E. Todesco

UPDATE ON THE MAGNETIC MODEL,PRECYCLES AND BETA-BEATING

P. Hagen, E. Todesco for the FiDeL teamMagnets, Superconductors and Cryostats Group

Technology Department, CERN

FiDeL meeting, CERN 16th February 2010

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 2

CONTENTS

Missing precycles in 2009

The issue of the minimum precycle current

The issue of magnets with negative current nominal settings

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 3

WHAT WAS MISSING IN 2009

The following magnets have not been precycled in 2009

The block of 6 MQTL around IR3 and IR7 in Q6

Spectrometer compensators (6 magnets) had no precycleWe had an issue in the bump correction of the spectrometers – it is not likely that this stems from the missing precycle, since they were operating at full current

Both families will be precycled in 2010

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 4

MINIMUM PRECYCLE CURRENT: foreword

In order to get on the ascending branch, the minimum current achieved during the precycle must be (much) smaller than the injection current Iinj>Imin

pre

A correct precycle with Iinj>Iminpre

We inject the beam hereWe have collisions here, then we dump the beam

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 5

MINIMUM PRECYCLE CURRENT: FOREWORD

If on the other hand Iinj<Iminpre we can make very

large errors since the model assumes we are on the ascending branch, whereas we are on the descending one

At low currents this error can be large (of the order of a few per mil)During the ramp the error disappears since one recovers the right branch

A wrong precycle with Iinj<Iminpre

But the magnet has this field

We inject the beam here

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 6

MINIMUM PRECYCLE CURRENT: MQWA

This wrong precycling happened for a few magnetsFirst case: MQWA – the precycle goes downto 35.1 A and we inject at 35.3/38.3 AAt 35 A the hysteresis gives a +4% errorbetween the model and the actual magnet

-300

-200

-100

0

100

200

300

400

500

600

0 100 200 300 400 500 600 700

TF

(uni

ts w

.r.t.

200

A v

alue

)

Current (A)

+4 %

Measurements of MQWA hysteresis

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 7

MINIMUM PRECYCLE CURRENT: MQWA

This fits with what guessed by Rogelio on the ground of beta beating measurements

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 8

MINIMUM PRECYCLE CURRENT: MQXA

This wrong precycling happened for a few magnetsSecond case: MQXA – the precycle goes downto 400 A and we inject at 408 A in IR1 and IR5But at 452 A in IR2 and IR8 – should be betterAt 400 A the hysteresis gives a -0.4% errorbetween the model and the actual magnet(30 units) in IR1 and IR5

Measurement of MQXA hysteresis

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 9

MINIMUM PRECYCLE CURRENT: MQXB

This wrong precycling happened for a few magnetsThird case: MQXB – the precycle goes downto 700 A and we inject at 708 A in IR1 and IR5But at 776 A in IR2 and IR8 – should be betterAt 700 A the hysteresis gives a -0.6% errorbetween the model and the actual magnet(60 units) in IR1 and IR5

-200

-150

-100

-50

0

50

100

0 5000 10000

TF

(Uni

ts)

Current (A)

MQXB - integral

MQXB - DCloop

Satur+magn FiDeL

Roxie model

MQXB1

Measurement of MQXB hysteresis

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 10

MINIMUM PRECYCLE CURRENT: MBW

This wrong precycling happened for a few magnetsFourth case: MBW – the precycle goes down to 41.1 and we inject at 40.9A

Fifth case: MBRC – injection 80 A above in IP1 and IP5

Probably the effect is not strong (less than 10 units ?) – to be checked

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 11

CONTENTS

Missing precycles in 2009

The issue of the minimum precycle current

The issue of magnets with negative current nominal settings

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 12

MAGNETS WITH NEGATIVE CURRENT NOMINAL SETTINGS

In 2009, the precycles for bipolar magnets have been implemented always as up-down to be on the ascending branch

Correctors will have either negative of positive current, according to the steering algorithm (orbit, tune, coupling)Some magnets with dipolar power converter belong to the optics and have negative current settings:

63 out of 128 MQTL in positions Q8 to Q134 out of 8 MQWB in positions Q4 and Q5

These magnets have been on the wrong hysteresis branch in 2009

Can this be a source of beta beating ?

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 13

MAGNETS WITH NEGATIVE CURRENT NOMINAL SETTINGS - MQTL

The MQTL have a very low residual magnetizationIt is not modelled in FiDeLTherefore this issue is not relevant, I would not take any action

-72

-54

-36

-18

0

18

36

54

5.1E-03

5.2E-03

5.3E-03

5.4E-03

5.5E-03

5.6E-03

5.7E-03

5.8E-03

-600 -400 -200 0 200 400 600

TF

(uni

ts)

TF

[T m

/ A

]

Current [A]

Measurement of MQTL hysteresis

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 14

MAGNETS WITH NEGATIVE CURRENT NOMINAL SETTINGS - MQWB

The MQWB have a non negligible residual magnetization

What is the effect of a wrong precycling ?Should one correct it?

Measurement of MQTL transfer function close to zero

-25%0%

25%50%75%

100%125%150%175%200%

0 10 20 30 40 50

Tra

nsfe

r fu

ncti

on v

aria

tion

I (A)

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 15

Summary of BETA BEATING SOURCES IN THE FIELD MODEL IN 2009

MQWA: magnets were on the wrong branch of the hysteresis due to a precycle minimum current larger than injection current (+4% error)MQXA in IP1 and IP5: magnets were on the wrong branch of the hysteresis due to a precycle minimum current larger than injection current (-0.4% error)MQXB in IP1 and IP5: magnets were on the wrong branch of the hysteresis due to a precycle minimum current larger than injection current (-0.6% error)

The block of 6 MQTL around IR3 and IR7 in Q6 has not been cycled – effect should be negligible

MQWB: half of them are on the wrong branch of the hysteresis since they have negative settings (error to be estimated - is it relevant?)MQTL: half of them are on the wrong branch of the hysteresis since they have negative settings (should be negligible)

E. Todesco 16th February 2010 – Magnetic model, precycles, beta beating - 16

TUNE DRIFT DURING RAMP

Tune drift during rampThe error in the transfer function of the MQWA (+4%), and in IR1 and IR5 MQXA (-0.4%) and MQXB (-0.6%) is present only at injectionDisappears during the ramp, where the right branch is recovered

What is the impact on the tune?Is it related to the observed tune drift?

With these changes the machine could be considerably different from 2009

Both in terms of tune trims and beta beating