Minutes of the EPU task force at the ALS

Taken from an e-mail I sent on 1/24/2001:

Hello everybody,

attached you will find the meeting minutes of the first meeting of the EPU (sector 4) taskforce plus a list of action items (with names and rough time schedules).

1/23/2001, 10:00-12:00, 80-234

Introduction, charge and welcome (Christoph Steier, Ben Feinberg)

A change of the longitudinal shift parameter of the EPU (to select different polarizations of the photons) creates a big change in the vertical beamsize of the ALS. The effect is larger than it used to be, has a very different qualitative dependence on the shift parameter than before and is a significant problem for some other beamlines. The charge for this taskforce is to understand the effect and come up with proposals for remedies.

In addition there is a problem with the motor of the gap servo system (vertical gap drive). The amplifier is tripping more and more often in recent weeks (last week about 10 times per day). This is a significant problem for beamline 4.0 (and sometimes for the operators if they want to refill the ring). The problem might be related to the first one or might not be related.

Presentation of results of accelerator physics studies concerning the EPU influence on the vertical beam size conducted so far

(Christoph Steier, with several plots from Laurent Nadolski who is in LA on a USPAS at the moment)
  1. vertical beamsize in general can depend on tunes (tune difference of hor. and vert.), strength of coupling resonance, and vertical dispersion
  2. effect of EPU on vertical beamsize is much larger than it used to be. We do not exactly know (so far) when it became worse, we noted it after the Superbend commissioning.
  3. dependence on shift parameter is sine like (asymmetric for positive and negative shift parameters), it used to be symmetric!
  4. tune shift is as expected and as it was in the past
  5. vertical dispersion does not change
  6. strength of coupling resonance does change (have not observed this in the past)

Conclusions so far

  1. this is not an effect of the finite dispersion lattice
  2. this is not an effect of orbit distortions in the arcs (in sextupoles or Superbends)
  3. this is not an effect of misalignment of the EPU
  4. The dependence of this effect on specific skew quadrupole settings is small.
  5. The dependence of the effect on deliberate offsets of the two moveable yaws of the EPU is negligible.

We did several other studies but I will not enumerate all of them here ...

One peculiar observation is that after maintenance periods in November and December, when the drive mechanism was lubricated, the effect initially became much smaller (in one case as small as it used to be several years ago). On the contrary, after the christmas shutdown where the device was off for weeks but was not lubricated, the effect was as big and bad as before the shutdown. (Note: This effect later turned out to be coincidental - note added: 2002-11-04)

Discussion:

  1. the aluminum holders are known to bend under the strong magnetic forces when moving the EPU longitudinally (should be symmetric for + and - direction), maybe check whether something has changed there
  2. understand influence of lubrication with a larger statistic than so far
  3. airflow to get hot air away from motors seems to be very small: investigate!
  4. there was an energy shift about a year ago which was consistent with the gap slipping by about 150 microns (field got weaker), mechanical measurements at that time showed that something changed ... maybe asymetrically (taper ?): understand this
  5. vertical gap+tapering was not studied as a function of shift parameter

Action Items:

  1. lubricate device (today or next AP shift or next maintenance) Steve Marks; Dennis Gibson
  2. measure horizontal deflection as function of shift parameter (next shutdown) Steve Marks
  3. measure motor, housing, undulator temperature with thermocouples (coordinate time with lubrication) Mike Chin
  4. measure taper + vertical gap (also as function of shift parameter), using micrometers or potentially laser interferometer (within a month) Steve Marks to check who could do it (Dave Anderson did it in the past) and how (maybe with help from Walter Barry if laser interferometer is to be used)
  5. field model B_{x,y} (x,y,s) - tracking - maybe error analysis (starting next week, first results within a few weeks) Laurent Nadolski, Ross Schlueter
  6. look at archive data to find when exactly it got worse and whether it was a sudden event (within a week) Christoph Steier
  7. quantify energy shift and peak width as a function of polarization (left/right), at a later time potentially measure polarization (or asymmetry) - use polarimeter or sample with large (known) dichroism. (first results based on old data within a week ...) Tony Young, Elke Arenholz

Problem of motor trips (presentation of Ken Fowler):

Explanation of what has been observed and how the servo motor feedback system works. Short description of remedies tried so far. Amplifier always trips when device is at rest (so far only observed for gaps smaller than 30 mm, not with gap open). The trip is an internal amplifier protection circuit, which reacts to integral(I**2 dt).

Discussion...

Action items:

  1. monitor the DAC ouput (amplifier input)
  2. measure motor temperature
  3. perform further tests with the integrator part of the feedback loop (e.g. try to disable it completely)

Next meeting will be in 2-3 weeks (probably just after the next shutdown, which will be February 11-12). That should give enough time to have some progress on most of the action items.

I hope I did not forget something important. If you have corrections to these minutes feel free to mail them to me.

Thanks a lot for your participation and help.

Cheers

Christoph

CSteier@lbl.gov