BPM6W1 Electronic Logbook 2006

26 Dec 2024 - 14:13

Hardware Layout

• Diagram of hardware layout:
  
• This diagram in pdf format: bpm_bsm_flm_PAC_diagram.pdf

January 2006

January 16, 2006 (J. Dobbins, C. Strohman, E. Tanke): DSP BPM6W1 mounted in box with heat sinks

• The DSP BPM6W1 has been mounted in a box together with heat sinks and a fan (the latter mounted on the outside of the box). In the future the fan speed will be controlled as a means of temperature control. With the DSP box in R101, from a "cold start" it takes about 40 minutes to reach its operating temperature (see the plots below).
  Error: (3) can't find bpm_temp_const_adc.jpg in ACC/Bunch
  Error: (3) can't find bpm_temp_const_tim.jpg in ACC/Bunch
• Temperature fluctuations at the operating temperature are now about 1 deg C peak to peak (was 2 deg C before the heat sinks were installed). In the future the fan speed will be controlled as a means of temperature control.

January 25, 2006 (E. Tanke): Timing calibration with single bunch e- (T1B1)

• With ADCclock=79 and gain=6 one finds the e- peak at TglobA=800 with the following channel timings: ch0=682, ch2=694, ch4=697 and ch6=710.

March 2006

March 13, 2006 (J. Codner, M. Billing, E. Tanke): Study of Differential Vert Orbit Shifts with SQ's Disabled

• Details of measurement are on the CESR MS web page
• Used option 5 in BPM_V1 program with ADCclock=79,gain=6,TglobA=TglobB=800, channel timings 682,690,694,697,697,717,710,720 and averaging over 1000 turns. Results are in the [cesr.palmer.ts101.bpm.saved_data] directory.
• Files #00135 correspond to data taken over ~ 15 min (first test), 18k averages 20k turns apart
• Files #00136 correspond to data taken over ~1.5 min (second test), 1.8k averages 20k turns apart

May 2006

May 16, 2006 (M.Palmer et al): Septum effect machine studies

• 1 bunch of electrons at 1.7 mA, adjusted the timing of the Vert pinger to be roughly at the time of the septum. TIM TRANSFER 21 = 17947. BSM measurement are described on the BSM23W twiki page.
• VERT PINGER ON cmd = 100
  • BPM Type II multi-file 00323
  • BPM Type II single-file 213
  • BPM Type I 3053
• Adjusted Coupling SQ4W = -765 --> 546, SQ4E = -1111 --> 289
• VERT PINGER ON cmd = 100
  • BPM Type II multi-file 00326
  • BSM file 1946
  • BPM Type II single-file 214
  • BPM Type I 3054
• E Septum 3 ON cmd = 1355
  • BPM Type II multi-file 00329
  • BSM file 1952
  • BPM Type II single-file 215
  • BPM Type I 3055
• Both OFF
  • BPM Type II multi-file 00331
  • BSM file 1959
  • BPM Type II single-file 216
  • BPM Type I 3056
• Preliminary results for the BPM Type II multi-files are in MS16May2006

June 2006

June 12, 2006 (J.Burrell): BPM6W1 pedestal values as a function of gain. Note that pedestal values are in "raw, 16-bit" units, the ADC is 14 bit.

• Experiment1.ppt: RMS Values wrt Gain as well as other variations.

June 29, 2006 (J.Burrell): Clock Setting Calibration.

• ClockSettings.pdf: Calibration for Lab (rm 215) Test BPM wrt Time Settings.

July 2006

July 07, 2006 (E.Tanke): Erroneous data on the first turn

• Not only on the last turn, but also on the first turn one can find erroneous data. Elimininating data for the first turn and the last turn when looking for possible "good" clocks, dozens of good clocks are found (millions of turns without error); this as opposed to one or no good clocks at all when eliminating only the last turn.
• The BPM DSP routines dsp_get_pedestal.c and dsp_pedestal.c in the bsm_v1 directory have been modified such as to add 2 turns to the number of turns requested, whereby the analyzed data ignore the first and last turns. Idem for the bsm_get_raw.f controls routine in the [CESR.PALMER.TS101.BPM] directory.

