Single Point Hall Probe Studies in SHMS for Scaling SHMS Golden Tune

Status of Single Point Hall Probe Studies in SHMS Q2 Sam Murri, Dave Mack 26 July, 2017

Motivation We would like to be able to scale the SHMS Golden Tune we determine this Fall from 2.2 GeV/c to other momenta with linearity errors less than 2E-4. (See slide 2 of https://hallcweb.jlab.org/doc-private/ShowDocument?docid=876 .) The SHMS tune is most sensitive to the Q2 setting. (2E-4 corresponds to ~4 Gauss near 5 GeV/c, or only 0.8 Gauss near 1 GeV/c.) Assuming the set current error is <= 1E-4, in general the sources of nonlinearity for an SHMS magnet are A small, O(10) Gauss remnant B field in the iron as I 0, i. A few percent iron saturation as I Imax, ii. An O(1)% change in Effective Field Length as I Imax (gotten model-dependently from TOSCA) iii. iv. Hall probe errors at the 1E-3 level, for which we supposedly have correction tables. Ahem. 1

Taking Q2 Data We went through two bi-polar hysteresis loops with Q2: one loop with 24 data points, and a second one with 14 data points to check the reproducibility. There was a small chance of a quench at full power. So as not to perturb a frail ESR or critical 12 GeV solenoid work in Hall B, our Imax was limited to 70% of 11 GeV/c current. We hoped that magnet history would be erased after ramping to 0.7*Imax. Our set currents were on the downward path of the hysteresis loop: for this coil-dominated magnet where saturation was not expected to be an issue, we stepped downward in coarse steps of 10%, then 5% as we approached the remnant field. Hysteresis Data 30000 y = 9.4776x - 7.613 20000 10000 Field_CMBox [G] First Loop Second Loop Linear (First Loop) 0 -3000 -2000 -1000 0 1000 2000 3000 -10000 -20000 -30000 2 I_Readout [A] (With sign of polarity)

2.5

Q2 Good News Very short settling time: Less than 1 minute after voltage stops changing, the field readings are stable up to a bit-flicker. Reproducibility is excellent: Field measurements at the same current and polarity on different loops are consistent to 5 digits. A little spooky actually. (Since the tempco is 10-4/degC and our measurements took hours, the temperature must be very stable inside the bore of Q2.) Direct cross-talk between Q2 Hall probe and Q3 magnet is negligible. (Still need to check for something more subtle: cross-talk from Q3 magnet to the Q2 power supply when the latter is on. And cross-talk with Q1. ) 3

Uncorrected Q2 Field Residuals Wrt A Line Only includes the first loop and one point for I=0. Relative Field Residuals (First Loop Only) Field Residuals (First Loop Only) 0.006 20 0.005 10 0.004 0.003 0 -3000 -2000 -1000 0 1000 2000 3000 0.002 Relative Field [unitless] -10 Field [Gaus] 0.001 -20 0 -3000 -2000 -1000 0 1000 2000 3000 -30 -0.001 -40 -0.002 -0.003 -50 I_Readout [A] I_Readout [A] The take-away: Left plot - Absolute departures from perfect linearity are only ~20 Gauss Right plot - Relatively speaking, nonlinearities are mostly below 2E-3 except at lowest fields (below 1 GeV/c) The Hall probe corrections when done correctly - should be of this magnitude and might completely change this picture. 4

Other Good News Assuming approximate symmetry for the + and polarities, we can determine the iron remnant field and the Hall probe offset at I = 0 without a zeroing can or a degaussing cycle. ==> PS on PS off Zero Offset 1.55 G 2.25 G Remnant Field for +- Polarity, respectively (preliminary since we only cycled to 0.7Bmax) -+11.95 G -+10.05 G Surprisingly, for Q2 we care! Non-surprising results: The remnant field has a reasonable magnitude. We have also confirmed Dave Gaskell s old HMS observation that the results at Iset = 0 depend on whether the power supply is on/off or +/- polarity. Clearly there is an O(0.1)Amp trickle current at Iset = 0 when the power supply is on. It s big enough to matter. Surprising result: for + polarity, the remnant field is negative, and vice versa. The working hypothesis is that the thick iron yoke surrounding the coil-dominated Q2 acts like a magnetic mirror. 5

Q2 Less-Good News The asymmetry between + and polarity is relatively large (0.2%) even after nominal Hall probe corrections. The corrections for the Q2 Hall probe are too small to account for this, even assuming a sign error. Working hypothesis: the nominal Q3 Hall probe has been installed inside Q2, and vice versa. And the polarity is reversed. Pathetic, but possible. Need another day or two to track this down. Absolute B Field Difference Between the Two Different Polarities vs Current Relative B Field Difference Between the Two Different Polarities vs Current 10 0.001 0 (B_pos + B_neg)/B_pos [unitless] 0.0005 0 1000 2000 3000 B_pos + B_neg [Gauss] -10 0 0 1000 2000 3000 -20 -0.0005 -30 -0.001 -40 -0.0015 -50 -0.002 -60 -70 -0.0025 I_set [Amps] I_set [Amps] Corrected Not Corrected Corrected Not Corrected 6

Summary Our goal is to scale the SHMS Golden Tune we determine this Fall at 2.2 GeV/c to other momenta with linearity errors less than 2E-4. Lots of good news, but looks challenging with these Hall probes. What we have learned about Q2 so far: Settling Time (after voltage stabilizes) < 1 minute (transparent to users) Good to 10-5 (for same Iset and polarity) Reproducibility (of B at a given Iset over hours) Hall Probe Offset +1.55 Gauss (nice and small) Remnant Field (after cycling to 0.7Bmax) -+10.05 Gauss (Sign is opposite of nominal polarity, interesting but possibly understandable.) Polarity Asymmetry 2E-3. The only seriously bad news so far. The expectation of polarity symmetry allows us to check the validity of the Hall probe corrections provided by the manufacturer, and have confidence that we re measuring the true nonlinearity in Field vs Current to few x 1E-4 (rather than getting fooled by the Hall probe nonlinearity). Maybe we re applying Q2 Hall probe error corrections to the Q3 Hall probe. Work in progress. 7

Backup Q2 s Nominal Hall Probe Error vs Field Magnitude: these corrections are far too small to symmetrize our Q2 data 20.000 y = 3E-08x6 + 1E-06x5 - 3E-05x4 + 0.0006x3 + 0.0129x2 - 0.4041x + 0.5242 15.000 10.000 5.000 UUT Error [G] Central [-22.856 kG, 22.855 kG] 0.000 Low End [-29.931 kG, -23.739 kG] -40.000 -30.000 -20.000 -10.000 0.000 10.000 20.000 30.000 40.000 High End [23.738 kG, 29.930 kG] -5.000 Poly. (Central [-22.856 kG, 22.855 kG]) -10.000 -15.000 -20.000 Field [kG] 8