Hi! I hope you have had a good Thanksgiving holiday.-- Here, as a status report on the measurements taken on the OTF quadrupole, I send you a draft of a memo on it.-- The outstanding puzzle at the moment is why the remaining field seems to depend on the time structure of the cycling (that is, why does it work for 30 cycles in a row, but not, if each cycle is followed by a ramp to 0, and then ~40 seconds for a measurement, as indicated in the table (Run 31 - strdat.r31), where the remnant GL does not go below 10.4 Gauss. (I now try to repeat this measurement, and also change the ramp rate). Also let me know if you would want any other measurements (to assess robustness of procedure). Cheers, Achim ----------------------------------------------------------------------- OTF Degaussing Draft of 28-Nov-2005 Summary Below I report measurements on the green quadrupole labeled OTF. The aim of these measurements was to find a degaussing procedure, with a remnant field strength /Integrated gradient of 10 Gauss or less. The recommended procedure is to cycle the magnet 30 times, with the current reduced by a factor 0.8 from the preceding cycle, starting with 10 A (that is, ramp, in sequence, to 10 A, -10 A [cycle 1], 8A, -8A[cycle 2], 6.4 A, -6.4 A [cycle 3]........) and then ramping to zero current. We used a linear ramp, with 10A/sec; 30 cycles were accomplished in about 250 seconds. An important finding is that turning off the power supply to the magnet produces a current spike of 2.5 A; this spike re-magnetizes the magnet to produce a remnant field of >39 Gauss. Therefore it is important to let the power supply stay on after degaussing. General considerations The raw data for this study are at http://www-group.slac.stanford.edu/met/MagMeas/MAGDATA/LCLS/quad/OTF/ (in files with extension .ru# or r## where # or ## refers to a Run Number) I also measured the field strength as a function of current (Run 10: srtdat.r10); as shown in the first two columns of this table: Field with current Remaining Field at ~ 1mA after I [A] GL [T] current cycle GL [T] ------- ------- ----------------- ------- 10.154 0.41581 0.01651 -10.153 0.41058 +- 10 0.01394 5.095 0.20859 -5.093 0.20677 +- 5 0.00965 4.076 0.16545 -4. +- 4 0.00793 2.036 0.08013 -2. +- 2 0.00383 1.019 0.03720 +- 1 0.00360 0.458 0.01393 +- .5 0.00421 . 0.177 0.00324 +- .2 0.00303 . 0.088 0.00027 +- .1 0.00259 0.043 0.00119 +- .05 0.00267 0.016 0.00201 +- .02 0.00243 0.008 0.00226 +- .01 0.00242 0.005 0.00234 +- .005 0.002 0.00242 ------- ------- ----------------- ------- In the measurements of this table, the current directions were alternated between measurements, and measurements at zero (order mA) currents, were taken after each cycle. One can see that if the current is just turned off after a ramp to 10 A, the remaining field is about 165 G. After cycling of currents a few times, one can easily obtain fields of the order of 25 Gauss. Development of Degaussing Procedure The proposed degaussing procedure has the magnet ramped to positive and negative currents in sequence, with the current of each cycle reduced by a factor e,g, for a 'Start Current' of 10 A and a factor 0.8, the sequence would be +10 A, -10 A, +8A, -8A, 6.4 A, -6.4 A. ( 3 cycles shown). After going through enough of these cycles, the current is ramped down to 0. Indeed, when the magnet had a Gdl of 34 G, ramping it through 30 cycles starting at 3A, each subsequent cycle with a current reduced by a factor of 0.8, reduced the remnant field to 1.4 Gauss (strdat.r17 shows 2 before and two sets of 'after' measurements). Examples of such results are shown in this table: Procedure Number Field Field Start of before after Run Current cycles [A] # Gauss Gauss -- ------- ------ ----- ----- 15 4 10 14.9 16 4 30 2.1 17 3 30 33.9 1.4 18 3 20 35.6 1.9 19 3 20 43.2 1.7 20 3 20 38.4 1.6 -- ------- ---- ----- ---- To assess the number of cycles needed, I ramped the current up ( to +x A) then down (to -x A), and then to 0, and took a measurement then. The results are shown in the following table (Run 31 - strdat.r31): Cycle # GL [G] ---------- ------ -1(before) 50.6 1 125.2 2 111.4 3 96.3 4 84.3 5 71.9 ... ..... 10 21.02 11 18.6 12 21.4 13 19.9 14 18.0 15 16.3 .... .... 20 11.7 21 11.4 22 11.1 25 10.7 ..... .... 30 10.5 35 10.5 40 10.4 ---------- ------ This seems to indicate that if you stop after each cycle, you cannot get below ~10.5 Gauss. Note that here after each up and down and back to zero cycle, a measurement was taken, taking about 2 minutes 40 seconds (actually, 4 measurements, which are averaged, and current measurements). However, if one ramps up and down without interruption, and takes a measurement only at the end of so many cycles, the remnant field can be much improved, as shown in the following table (based on Runs 32-42): ---------------------------------------------------------------- Number of 10 15 20 20 25 28 30 30 32 32 35 40 cycles ------ --------------------------------------------------------- GL [G] 22.3 11.9 5.1 5.2 3.8 3.8 3.4 3.7 6.1 3.7 5.3 3.3 ---------------------------------------------------------------- Here, the current was ramped linearly at 10 A/sec, and 10 cycles of ramping up and down took about 85 sec. So one can conclude that by ramping up and down 30 times, a remnant field of 3.5 G can be achieved. Remaining current and current spikes A concern had been raised about the remaining current, at the mA level, so after ramping the current to 0, and measuring there, I also made two more measurements with the Power supply turned off. Surprisingly, the new measurement gave a field of about 37 Gauss (at a current measured at 0.03 mA; when PS was on, it was 1.5 mA). This effect could be reproduced (Runs 17, 18, strdat.r18). Also, ramping to +-2.5 A and back to zero gave similar remaining fields (GL): Procedure Number Field Field Start of before after Run Current cycles [A] # Gauss Gauss -- ------- ------ ----- ----- 21 2.5 1 45.2 38.7 22 3.125 2 45.2 39.4 -- ------- ---- ----- ---- Looking at the transducer output with an Oscilloscope, it was found, that turning off the power supply, a Kepco Bipolar Power Supply/Amplifier, created a current spike of 2.5 A, which apparently re-magnetized the magnet. (A figure of the current spike is at http://www-group.slac.stanford.edu/met/MagMeas/MAGDATA/LCLS/quad/OTF/Spike-fig.pdf ) After ramping the magnet to +-3.125, +-2.5A, the remaining field is also 39 Gauss, and increases to 59 Gauss, after the PS is turned off. (Run 21, 22 in Table above) To address the problem that the remaining current affects the magnetic strength measurement, I tried to measure at several small currents (Run 20,44). The DAC does not have enough resolution to properly set +-1 mA currents, so the setting a current is hit-and miss, but the current readout is better. I [mA] | -7 -1 2 10 ------------------------------------ GL [G] | 4.1 2.4 1.6 0.9 I [mA] : -16. -9.9 -6.9 -3.9 -1.1 -1 -.9 1.8 4.8 7.8 13.7 ------------------------------------------------------------------ GL [G] : 7.9 6.2 5.4 4.5 3.7 3.7 3.7 2.9 2.1 1.2 0.45 From these limited measurements, there seems to be a reduction of the field for small positive currents. However the change in GL from, say -1.1 to +1.8 mA is less than 1 Gauss. So, at that scale, we can ignore remaining currents of a mA. --------------------------------------------------------------------