Paul Scherrer Institute Pierre-André Duperrex. On-line calibration schemes for RF-based beam diagnostics

Size: px

Start display at page:

Download "Paul Scherrer Institute Pierre-André Duperrex. On-line calibration schemes for RF-based beam diagnostics"

Stephen Glenn
5 years ago
Views:

1 Paul Scherrer Institute Pierre-André Duperrex On-line calibration schemes for RF-based beam diagnostics HB2012 Beijing, Sept. 2012

Motivation Current monitor Some difficulties related to RF signal measurements: BPM Sensors 1 to N

Difference in the overall gain between measurement chains introduces error.

Calibration may require some large effort, be time-consuming and possibly be required after repairs.

2 Motivation Current monitor Some difficulties related to RF signal measurements: BPM Sensors 1 to N might have slightly different sensitivity. Difference in the overall gain between measurement chains introduces error. Temperature drifts may affect differently the electronic elements. Calibration may require some large effort, be time-consuming and possibly be required after repairs. Also repeated calibrations may be needed to confirm the gain. On-line calibration schemes may remove some of these difficulties. HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 2

3 First application: on-line calibration scheme for beam current monitors HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 3

4 HIPA Beam Current Monitors Measurement principles: coaxial resonator magnetic field directly proportional to the beam current tuned at MHz, the RF 2 nd harmonic dimension: outer diameter: 40cm, length: 20cm Advantages: simple design radiation resistant Disadvantages: no absolute measurements temperature drifts Challenge: A beam current monitor (called MHC5) is in vacuum 8 m behind a 4cm graphite target. This monitor is subject to a heat load due to scattered particles ( 250W for a 2mA beam). The resonator gain drift due to the heat load makes the beam current measurement challenging. HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 4

5 Resonance Condition coaxial line with a capacitor shunt: Coaxial line loss-less impedance with a shorted load: an inductive effect for: 2π fml π L 1 < ie.. < c 2 λ 4 m Z i 2π fml = jzotan c With the capacitor reactance: X c = 1 j2π f C m capacitor shunt The parallel circuit is resonant when: Z i = X c Resulting relation for resonance with a capacitor shunt: coaxial line L : resonator length C : capacitor shunt Z o : characteristic impedance λ : resonant wavelength m 2π L tan = λm λ m 2π ccz o In our case: L ~ 15 cm HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 5

6 Universal Tuning Curve The universal tuning curve as relation between the resonator length and the required C value. 2π L λ m 1 λ m = tan 2π cczo The red line corresponds to the MHC5 conditions. Temperature drift will affect the length of the resonator, the shunt capacitance and the line impedance. HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 6

7 Temperature Effects Optimization on the test bench to minimize the temperature drifts 30 deg. C. 1dB 70 deg. C. Measurement of the transfer function around MHz (2 nd harmonic) Temperature coefficient: <0.01dB/ C (30 C..70 C) 0.3dB 3.5% in amplitude MHz 500kHz But the drifts observed during operation are even larger, due to the non uniform temperature distribution. HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 7

8 MHC5 measured signal measured: S = K. S (2 f ). I m 21 RF beam K : constant S (2 f ) : resonator gain at the RF 2nd harmonic I 21 beam RF : beamcurrent beam signal another S ( f) 21 resonator transfer function S ( f) MHz frequency Change of the measured signal reflects either a change of beam current or a change of the resonator gain or both HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 8

9 Calibration Scheme Idea S. (2 ). m = K S21 frf Ibeam How to distinguish or to separate the two contributions? Basic idea: measure some pilot signals at frequencies very close to the beam signal frequency (101.26MHz) to monitor the gain drift. Pilot signals beam signal resonator transfer function S ( f) MHz frequency HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 9

10 Calibration Scheme Idea S21 P1 B B 0.5*(P1+P2) P2 Pilot signals resonator transfer function S ( f) MHz frequency HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 10

11 Calibration Scheme Idea S21 B P2 B 0.5*(P1+P2) Pilot signals P1 resonator transfer function S ( f) MHz frequency HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 11

12 Calibration Concept Implementation concept Transmitter MHC5 Receiver estimate of the resonator gain: gain resonator S α S Pilot Reference HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 12

13 I/Q Demodulation & Image Rejection gain MHC5 P α S + P 1 2 Reference S = P.cos( ω. t+ φ ) + P.cos( ω. t+ φ ) + S.cos( ω. t) m beam SQ () t = P10.sin ( Δω. t φ1 ) P20.sin ( Δ ω. t+ φ2 ) SI () t = P10.cos ( Δω. t φ1 ) + P20.cos ( Δ ω. t+ φ2 ) SQ+ 90deg () t = P10.cos Δω. t φ1 P20.cos Δ ω. t+ φ SQ 90deg () t = P10.cos Δω. t φ1 + P20.cos Δ ω. t+ φ2 2 2 ( ) ( ) ( ) ( ) Δ ω = ω1,2 ω0 ( ω φ ) S () t + S () t = P.cos Δ. t I ( ω φ ) S () t + S () t = P.cos Δ. t+ I Q 90deg 20 2 Q+ 90deg 10 1 HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 13

