Difference between revisions of "UITF Notes"

From Ciswikidb
Jump to navigation Jump to search
 
(37 intermediate revisions by the same user not shown)
Line 1: Line 1:
−
== Opportunistic tests and long-term ideas off the top of my head ==
+
== MeV beam time ==
 +
* Test IHAM505
 +
* TM beam to 703 to see RF drift (full shift)
 +
* Verify optics setup for HKDL
 +
** To produce small beam at M703, first measure Twiss at 501 and then optimize beta_x with Elegant: try to make it small in the dipole AND at the harp if possible (keep the dipole from focusing)
 +
* Repeat buncher study, booster crested, 8 MeV, multiple scans per point
 +
** It may be interesting to do the buncher study with the 7-cell off.
 +
** This reduces microphonics impact, but optics of the low-energy beam are a problem
 +
** Crank 2-cell up?
 +
* Buncher power vs. GSET
 +
** Once correct buncher power is found, measure displacement at +/- 90 degrees on K501 again
 +
* Test booster off crest, scan buncher if results good
  
−
* Measure p spread on 703
+
== keV beam time ==
−
** Not trivial because contribution of beta function is significant
+
* LVQE scan at different voltages. Expect same result, provided different laser powers at same voltage give same result (avoid SCL).
−
** eta = length * tan(theta) = about 1.2 m
+
** Results are different, don't know if related to instrumentation (battery resistance etc.) or physics. If it's physics:
−
** sigma_x at 700 harp is about 0.5 mm. The intrinsic momentum spread is smaller than it looks; the beam moves visibly due to RF jitter, and the harp averages over that. In a sense, this is the "true" momentum spread; it depends on what we want to learn.
+
*** CST: Does thermal energy affect collection efficiency?
−
** Assuming Gaussian distributions: momentum spread = sqrt(sigma^2 - epsilon*beta) / eta, yes?
+
*** Ion current? probably too low to matter
−
** epsilon*beta = 0 would give an upper bound of 4.6e-4.
+
*** Try adjustable bias voltage source (Keithley)
−
** We know the emittance (7.3 nm at 8 MeV/c). How do we get the beta function?
+
* HVQE scan at different voltages. Expect different result.
−
*** It would be trivial if the 600 harp were placed at the same distance from the dipole as the 700 harp, but it is not.
+
* High current to FC2 for 12+ hours, check if drift is gone. Check with buncher ON and buncher OFF.
−
*** Elegant can calculate the betas directly from first principles, but we can't trust the model until we validate it.
+
* Look at [https://logbooks.jlab.org/entry/4022870 mains-harmonic BPM signals] as a function of parameters (buncher etc.)
−
*** qsUtility can measure alpha and beta upstream of Q504 using IHAM601 (already done in a test case). Elegant can predict beta at IHAM703 based on that. Sounds good enough to me? The prediction of beta_y is verifiable as the spot size in y is unaffected by dispersion.
+
* Measure bunch length upstream (chopper or Brock)
−
*** First, using Elegant, come up with a set of 501..504 quad values that will give a reasonably small beta_x at IHAM703, and verify that it is not too sensitive to the setpoints in practice. Making the intrinsic transverse size as small as possible (without it becoming unpredictable) will make sure the errors of beta and epsilon won't contribute too much to the error of the deconvolution result.
 
−
*** Do the optics study for HKDL ''first''. As soon as we know the Twiss parameters upstream of the quads and we convince ourselves that the quads do what they should, this special case becomes trivial.
 
−
* Is there anything we can learn from the BPMs / correctors to supplement the gun kick study at CEBAF?
 
−
* It would be nice to have an extra corrector before the booster to get a nicer axis through both cavities. Prefer the duct-tape variety to nothing at all
 
−
* Permanently incorporate prep chamber stuff into EPICS
 
−
** stalk heater PS, temperature readback, ion pump current, anode current. Consider protection logic to disable heater if pressure or temp gets too high. Add oven timer. It should automagically post a completion notice to UITFLOG.
 
−
** maybe also an EPICS-switchable DC voltage for the auxiliary laser diode (replaces manual "beam shutter")?
 
−
** have a second PS for the dispenser; could be remote-controlled or not, don't really care... we can't automate the whole process anyway because of the manual valve
 
−
* Can we get decent, auto-aligning corrector mounts throughout the machine? The multipole moments are currently uncontrollable, presumably large, which is a bigger deal than one might think because the beam line alignment also looks terrible.
 
−
* Measure beam parameters downstream of booster as a function of gun energy. Maybe 4 or 5 energies. (Emittance, Energy spread)
 
  
−
== Preparation for HKDL ==
+
== Random things to change when it makes sense ==
  
−
* The keV emittance can be measured at 501 provided intrinsic energy spread is negligible (buncher off). I'm not sure if it is, but at least in y the result may still be meaningful. The value of this measurement is largely academic, though, in that what ultimately matters is the beam after the booster.
+
* The way MFQK403 is supposed to focus into both Aperture A4 and the booster is flawed for multiple reasons:
−
* In preparation for the simulation of the HKDL optics, we need to know the beam ellipse at the entrance to the quads.
+
** The lens would need to produce multiple waists (DP can, A4, Brock cavity, RM11), so the focal length is a compromise. But interestingly, it is the same at CEBAF.
−
** We don't know what the booster does to the transverse phase space, and we also don't know if our model of the quads is correct. These problems need to be decoupled as follows:
+
** The orbit needs to be very straight for the beam to make it all the way downstream, but to adjust this, we need both upstream correctors, 401A and 402, so centering in the lens at the same time is very difficult.
−
*** Perform any reasonable pair of quad scans to determine the beam ellipse upstream of M501. Ignore the booster, just take the beam ellipse at its exit for what it is. What exactly "reasonable" means will be determined empirically on the first try. It can be refined once we know roughly what we are working with. Based on previous simulations of the whole beam line using Dennis's model, we should assume that the booster overfocuses at high output energy and therefore produces large betas at the first quad, so the first two quads probably can't be scanned too liberally. In principle, any quad or set thereof can be used to do this.
+
** We should do a GPT study and play with a potential extra lens, but I would intuitively suggest something like this:
−
*** Put this beam ellipse into an Elegant model starting at M501 to predict the x and y beam size at the harp as a function of all quad strengths within reasonable parts of the parameter space. Confirm these predictions experimentally. If they are consistent, that means the quads are modeled correctly. This model will then give an accurate prediction of the beta function anywhere downstream (at this energy). Otherwise, find out what is wrong (distances, wiring issues) and iterate.
+
*** Remove Brock cavity, not needed
−
** It may be useful to perform this measurement methodically for a variety of booster output energies. This could result in a separate tech note that contrasts the optical effect of the booster with what was measured at CEBAF (JLAB-TN-15-052).
+
*** Add corrector between MFQK403 and A4
−
 
+
*** I think having the beam converge slowly into the booster is good, so I like that the lens is far away. Maybe the aperture should just be closer to the booster? At the very least, put it behind the BPM so we can see where we're at.
−
== Random ==
+
* MFAK303 has a similar problem. Being single-wound, it is supposed to be equal and opposite to MFAK301, but this fixes its focal length, while it is supposed to focus in both the buncher and the dipole. This issue does not seem critical, but it could be pondered sometime.
−
 
 
−
* Harp axis calibration does not matter: It only changes the measured emittance but not alpha/beta, and its effect does not depend on the quad in use.
 
−
* Adding reasonable quadrupole moments to correctors is not enough to explain the inconsistency. It would need an extra quad with K1 ~ 5.
 
−
* At dp/p = 1e-3 (which is higher than what we observe unless the measurement is flawed), seeing significant inconsistencies in the quad scans needs a dispersion of many cm. In y, the only dispersion in this part of the lattice should come from the earth's field; this gives about 8 mm at the harp, much too low to see anything.
 
