output of the comparator (green).

-- DietrichBeck - 4 Dec 2020
I Attachment Action Size Date Who Comment
drNov2020_ESR18B2E.JPGJPG drNov2020_ESR18B2E.JPG manage 547 K 27 Nov 2020 - 17:43 DietrichBeck ESR - bunch to extraction
drNov2020_ESREKS.JPGJPG drNov2020_ESREKS.JPG manage 640 K 03 Dec 2020 - 11:26 DietrichBeck ESR: trigger kicker by event
drNov2020_SIS18EKS.JPGJPG drNov2020_SIS18EKS.JPG manage 602 K 03 Dec 2020 - 11:21 DietrichBeck SIS: trigger kicker by event
drNov2020_SIS18ESRB2C.JPGJPG drNov2020_SIS18ESRB2C.JPG manage 669 K 27 Nov 2020 - 17:48 DietrichBeck SIS->ESR - bunch to coasting beam
drNov2020_noRF.JPGJPG drNov2020_noRF.JPG manage 1 MB 03 Dec 2020 - 11:44 DietrichBeck SIS: no RF signal from H=1 group DDS
kickerSISMonitorSignals.pngpng kickerSISMonitorSignals.png manage 37 K 03 Dec 2020 - 17:29 DietrichBeck SIS: kicker monitor signals
kickerSignalsDiagnostic.pngpng kickerSignalsDiagnostic.png manage 640 K 03 Dec 2020 - 16:59 DietrichBeck SIS: kicker signals and diagnostic
kistTestNov2020_ESR.JPGJPG kistTestNov2020_ESR.JPG manage 705 K 03 Dec 2020 - 18:39 DietrichBeck ESR: kicker monitor signals
Topic revision: r6 - 22 Feb 2021, DietrichBeck
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback