beamline experiments
  Beamline Logbook  Logged in as "Eric Dufresne" ELOG logo
Message ID: 59     Entry time: Sun Jun 13 00:09:31 2004
Author: Eric Dufresne 
Type: Other 
Category: BPM 
Subject: I turned off the valve V6 on the cryocooler and started the heater again to poss. improve stability 
Hi all, 

this entry predates the last entry of tonight i.e.
beamline-Elog 58

Last night 6/12/04 at 1h25, there was a beam dump during a bad thunderstorm.
(It rained so much that the rain water seaked through the cracks of the door
to the experimental floor between S7 and S8(most water around there) and S6
and S7.) I looked at data I had taken from dinner time and something I noticed 
pointed to something weird on the new pressure sensor of the closed loop. 

See attachment 1. In this figure, the pressure on average reads 17.7 psi, but it
suddenly dropped by 0.3 psi after the power bump and beam dump. I was concerned
that this would have meant a beam motion. More surprising was the average itself
which is now down to 17.7 psi instead of the usual 18.7 noted in the past few 
days (see /Sectors/Sector7/Operations/FY04/run3/#06-10). So not
understanding why the pressure had dropped, I decided to turn off the Omega
regulating circuit for now until I understand this better. (Note that the flux
problem has since been identified by Linda and I on
:8080/beamline/58 to a L5-20 motion). 

Attachment 2 shows time series during dinner time on 6/11/04 started at 17h49. I
modified the time series program to record data even when the 7ID-C hutch has no
X-rays inside.Attachment 2 shows the ring current and 7ID-B flux which corresponds
to approx. 9.6 x 10^12 ph/s. Note that the ring current goes down after t=7.3 hrs.
We had up to 1.4x10^13 ph/s a day ago.

Attachment 3 shows the projected beam positions in the L5-20 plane during the same
time period. I modified tserie_bpm to project the APS P1 and P2 X-ray beam
position monitor data to the L5-20 position at 26.5 m instead of the BPM in 7ID-C
at 49.2 m. The X-ray beam seems to move more in the horizontal than the vertical
with a 1 hour period which ressembles the cogging problem (every 55 min) seen in
24 bunch mode. 

After analysing the data discussed above, I closed valve V6 of the crycooler,
emptied the buffer to 81.6 % by opening briefly valve V8 (I should have done it
through V5 now that I think about it since V5 has a one way relief valve on it: it
prevents water from coming inside the buffer). I then set the pressure of the heater
to 20 psi and started a time series. 

After the power failure, the flux started to drop quickly. It dropped by 20% in
less than 2 hours. (See attachment 4, early time 0 - 1.7 hrs drop. In this attachment,
all the usual signals are present, BPM, ICs, Iring, the time series starting on 06/12
at 2h12am). 

I tweaked the second crystal at around 1.68 hours. This is shown in Attachment 5 and
6. On Attchment 5, the piezo voltage is shown versus time, in Att. 6, the 7ID-B ion
chamber signal is shown versus time. Note that I had to tweak the piezo several time
(about 5 times) before the intensity saturated again to 2.2 E5, from 1.4e5 so a gain
of 57% was obtained but only through tweaking the piezo several times. This might be an
early indication that the piezo is starting to go. We have to follow it in the future. I
was actually concerned enough that I started to plan to turn off the piezo and use the
picomotor instad. I did not do it in the end. 

The last 4 attachments show the same time series started at 2h12 on 6/12. Attachment 7
shows the cryocooler and pressure data. This plot is strange again as the pressure changes
when there is beam in the 7ID-C hutch! Note that from t = 2 to 7, and from t = 10-10.5,
the pressure is back  to 18.7 PSI whereas it is 17.7 when there is no beam in 7ID-C. This
must come from crosstalk between the channels on the IP-330 on the 7ida crate. I am pretty
convinced that this is the reason and attachment 1 is also consistent with this since the
P5 was off during attachment 1. So this is something we can look into in the future, i.e.
why is there crosstalk between channels on our IP-330? Attachment 8 shows that the X-ray BPM
data which shows that the X-rays were in the hutch only in the period discussed above. This
crosstalk might also explain why the pressure sensor never quite give 20 PSI (only 18.7PSI),
the pressure recorded on the cryocooler Love controller (pressure of the closed loop). Note that
there is a small beam position bump during the cryocooler fill at t = 4.3hrs. It not as large as in
the past when we typically had 40 um beam motion during a fill.

Attachment 9 shows the APS X-ray BPM data, projected  on the L5-20 plane 26.5 m from the source. 
The vertical projected position wanders by 8 um, while the horizontal does by 45 um. Attachment 10
shows the usual intensity plots.

So in summary, I observed some weird behavior with the piezo, requiring several tweak from 0V to 2.8V
to optimize the intensity in 7ID-B. I noted some strange behavior of the crycooler pressure, which
changes when beam is on or off in 7ID-C, which points to crosstalk with the 7ID-C X-ray BPM.
I now record the APS XBPM data and the piezo voltage in the time serie program.

Attachment 1: 061104_17h49m16s_mono.png  7 kB
Attachment 2: 061104_17h49m16s_int.png  7 kB
Attachment 3: 061104_17h49m16s_apsbpm.png  5 kB
Attachment 4: 061204_02h12m09s_int.png  8 kB
Attachment 5: 061204_02h12m09s_piezo.png  8 kB
Attachment 6: 061204_02h12m09s_7IDB-IC.png  11 kB
Attachment 7: 061204_02h12m09s_mono.png  6 kB
Attachment 8: 061204_02h12m09s_bpm.png  5 kB
Attachment 9: 061204_02h12m09s_apsbpm.png  8 kB
Attachment 10: 061204_02h12m09s_int.png  8 kB
ELOG V2.5.2