Digital Low-level RF Test at Diamond Booster Digital Low-level RF Test at Diamond Booster LLRF9...

Digital Low-level RF Test at Diamond BoosterLLRF9 demo, March 16–20, 2015

D. Teytelman2, et. al.

2Dimtel, Inc., San Jose, CA, USA

April 15, 2015

(Dimtel) iGp12 DLS 1 / 16

Setup LLRF9 Introduction

Outline

1 SetupLLRF9 IntroductionDemo Setup and Schedule

2 System Operation

3 Performance Tests

LLRF9 System

A single 2U chassis for one-and two-cavity RF control;9 input RF channels, 2 RFoutputs;Tuner motor control viaRS-485/Ethernet/EPICS;External interlock daisy-chain;Two external trigger inputs;Eight opto-isolated basebandADC channels for slowinterlocks.

LLRF9 System

Setup Demo Setup and Schedule

Outline

1 SetupLLRF9 IntroductionDemo Setup and Schedule

2 System Operation

3 Performance Tests

Demo Setup

Set up LLRF9 to run the booster RF with the following signals:RF reference (500 MHz)Three cavity probe signals (500 MHz)Cavity forward signal (500 MHz)Cavity reflected signal (500 MHz)Drive output (500 MHz)Interlock input (TTL)Ramp trigger (TTL)

Used EPICS interface to control cavity tuning;LLRF9 drive was connected to RF Safety Crate (113S) to maintaininterlocks;Bypassed existing LLRF chain, RF Safety Crate output connecteddirectly to preamplifier (222S).

Progress

MondayConnected LLRF9 to inputs only, established signal levels andtransferred calibrations;Connected drive output, configured feedback loops;Established closed-loop operation in CW mode.

TuesdayInterfaced LLRF9 tuner control loops to booster motor control;Established closed-loop operation of tuner and field balance loops.

WednesdayTransitioned to closed-loop operation with ramping and beam;Characterization of open and closed-loop spectra.

ThursdayStep response measurements;System set up from scratch by Alun Watkins.

FridayPerformance testing;Discussion.

System Operation

Open Loop Transfer Function

∑ ErrorFeedback

RFcavity

DisturbancesSetpoint and excitation

Cavity field probe

Measured from setpoint to thecavity probe;Feedback block in open loop hasno dynamics, just gain and phaseshift;Open loop cavity response;Fit resonator model to extractgain, loaded Q, detuning, delay,phase shift at ωrf;Faster than expected gain roll-offabove the resonance.

System Operation

∑ ErrorFeedback

RFcavity

Cavity field probe

System Operation

−250 −200 −150 −100 −50 0 50 100 150 200 250−40

Frequency offset (kHz)

−250 −200 −150 −100 −50 0 50 100 150 200 250−200

−100

Phase (

System Operation

−250 −200 −150 −100 −50 0 50 100 150 200 250−40

Gain = 0.2, Q = 15739.5, (wr − w

rf) = 0.55 kHz

−250 −200 −150 −100 −50 0 50 100 150 200 250−200

−100

Phase (

τ = 750.834 ns, φ = 359.6 deg

System Operation

−250 −200 −150 −100 −50 0 50 100 150 200 250−40

Gain = 0.2, Q = 15739.5, (wr − w

rf) = 0.55 kHz

−250 −200 −150 −100 −50 0 50 100 150 200 250−200

−100

Phase (

τ = 750.834 ns, φ = 359.6 deg

System Operation

Wideband Open Loop Transfer Function

−4000 −3000 −2000 −1000 0 1000 2000 3000 4000−80

−4000 −3000 −2000 −1000 0 1000 2000 3000 4000−600

−400

−200

Phase (

Wider sweep reveals a parasiticmode at 2.8 MHz above the πmode;Negative feedback for the π modeis positive for the parasitic mode;This positive feedback limits directloop gain;The simplest way around theissue is to use digital delay toequalize the modal phase shifts(230 ns).