July 10, 2006 (E.Tanke): Erroneous data on the first turn (bis)

• The BPM DSP routines dsp_get_navg.c, dsp_get_lavg.c and dsp_get_apos.c in the bsm_v1 directory have also been modified such as to add 2 turns to the number of turns requested, whereby the analyzed data ignore the first and last turns.

July 13, 2006 (J.Burrell): RMS Calculations. Note that pedestal values are in "14-bit" units.

• RMSCalc.pdf: RMS Calculations wrt differing timing settings (2.75k turns and 65.5k turns).

July 14, 2006 (J.Burrell): RMS Calculations. Note that pedestal values are in "14-bit" units.

• 800800Calc.pdf: RMS Calculations wrt differing Gain settings (800/800 Timing).

• 200800Calc.pdf: RMS Calculations wrt differing Gain settings (200/800 Timing).

July 17, 2006 (J.Burrell): More RMS Calculations

• RMSCalcvGain.pdf: RMS Calculations wrt differing Gain settings (different look).

July 17, 2006 (J.Burrell): Histoplots

• Histoplot1.pdf: Channel 0, 2nd Board Still disabled, 800/800 Timing Settings.

• Histoplot2.pdf: Comparison of Channel 0 with 2nd board unenabled and enabled. (Also, first Ch.1 Data)

July 18, 2006 (J.Burrell): Channel 0 comparison.

• ch0.pdf: Comparison between Channel 0 with the 2nd board enabled as well as disabled.

July 18, 2006 (J.Burrell): First RMS Calculations for room 101. Note that pedestal values are in "14-bit" units.

• r101RMSCalc.pdf: Channel 0 RMS Calculations. (183 Bunches and 5 Bunches)

July 24, 2006 (J.Burrell,E.Tanke): Bunch Testing.

While running tests on the Gain 14 setting for the room 101 test BPM, we ran into rather strange findings. Although details still aren't clear as to why there is such a fluctuation in RMS values, the first tests are available. In these tests, with a variance of bunches selected, we found a fluctuating RMS value anywhere from 42 to 60. Certain scenarios were tested, such as every other 30 bunches, and found that the major problem occured when a series of 20 bunches were selected. Here are a few examples of tests ran. More details will be obtained in the future.

• BunchTests.pdf

July 25, 2006 (J.Burrell): More Bunch Testing. Note that pedestal values are in "raw, 16-bit" units, the ADC is 14 bit.

The picture above is a culmination of both today's and yesterday's tests. In it, we see both the high and low RMS values for each number of bunches omitted. Diamond shapes represent the high average of RMS values and empty squares represent the Low average RMS values. The full squares represent tests where the high and low averages were the same. For more detail, please read further.

Today's set of tests consisted of running on all of the odd pairings. Since we just ran the situations where we had the last 21 bunches, 20 bunches, no bunches and 1 bunch tests, the plan was to run all of the odd bunch tests to see what would be produced. After running these tests from 19 bunches omitted to 3 bunches omitted (i.e. 164-182, 166-182, etc) it became apparent that all of the odd numbers produced rather smooth RMS values. With this, I used a bit of information from yesterday (where we got our funny values at 20 and 10 bunches omitted) to test all of the even number of bunches left (i.e. 18-2). After doing this, I produced nothing but the irregular data. Here are all of the tests.

• BunchTestsv2.pdf

July 26, 2006 (J.Burrell): More Bunch Testing...

Today's tests included bunch testing for both Gain 14 and 6 for channels 6 and 7 (note: these tests were run on the room101 BPM). First, we found the two to three best clock settings for channels 0 and 1, then re-tested the bunches (20 missing) for the two gain settings. When these were found to be fruitless (aka, the same patterned error), we tested in the same fashion for channels 6 and 7. These tests were also found to have the same type of errors. The theory at this point is that there may be a problem with the first board (channels 1 and 2), which may be causing interference in the other boards. With this theory, we tested on the Lab215 BPM. Running the same types of tests with clock setting 79, both channels 0 and 1 showed to be normal. The next test was to install this card on the last board (6 and 7), but very wild data was produced. More tests will be run tomorrow.

