[BTeam] Hall A Energy discussion

Sat Aug 29 15:09:39 EDT 2020

Through some discussion about different ways to lock in this energy offset, I noticed that the dipole setting has been changed a few times throughout the run.  The change is on the order of 0.2% but not sure if it's significant, but wanted to bring it to the attention of those listed here.  Striptools attached.  One indication of this is having to set an energy offset in ARC2 of +1E-4 vs -1E-4 to give a similar HALLA:dpp.

The second stripchart shows changes of a similar magnitude during the run before covid hit.
________________________________
From: Matthew Poelker <poelker at jlab.org>
Sent: Tuesday, August 25, 2020 10:15 AM
To: Michael Tiefenback <tiefen at jlab.org>; Brian Freeman <bfreeman at jlab.org>; Daniel Moser <moser at jlab.org>; bteam at jlab.org <bteam at jlab.org>; Michael Aiken <aiken at jlab.org>; HyeKyoung Park <hkpark at jlab.org>; Paul Vasilauskis <vasilaus at jlab.org>; ciprian at jlab.org <ciprian at jlab.org>; Eric Forman <eforman at jlab.org>
Subject: Re: Hall A Energy discussion

with some levity....

our job: put the beam in the right place at the right time.  How's that for over simplification?

and then don't forget about "the C laser"

I liked your email Mike

Matt
________________________________
From: BTeam <bteam-bounces at jlab.org> on behalf of Michael Tiefenback <tiefen at jlab.org>
Sent: Tuesday, August 25, 2020 9:05 AM
To: Brian Freeman <bfreeman at jlab.org>; Daniel Moser <moser at jlab.org>; bteam at jlab.org <bteam at jlab.org>; Michael Aiken <aiken at jlab.org>; HyeKyoung Park <hkpark at jlab.org>; Paul Vasilauskis <vasilaus at jlab.org>; ciprian at jlab.org <ciprian at jlab.org>; Eric Forman <eforman at jlab.org>
Subject: Re: [BTeam] Hall A Energy discussion

These words were heard, floating sweetly through the summer aether:

"Can anyone explain why the compression knob is effective? Is it because the Compton is in a chicane and this magnetically compresses too? It would be interesting to perhaps have an SLM or even a viewer with in the chicane on the Compton beam lines to help? "

As I am fond of saying, and I believe to be accurate, the CEBAF accelerator runs on an "unstable fixed point."  This means that if we had a single electron at a well-defined trajectory and energy, it would generally go where we want it through the several passes and all of the vibrating and otherwise not quite perfect hardware.  The optics and RF are configured to handle, to first order, deviations in timing, trajectory and energy, and bring not quite perfectly aligned beam particles through the machine in close proximity to the "reference particle."

Largely due to M56 longitudinal residual errors, inflated energy spread results in increased bunch length, which cascades into further increased energy spread (particles falling off the RF crest).  This is an avalanche kind of thing.  When once sufficiently seeded, these particles fall out of the "bucket" and cause miscellaneous grief.  One particularly efficient "seed" in this process is bunch length tails from the injector.  The off-time particles quickly deviate in energy from the reference particle, and rapidly grow more distant from it longitudinally.  Deteriorating beam properties result in beam loss at tight apertures in the accelerator, particularly at dispersive regions.  Each Compton chicane is such an area, but there are others.

Within limits, the machine is more tolerant of energy errors at the front end than it is of excessive bunch length, but this is really just a statement that it is harder for our injector hardware configuration to control the bunch length than the energy spread.  Valerie Lebedev found that compressing the errant longitudinal particles by trading against energy error improved beam transport to the halls.  We still find this.

It is helpful to separate emittance from energy spread.  One reason I want to site high-arc SLMs in dispersive regions is to provide a non-invasive monitor allowing observation of the effects of this cascade.  The Arc 7A SLMs are targeted at an alternate and much more time-efficient means of matching the beam in the arcs, because transverse matching avoids unnecessary emittance growth.  They are at regions of minimal accessible dispersion because that helps target the transverse matching goal.  But having SLMs at dispersive locations reveals specifically a creep in energy spread.  (This is also why the 1A/2A SLMs are at specifically highly dispersive locations.) The Compton chicane counters are sensitive to particles significantly separated spatially from the reference particle.  The accelerator is less able to put longitudinally errant particles from the injector at The Right Place in that chicane than energetically errant particles.

I hope this is adequately descriptive and not uselessly simplistic.  I tend to think in geometric terms. and am prone to over-simplifying things.

Michael

________________________________
From: BTeam <bteam-bounces at jlab.org> on behalf of Brian Freeman <bfreeman at jlab.org>
Sent: Tuesday, August 25, 2020 08:14
To: Daniel Moser <moser at jlab.org>; bteam at jlab.org <bteam at jlab.org>; Michael Aiken <aiken at jlab.org>; HyeKyoung Park <hkpark at jlab.org>; Paul Vasilauskis <vasilaus at jlab.org>; ciprian at jlab.org <ciprian at jlab.org>; Eric Forman <eforman at jlab.org>
Subject: Re: [BTeam] Hall A Energy discussion

Thanks for all the great discussion everyone!

I appreciate all of the responses and helps us all to understand what is being requested and why.

Adding to the list of Daniels links is when I was able to opportunistically measure what the max compression is:
https://logbooks.jlab.org/entry/3828665
I found that between 7 and 8 degrees is max compression, going further starts to de-bunch the beam, as apparent in Arc 1. Running the the NL off crest as well certainly will confuse things even more.

Can anyone explain why the compression knob is effective? Is it because the Compton is in a chicane and this magnetically compresses too? It would be interesting to perhaps have an SLM or even a viewer with in the chicane on the Compton beam lines to help?

I plan to have a closer look at global beam loss as a function of 0L04 phase, this is also maybe useful to determine what we are doing to the machine. Perhaps if it is just reduced size through the Compton it would be best to go back and try an Optics solution again?

Thanks again,
-B
________________________________
From: Daniel Moser <moser at jlab.org>
Sent: Monday, August 24, 2020 11:08 PM
To: Brian Freeman <bfreeman at jlab.org>; bteam at jlab.org <bteam at jlab.org>; Michael Aiken <aiken at jlab.org>; HyeKyoung Park <hkpark at jlab.org>; Paul Vasilauskis <vasilaus at jlab.org>; ciprian at jlab.org <ciprian at jlab.org>; Eric Forman <eforman at jlab.org>
Subject: Re: Hall A Energy discussion

1) The PID lock used to keep HALLA:dpp at -1.1e-4, its setpoint, only adjusts the gradient in 2L19-1, its output.
2) Yes, the Inj SLM spot with 0L04 +10.4deg off-crest is ~50% of the entire horizontal width. With the NL on-crest, the spot is not quite as long on the 1A SLM, but with it off-crest further negative, as it was taken at ~two points on Saturday for Hall A Compton background rates, the spot elongates further. Here are some screen grabs of what we see on the SLMs right now, NL crested,  https://logbooks.jlab.org/entry/3833462
3)  Back on 8/19, when trying to reduce xA Compton rates, I did first try to take the 0L04 gang phase closer to crest. Rates sky-rocketed. https://logbooks.jlab.org/entry/3831043 Interested to see what would happen if we tried again right now, given the optimized steering solution (see 5)).
4) 0L04 gang phase throughout run thusfar https://logbooks.jlab.org/entry/3832620
5) Eric Deir did excellent sleuth work Saturday evening, finding times when Compton rates were best, determining what orbit we had then, and then steering to it, bringing Compton rates down to ~5kHz. https://logbooks.jlab.org/entry/3832602

Of course I typed all this up hours ago, and now we are bringing Hall C online and finding 0R08 BLM trips are a big limiter to their current. Going to send this w/ the caveat of more to come. 0L04 gang phase has been reduced by 4degs, closer to crest.

________________________________
From: Brian Freeman <bfreeman at jlab.org>
Sent: Monday, August 24, 2020 8:21 AM
To: bteam at jlab.org <bteam at jlab.org>; Michael Aiken <aiken at jlab.org>; HyeKyoung Park <hkpark at jlab.org>; Paul Vasilauskis <vasilaus at jlab.org>; ciprian at jlab.org <ciprian at jlab.org>; Daniel Moser <moser at jlab.org>; Eric Forman <eforman at jlab.org>
Subject: RE: Hall A Energy discussion

Just some clarification  about the Hall A energy discussion of this morning.

  1.  Hall A is requesting a certain energy error (dp/p) -1E-4. This is an abnormal  request.
  2.  With FFB on in Energy mode,  or when running Arc 2 it does as designed and brings the error even closer to zero down to the E-5 range. So not sure what this request is about, and I believe this abnormal request is the root of the Hall A issue. (Cipran ?)
  3.  This issue with monitoring the SLMs, is that we are running 0L04 so far off crest that the spots almost encompass the entire image frame:
     *   https://logbooks.jlab.org/entry/3832990
     *   This was done to reduce Compton rates.
     *   I suggest we back out of the 0L04 change.

BTeam items for tomorrow's meeting? Yves you still on vacation?

Any others chime in, please.

-Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20200829/90edd58b/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HADipole.png
Type: image/png
Size: 45323 bytes
Desc: HADipole.png
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20200829/90edd58b/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HADipole1.png
Type: image/png
Size: 42760 bytes
Desc: HADipole1.png
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20200829/90edd58b/attachment-0003.png>