System Operation

−400 −200 0 200 400−50

Gain −13.7 dB

−400 −200 0 200 400−200

−100

Phase (

Phase −0.4 deg

2600 2800 3000 3200−35

Gain −12.6 dB

2600 2800 3000 3200−300

−200

−100

Phase (

Phase −121.0 deg

System Operation

−400 −200 0 200 400−50

Gain −13.7 dB

−400 −200 0 200 400−200

−100

Phase (

Phase −0.4 deg

2600 2800 3000 3200−35

Gain −12.6 dB

2600 2800 3000 3200−300

−200

−100

Phase (

Phase −121.0 deg

System Operation

−400 −200 0 200 400−50

Gain −13.7 dB

−400 −200 0 200 400−200

−100

Phase (

Phase −0.4 deg

2600 2800 3000 3200−35

Gain −12.6 dB

2600 2800 3000 3200−200

−100

Phase (

Phase −2.5 deg

System Operation

Proportional Loop Gain and Delay

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−80

Gain margin 12.3 dB at 2874.4 kHz

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−200

−150

−100

Set up minimum delay and equalizedtransfer functions for identical 3 dBclosed-loop peaking.

Minimum delay: peak gain at RF is−9.2 dB, gain margin 12.3 dBEqualized: peak gain at RF is+8 dB, gain margin 11.8 dB, phasemargin 88 degrees

More sophisticated parasitic modesuppression methods can improvethe performance only slightly, around2-3 dB.

System Operation

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−60

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−200

−150

−100

200Phase margin 88.5

° at 2904.3 kHz

System Operation

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−60

−1000 −500 0 500 1000 1500 2000 2500 3000 3500 4000−200

−150

−100

200Phase margin 88.5

° at 2904.3 kHz

System Operation

Closed Loop Transfer Function

∑ ErrorFeedback

RFcavity

Cavity field probe

Measured from setpoint to theerror signal;Quantifies closed-loopdisturbance rejection vs.frequency offset from fRF;Proportional and integrator loopsproduce high rejection at lowfrequencies;Magnitude on log-log scale, fieldsetpoint of 1 MV.

System Operation

∑ ErrorFeedback

RFcavity

Cavity field probe

System Operation

Frequency offset (Hz)

Direct

Direct+integral

System Operation

Frequency offset (Hz)

Direct

Direct+integral

System Operation

Step Response

0 50 100 150 200 250190

Time (µs)

Cavity probe

0 50 100 150 200 25054

) Use ramp start signal to triggerwaveform acquisition;Ramp profile loaded with a 10%amplitude step (200 to 220 kV);Some coupling between amplitude andphase;All input channels are captured on thesame trigger, observing reflectedpower here;Characterized a 5 degree phase stepas well.

System Operation

Step Response

0 50 100 150 200 250190

Time (µs)

Cavity probe

0 50 100 150 200 25054

System Operation

Step Response

0 50 100 150 200 250190

Time (µs)

Cavity probe

0 50 100 150 200 25054

System Operation

Step Response

0 10 20 30 40 50190

Time (µs)

Cavity probe

0 10 20 30 40 500

Time (µs)

Reflected power

0 10 20 30 40 5054

0 10 20 30 40 50−100

Use ramp start signal to triggerwaveform acquisition;Ramp profile loaded with a 10%amplitude step (200 to 220 kV);Some coupling between amplitude andphase;All input channels are captured on thesame trigger, observing reflectedpower here;Characterized a 5 degree phase stepas well.

System Operation

Step Response

0 50 100 150 200 250198

Time (µs)

0 50 100 150 200 25050

Performance Tests

IQ Circles

Frequency offset input, IQ circles, decibel radius scale

Input signal at −60 to 0 dBm in 10 dBsteps, ≈ 1 Hz offset;Log scale polar plot (r = 20 log10|x|);Center offset, more prominent at lowamplitudes, is due to RF feedthrough(coupled from the reference channel);Can numerically remove the offsets;Good linearity (uncorrected forsynthesizer errors), consistent offsetsat −73 dBFS.

Performance Tests

IQ Circles

Performance Tests

IQ Circles

Performance Tests

IQ Circles

Performance Tests

IQ Circles

−60 −50 −40 −30 −20 −10 0−80

Input power (dBm)

IQ amplitude vs. input power

−60 −50 −40 −30 −20 −10 0−0.2

Input power (dBm)

Error between Pin

and IQ amplitude, linear fit from −50 to −10

−60 −50 −40 −30 −20 −10 0

Input power (dBm)

RF reference feedthrough (circle offset)

Performance Tests

IQ Circles: On the Bench

Similar to the test at DLS with severalmodifications:

Reference level at +10.5 dBm vs.+8 dBm at DLS;Input power level monitored by aspectrum analyzer;5 dB steps.