July 28, 2006 (E.Tanke): Further investigation of the pedestal problem (see July 26)

• Have taken the BPM6W1 upstairs; took the DSP out of the box and made a series of tests, which can be found in this PPT file or PDF file

August 2006

August 02, 2006 (J. Burrell): Pulse Tests

Conducted pulse tests with respect to gain as requested by J. Dobbins. Single Bunch, Single Channel, 1000 turns, 10-step timescans. Two different graphs are shown for each respective gain. 165 mV, 12 nm width.

Pulse Gain Tests

Next, tests were run for both channels 1 and 2 (one on the same card, one on a different card). Same pulse characteristics. No significant differences shown. Here is a summary of the linearity for the three tests.

Linearity wrt Gain

August 03, 2006 (J.Burrell): Linearity Tests

Conducted ADC Count tests wrt pulse for gain 6, channel 0. Here is a summary of the values, showing a rather linear output.

Linearity Tests

October 2006

October 02, 2006 (E.Tanke): New BPM_V1 executable

• There is now a new bpm_v1 executable, which generates cur_raw.dat and cur_nav_xxxxx.dat files that contain more info about the machine status and saveset numbers. The "old" executable has been renamed BPM_V1_15SEP2006.EXE
• Note that the [cesr.palmer.ts101.bpm.dev] directory is the development area for this code

October 31, 2006 (E.Tanke): Testing of BPM6W1 in R101, relays connected (EPT)

• Two types of scans:
  • Time scan: with fixed channel timings, scan global timings
  • Channel time scan: with fixed global timings, scan channel timings
• Single e- bunch, not in HEP
  • Time scan, gain 6, 3 mA, pretzel on, file 1522
  • Time scan, gain 6, 4 mA, pretzel off, file 1523
  • Channel time scan, gain 6, pretzel off, file 1524
• Single e- bunch 4 mA, HEP, pretzel off
  • Time scan, gain 6, file 1525
  • Channel time scan, gain 6, file 1526
  • Time scan, gain 8, file 1527
  • Channel time scan, gain 8, file 1528, 1529
  • Time scan, gain 10, file 1530
  • Channel time scan, gain 10, file 1534
  • Time scan, gain 4, file 1536
  • Channel time scan, gain 4, file 1537

November 2006

November 01, 2006 (E.Tanke): BPM6W1 recabled

• BPM6W1 has been recabled such as to have the BPM cables connected directly to the DSP module (bypassing the relays).
• Apart from new timings (see below), with the signal cables directly connected to the DSP box also the PBUNCHES.DAT and EBUNCHES.DAT files need to be changed. The reference ones are now stored in PBUNCHES_NORELAY.DAT and EBUNCHES_NORELAY.DAT respectively. The old ones, for when we use the relays, are PBUNCHES_RELAY.DAT and EBUNCHES_RELAY.DAT respectively. These are in the [CESR.TS101.BPM] directory.
• Suggested timings for GAIN 6 are:
  • TglobA=TglobB=1090
  • Tch0=682
  • Tch1=700
  • Tch2=697
  • Tch3=708
  • Tch4=700
  • Tch5=729
  • Tch6=710
  • Tch7=731
• Note that the even channels have been optimized for e-, the odd ones for e+
• These timings were established during multi-bunch conditions; fine tuning still needs to be done under single bunch conditions.

November 03, 2006 (E.Tanke): Channel timing scans

• Channel timing scans taken under CHESS conditions and results from November 1,2006 can be found here

November 06, 2006 (E.Tanke): RMS of positons under CHESS conditions

• Channel timing scans were taken under CHESS conditions for GAINS 6 and 8, yielding optimum channel timings for this gain; results are stored in the [CESR.TS101.BPM.DEV] directory. Measurements are for T1B1 e- and T1B1 e+
  • Optimum channel timings for gain 6 are in BPM_V1_PROMPT_G6_NORELAY.INI
  • Optimum channel timings for gain 8 are in BPM_V1_PROMPT_G8_NORELAY.INI
• 1k raw data was taken simultanuously for e- and e+ under CHESS conditions, ~ 4mA/bunch. Positions were calculated without pedestal subtraction (pedestal subtraction increases the RMS; see here. Data was taken for gain 6 and gain 8. Note that for gain 10 with these bunch currents one starts to saturate (see here).