14 Calibration: Off-line comparison MHC5 cooling switched on 10% variation pilot drift compensation MHC5 cooling switched off MHC6: another monitor on the same beam line, expected: MHC5 MHC 6 I Possibility to use MHC6 for the MHC5 calibration I The pilot drift compensation matches the calibration deduced from MHC6 HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 14

15 Results during beam operation (2010) Drift compensation Temp MHC5 MHC5 MHC6 HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 15

16 Initial Implementation Transmitter MHC5 Receiver Phase drifts! HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 16

17 Improved Implementation Transmitter MHC5 Receiver Upgrade: I/Q demodulation & image rejection for the reference too HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 17

18 2nd application: on-line calibration scheme for beam position monitors HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 18

26MHz) signals Position calculation based on the difference between the signal level of probes located on

19 HIPA Beam Position Monitors BPM probe magnetic pickup size: 4 x 8 cm x/y systems integrated with x/y profile monitors in a single box measurement of the RF 2nd harmonic (101.26MHz) signals Position calculation based on the difference between the signal level of probes located on opposite side to the beam S Δx α S S + beam beam + beam + Sbeam HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 19

20 standard Implementation Overview based on digital receiver technology direct frequency down-converting of the RF 2nd harmonic (101.26MHz) signals (no analogue LO) max. S/N full ADC range AGC Issues related to absolute RF measurements requires tuning temperature drift may be a problem (electronics, cable attenuation) HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 20

21 Implementation with Pilot Signals based on digital receiver technology direct frequency down-converting of the RF 2nd harmonic (101.26MHz) signals (no analogue LO) online measurement of individual channel overall gain using MHz pilot signals max. S/N full ADC range AGC Using a pilot signal the whole measurement chain is calibrated. HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 21

22 dbpm electronics HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 22

23 Position Calculation including the pilot signals: S Δx α S S + n n + n + Sn + ( ) S with Sn = normalized signal of sensor + or S S : beam component of sensor + or + ( ) beam S : pilot signal of sensor + or + ( ) pilot + ( ) beam + ( ) pilot Some concerns about the degradation of the signal-to-noise ratio (SNR) HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 23

Signal-to-Noise Ratio Signals analyzed: raw, pilot and normalized signals from the right pickup of MXS3 S + σ 4% beam S + pilot σ 4% S + n = S S + beam + pilot σ 1% The relative standard

24 Signal-to-Noise Ratio Signals analyzed: raw, pilot and normalized signals from the right pickup of MXS3 S + σ 4% beam S + pilot σ 4% S + n = S S + beam + pilot σ 1% The relative standard deviation decreases by a factor 4 for the normalized signal! SNR has improved for the normalized signal! HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 24

25 Correlation between Pilot and Beam Signals Coherence Spectra for different beam, pilot and normalized signal combinations C xy ( f) = xy ( f) 2 P ( f) P ( f) xx P yy The Welch method using a Hanning window and 50% overlap has been applied. Measurement conditions 1.97mA beam current BPM MXS3 & MXS4 a) b) c) HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 25

26 Correlation between Pilot and Beam Signals Coherence Spectra for different beam, pilot and normalized signal combinations C xy ( f) = xy ( f) 2 P ( f) P ( f) xx P yy The Welch method using a Hanning window and 50% overlap has been applied. Measurement conditions 1.97mA beam current BPM MXS3 & MXS4 d) e) f) HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 26

27 Correlation between Pilot and Beam Signals Coherence Spectra for different beam, pilot and normalized signal combinations a) b) d) e) c) f) Observations horizontal and vertical signals are not correlated (a,b) beam signal and pilot signal noise highly correlated in the horizontal or vertical direction (c,e,f) mainly dominated by instrumental broadband noise normalized signals (d): only the 50Hz harmonics are correlated, real origin possible Because pilot and beam signals are well correlated, the normalization improves the SNR! HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 27

28 Pilot Scheme as (adaptive) Interference Cancelling The dbpm pilot scheme can be seen as interference canceling system. This idea could be possibly further develop so to have an adaptive interference cancelling system. Reminder: an IEEE 1975 paper from B. Widrow and al. the concept an application HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 28

29 Summary On-line calibration schemes: offer some clear accuracy improvements for sensors such as resonators affected by gain drifts due to temperature effects. do not require extensive calibration procedure may improve the signal-to-noise ratio (SNR) in some cases Drawback: more elaborate instrumentation Outlook: with the latest FPGA und electronic technology, more elaborate schemes may be implemented such as adaptive filtering or interference cancelling HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 29

30 Thanks for your attention PSI, HB2012, Beijing, 20 September 2012 On-line calibration schemes for RF-based beam diagnostics Page 30

Chapter 12: Transmission Lines. EET-223: RF Communication Circuits Walter Lara

Chapter 12: Transmission Lines EET-223: RF Communication Circuits Walter Lara Introduction A transmission line can be defined as the conductive connections between system elements that carry signal power.