−
 
 
−
== Urgent ==
 
−
 
 
−
* Edge focusing of the 601 dipole is not well-understood yet. Solve with tracking.
 
−
* The dispersion at 703 can be measured directly (though there's not really any reason to assume it's different from theory):
 
−
** Vary momentum, steer back with dipole. This determines delta p as a function of delta GSET around the operating point without any effects from edge focusing (constant path length through the field).
 
−
** Then, apply a well-known delta p to move the beam off-center and measure the displacement with viewer and harp.
 
−
* Make sure 700 quads are at zero field! Consider degaussing, measure with probe
 
−
* Make sure 700 correctors have no quadrupole moments (realign if necessary).
 
−
* Same for 201 and 401, can be verified by measuring beam size x/y vs. steering strength
 
−
* Take home:
 
−
** an allsave of the 500 correctors (note polarity errors)
 
−
** DL field map (per Jay's email)
 
−
** harp/viewer data from the other day
 
−
* Measure earth's field, I don't recall the z component being 40 µT anywhere in the UITF
 
−
 
 
−
== Changes I would make to the machine ==
 
−
 
 
−
* The 504 viewer is unnecessary. Can I get a harp there instead?
 

Latest revision as of 13:43, 12 November 2022

MeV beam time

  • Test IHAM505
  • TM beam to 703 to see RF drift (full shift)
  • Verify optics setup for HKDL
    • To produce small beam at M703, first measure Twiss at 501 and then optimize beta_x with Elegant: try to make it small in the dipole AND at the harp if possible (keep the dipole from focusing)
  • Repeat buncher study, booster crested, 8 MeV, multiple scans per point
    • It may be interesting to do the buncher study with the 7-cell off.
    • This reduces microphonics impact, but optics of the low-energy beam are a problem
    • Crank 2-cell up?
  • Buncher power vs. GSET
    • Once correct buncher power is found, measure displacement at +/- 90 degrees on K501 again
  • Test booster off crest, scan buncher if results good

keV beam time

  • LVQE scan at different voltages. Expect same result, provided different laser powers at same voltage give same result (avoid SCL).
    • Results are different, don't know if related to instrumentation (battery resistance etc.) or physics. If it's physics:
      • CST: Does thermal energy affect collection efficiency?
      • Ion current? probably too low to matter
      • Try adjustable bias voltage source (Keithley)
  • HVQE scan at different voltages. Expect different result.
  • High current to FC2 for 12+ hours, check if drift is gone. Check with buncher ON and buncher OFF.
  • Look at mains-harmonic BPM signals as a function of parameters (buncher etc.)
  • Measure bunch length upstream (chopper or Brock)

Random things to change when it makes sense

  • The way MFQK403 is supposed to focus into both Aperture A4 and the booster is flawed for multiple reasons:
    • The lens would need to produce multiple waists (DP can, A4, Brock cavity, RM11), so the focal length is a compromise. But interestingly, it is the same at CEBAF.
    • The orbit needs to be very straight for the beam to make it all the way downstream, but to adjust this, we need both upstream correctors, 401A and 402, so centering in the lens at the same time is very difficult.
    • We should do a GPT study and play with a potential extra lens, but I would intuitively suggest something like this:
      • Remove Brock cavity, not needed
      • Add corrector between MFQK403 and A4
      • I think having the beam converge slowly into the booster is good, so I like that the lens is far away. Maybe the aperture should just be closer to the booster? At the very least, put it behind the BPM so we can see where we're at.
  • MFAK303 has a similar problem. Being single-wound, it is supposed to be equal and opposite to MFAK301, but this fixes its focal length, while it is supposed to focus in both the buncher and the dipole. This issue does not seem critical, but it could be pondered sometime.
Retrieved from "https://wiki.jlab.org/ciswiki/index.php?title=UITF_Notes&oldid=28745"