Center offset correction;Linearity good to ±0.05 dB down to-50 dBFS;Offsets/feedthrough at −72 dBFS.

Performance Tests

−70 −60 −50 −40 −30 −20 −10 0−80

Input power (dBm)

−70 −60 −50 −40 −30 −20 −10 0−0.5

Input power (dBm)

Error between Pin

and IQ amplitude, linear fit

−70 −60 −50 −40 −30 −20 −10 00

Input power (dBm)

Performance Tests

−70 −60 −50 −40 −30 −20 −10 0−80

Input power (dBm)

−70 −60 −50 −40 −30 −20 −10 0−0.5

Input power (dBm)

Error between Pin

and IQ amplitude, linear fit

−70 −60 −50 −40 −30 −20 −10 00

Input power (dBm)

Performance Tests

Single Channel Phase Noise

−146

−144

−142

−140

−138

−136

−134

−132

Frequency (Hz)

Reference channel at DLS,+8 dBm input level;Sources of lines at 65.2and 111.8 kHz are known,already fixed;Same channel measuredon the bench, HP8664A at+11.5 dBm.

Performance Tests

−146

−144

−142

−140

−138

−136

−134

−132

Frequency (Hz)

Performance Tests

−146

−144

−142

−140

−138

−136

−134

−132

X: 5.592e+04Y: −134

Frequency (Hz)

Performance Tests

Thermal Stability

−12 −10 −8 −6 −4 −2 0

Time (hours)

Ambient temperature measurement

On chassis

In air

−12 −10 −8 −6 −4 −2 0

Time (hours)

Out of loop sensors

−12 −10 −8 −6 −4 −2 020.98

Time (hours)

In−loop sensors

LLRF:BRD1:NTC1:TEMP [degC]

LLRF:BRD1:DS1822:TEMP [degC]

9 internal sensors on cold plate: 6NTCs, 3 DS18B20 digital sensors;Three temperature stabilization loopsusing thermoelectric coolers;Two external sensors, in air andattached to chassis;Tight stabilization of in-loop sensors;Residual sensitivity of out-of-loopsensors is 0.09–0.12 ◦C/◦C.

Performance Tests

Thermal Stability

−12 −10 −8 −6 −4 −2 0

Time (hours)

On chassis

In air

−12 −10 −8 −6 −4 −2 0

Time (hours)

Out of loop sensors

−12 −10 −8 −6 −4 −2 020.98

Time (hours)

In−loop sensors

Performance Tests

Thermal Stability

−12 −10 −8 −6 −4 −2 0

Time (hours)

On chassis

In air

−12 −10 −8 −6 −4 −2 0

Time (hours)

Out of loop sensors

−12 −10 −8 −6 −4 −2 020.98

Time (hours)

In−loop sensors

Performance Tests

Thermal Stability

−12 −10 −8 −6 −4 −2 0

Time (hours)

On chassis

In air

−12 −10 −8 −6 −4 −2 0

Time (hours)

Out of loop sensors

−12 −10 −8 −6 −4 −2 020.98

Time (hours)

In−loop sensors (32 sample average)

Performance Tests

Thermal Stability

−12 −10 −8 −6 −4 −2 0

Time (hours)

On chassis

In air

−12 −10 −8 −6 −4 −2 0

Time (hours)

Out of loop sensors

−12 −10 −8 −6 −4 −2 020.98

Time (hours)

In−loop sensors (32 sample average)

Summary

Successfully operated booster RF station with beam;Demonstrated powerful user interface tools necessary for faststation configuration;Demonstrated field control at forward power 60 W to 60 kW;Showed high rejection of disturbances, good tuner loop responsewith low reflected power.

Summary

Digital Low-level RF Test at Diamond Booster Digital Low-level RF Test at Diamond Booster LLRF9...

Documents

Transcript of Digital Low-level RF Test at Diamond Booster Digital Low-level RF Test at Diamond Booster LLRF9...