	electron RMS on positions		positron RMS on positions
file	RMS_H (mm)	RMS_V (mm)	RMS_H (mm)	RMS_V (mm)	Comment
1607	0.1024	0.0863	0.1324	0.1211	gain 6, HW trigger=0 (no bit set)
1609	0.1076	0.0877	0.1452	0.1241	gain 6, HW trigger=1 (bit 1 set)
1610	0.1057	0.0925	0.1392	0.1234	gain 6, HW trigger=2 (bit 2 set)
1606	0.0587	0.0419	0.0729	0.0607	gain 8, HW trigger=0 (no bit set)
1608	0.0618	0.0439	0.0775	0.0646	gain 8, HW trigger=1 (bit 1 set)
1612	0.0667	0.0442	0.0766	0.0653	gain 8, HW trigger=2 (bit 2 set)

November 07, 2006 (E.Tanke): RMS of positons under CHESS conditions, single bunch, separators off

• Single bunch of e-, 0.5 mA
  • Timescan at gain 8, file 1641
  • Timescan at gain 10, file 1643
  • Timescan at gain 12, file 1645
  • Timescan at gain 14, file 1646
  • Channel Timescan gain 14, files 1647,1648
  • 1k turns Raw data, gain 14, file 1650
• Single bunch of e-, 4.0 mA
  • Timescan at gain 8, file 1651
  • Channel Timescan gain 8, files 1652,1653
  • 1k turns Raw data, gain 8, file 1655
• Single bunch of e+, 0.5 mA
  • Timescan at gain 14, file 1658
  • Channel Timescan gain 14, files 1659
  • 1k turns Raw data, gain 14, file 1660
• Single bunch of e+, 4.0 mA
  • Timescan at gain 8, file 1661
  • Channel Timescan gain 8, files 1662
  • 1k turns Raw data, gain 8, file 1663

	electron RMS on positions		positron RMS on positions
file	RMS_H (mm)	RMS_V (mm)	RMS_H (mm)	RMS_V (mm)	Comment
1650	0.1135	0.0749			0.5 mA, gain 14
1655	0.0918	0.0382			4.0 mA, gain 8
1660			0.1082	0.0464	0.5 mA, gain 14
1663			0.0668	0.0359	4.0 mA, gain 8

November 13, 2006 (E.Tanke): RMS of positons under CHESS conditions: dependance on pedestal

• 1k turns of raw data take for T1B1 electrons under 6x5 CHESS conditions. Measurements without pedestal subtraction and with pedestal subtraction, where different pedestal "bunches" have been selected. Results can be found here
• Conclusion is that the RMS on position is larger with pedestal subtraction than without and that all of the selected pedestals yield approximately the same RMS.

November 30, 2006 (M.Billing, G.Codner, E.Tanke): MS: 4 ns spaced bunches

• Find optimum timings for e- and e+ at CLEO energy, gain 8:
  • TglobA=!TglobB=1080
  • Tch0=680
  • Tch1=690
  • Tch2=690
  • Tch3=700
  • Tch4=700
  • Tch5=715
  • Tch6=705
  • Tch7=720
• Assume a train of bunches with 4 bunch spacing B1,B2,B3.....Bn
  • With the even channels having TglobA=1080, DSPe bunch T1B1 will correspond to B1, DSPe bunch T1B3 will correspond to B8 etc.
  • DSPe T1B2, T1B4 etc will not see any beam
  • The odd channels can be set to TglobB=650; then DSPo bunch T1B1 will correspond to B2, DSPo bunch T1B3 will correspond to B9 etc.
  • DSPo T1B2, T1B4 etc will not see any beam
• Timescan (file #1740) for 2 bunches shows they are ~4 ns apart
• For proof of principle, raw data over 10 turns was taken for these 2 bunches (file # 1738)